Dear Son..

About Dear Son..

The books, frameworks and ways of thinking I most want my son to encounter, preserved before I forget why they mattered. tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast)

Publisher

tiwaryshailesh

Category

education · science

Language

en

Explicit

No

First episode

12 Jun 2026

Latest episode

24 Sept 2026

Latest episodes

15 episodes in the feed.

1. 

   24 Sept 2026

   On Military Preparedness in AI Era

   There is a mode of thinking that treats military training not as preparation for war but as the only mechanism an armed force possesses for converting information about the world into behaviour under fire. Most people, including most soldiers, see training as a pipeline - identify the tasks, write the standards, run the repetitions, certify the unit, deploy. On that view the arrival of drones, machine learning and offensive cyber operations is a content problem. New tasks appear, new standards are written, new courses are stood up and the pipeline absorbs them. The person who has internalised this field sees something else entirely. Training is the process by which an organisation manufactures experience in advance of the events that would otherwise supply it and the value of manufactured experience depends entirely on whether the world it simulates still exists by the time the trainee meets it. The central insight, the one that unlocks most of what follows, is this: expertise can only form where cues reliably predict outcomes and where feedback arrives fast enough to correct error; the defining feature of drone, cyber and algorithmic warfare is that an intelligent adversary is deliberately and continuously breaking that relationship. Everything else in this field is a consequence. Once the environment stops repeating itself, the quantity of skill a force holds matters less than the rate at which it can generate, validate and distribute new skill. Training stops being a pipeline and becomes a learning rate and a learning rate is an organisational property rather than an individual one. The Manufacture of Experience: What Military Training Actually Produces Before anything else can be said, the field needs a working account of what happens inside the head of a competent operator, because training is defined by what it is trying to build. The most useful answer comes from Gary Klein, whose Sources of Power: How People Make Decisions established that experts in time-pressured domains do not generate and compare options. They recognise a situation as an instance of a type, retrieve the course of action typically associated with that type, mentally simulate it briefly and act. Klein called this recognition-primed decision making and it reframes the purpose of training completely. Training is not the installation of procedures. It is the stocking of a pattern library, the accumulation of cases against which a novel situation can be matched, plus the capacity to run a quick simulation of what would happen next. This connects directly to the work of K. Anders Ericsson, whose 1993 paper on deliberate practice and later synthesis in Peak: Secrets from the New Science of Expertise identified the conditions under which such libraries grow: tasks slightly beyond current capability, immediate and specific feedback, repeated attempts with correction and a coach or standard that defines what correct looks like. Ericsson’s framework explains why so much military time labelled training produces so little learning. Standing in a queue at a range firing at static targets with a scorecard delivered an hour later satisfies almost none of these conditions. It produces exposure, which is not the same thing. Science, Strategy and War: The Strategic Theory of John Boyd by Frans PB Osinga supplies the third piece of the vocabulary, though he is routinely misread. The OODA loop is popularly treated as a claim about speed, as if the side that cycles faster wins. Osinga shows that Boyd’s actual argument centres on orientation, the phase in which observation is interpreted and that orientation is constituted by prior experience, cultural tradition, training and analytical habits. Boyd’s point in Destruction and Creation is that orientation must be repeatedly broken apart and reassembled, because the schemas that make rapid recognition possible are the same schemas that make an operator blind when the world changes. Training, on this reading, both builds and endangers. It is the deposit of patterns that permit fast action and the source of the frames that will eventually mislead. Sitting underneath all three is a distinction the field treats as foundational and that institutions persistently blur: the difference between training and education. Training prepares a person for problems whose shape is known. Education prepares a person for problems whose shape is not. Donald Schön’s The Reflective Practitioner describes the second as reflection in action, the capacity to reframe a problem while inside it when the standard framing fails. Every serious argument about military learning eventually reduces to the ratio between these two and the unconventional era shifts that ratio sharply toward the second without any corresponding shift in institutional budgets, which remain overwhelmingly committed to certifying known tasks. Validity and Its Collapse: Why the Environment Decides What Can Be Learned The single most important paper in this entire field was not written about the military. In 2009 Daniel Kahneman and Gary Klein published Conditions for Intuitive Expertise: A Failure to Disagree, an adversarial collaboration between the leading sceptic of intuition and its leading defender. They resolved their dispute by locating the disagreement in the environment rather than the person. Intuitive expertise develops when two conditions hold. The environment must have high validity, meaning stable and lawful relationships between available cues and later outcomes. And the practitioner must have the opportunity to learn those regularities through adequate feedback. Where validity is high and feedback is fast, as in firefighting, anaesthesia or chess, intuition becomes reliable and experience is worth what it costs. Where validity is low, as in long-horizon political forecasting or stock picking, confidence grows with experience while accuracy does not. Apply that test to the emerging domains and the result is stark. Flying a First-Person-View (FPV) drone through a treeline is a high-validity task with instant feedback and it can be trained to a very high standard very quickly, which is exactly what has happened across Ukrainian and Russian units since 2022. Choosing what to do about an adversary’s electronic warfare posture is not. In cyber operations the mapping between cue and outcome is not merely noisy, it is authored by an opponent who observes your responses and changes it deliberately. Ben Buchanan’s The Hacker and the State makes the texture of this clear: cyber operations are slow, intelligence-driven, craft-based and heavily dependent on accumulated familiarity with specific target systems, which means the transferable component of the skill is smaller than institutions want to believe. The same non-stationarity applies to counter-drone work, where effective countermeasures have repeatedly had operational lifetimes measured in weeks before the opposing side re-engineers around them. The consequence is a division that should govern every training investment decision. Sub-skills within these domains are frequently high-validity and should be trained hard, fast and to genuine mastery using Ericsson’s conditions. Operational judgement within the same domains is low-validity and cannot be trained to intuition at all, because the regularities the intuition would encode will not survive contact. For that layer, what is being built is not a pattern library but the capacity to detect that one’s pattern library has expired. That capacity is not a skill in the ordinary sense. It is a disposition, sustained by institutional mechanisms rather than by individual repetition, which is why the rest of this essay is largely about organisations. The Modern System: Why Force Employment Is Fundamentally a Training Problem The strongest empirical argument that training determines combat outcomes comes from Stephen Biddle’s Military Power: Explaining Victory and Defeat in Modern Battle. Biddle tested numerical preponderance, technological advantage and force employment against the historical record and found that how forces fight explains outcomes better than what they have or how many they are. He named the winning approach the modern system: cover, concealment, dispersion, suppression, small-unit independent manoeuvre, combined arms integration, depth of defence and the withholding of reserves. The system dates from the tactical crisis of 1917 and its central property, for present purposes, is that it is extraordinarily difficult to execute. It demands junior leaders who act without orders, sub-units that coordinate weapon effects under stress and formations that accept dispersion at the cost of control. This is why it diffuses so slowly. It cannot be purchased. It can only be trained and the political preconditions for training it, including tolerance of junior initiative, are absent in many states. Biddle’s framework also explains what the drone era does and does not change. Cheap persistent sensors attack concealment, which is one pillar of the modern system and the result across Ukraine has been a visible intensification of the other pillars: greater dispersion, deeper defences, smaller assault elements and more reliance on decoys and signature management. The modern system is being stressed at one point and reinforced at every other. It has not been overturned, which means that the training burden it imposes has increased rather than disappeared. Martin van Creveld’s Fighting Power: German and U.S. Army Performance 1939 to 1945 supplies the organisational mechanics of how such capability gets built or squandered. His comparison of the German and American armies of the Second World War attributes German tactical resilience not to doctrine on paper but to the machinery around it: unit-based rather than individual replacement, regimental training depots, officers who had commanded at the level below the one they taught and a personnel system that treated combat units as the centre of gravity rather than the administrative periphery. Anthony King’s The Combat Soldier: Infantry Tactics and Cohesion in the Twentieth and Twenty-First Centuries extends the point into the present and corrects a persistent myth. Cohesion in professional forces is not primarily an affective bond between friends. It is task cohesion produced by shared, rehearsed and mutually legible drills. Soldiers trust each other because they know precisely what the other will do. That is a training product, not a social accident and it degrades in exactly the situations where training has been thin. The doctrine most often invoked as the answer to uncertainty is mission command and Eitan Shamir’s Transforming Command: The Pursuit of Mission Command in the U.S., British and Israeli Armies is the necessary corrective. Shamir traces how Auftragstaktik was adopted in name by the American, British and Israeli armies while the underlying conditions were not transplanted: shared tactical schooling that made subordinates predictable to superiors, a culture that tolerated disobedience in pursuit of intent and promotion systems that rewarded it. Stephen Bungay’s The Art of Action names what mission command is meant to close, three gaps created by friction, between what you would like to know and what you do know, between what you want people to do and what you tell them and between what you expect to happen and what actually happens. Bungay’s insight is that these gaps cannot be closed by more detailed instruction and that attempts to do so widen them. Cheap, pervasive video from the forward edge makes the temptation to close them by supervision nearly irresistible. The literature predicts the result: higher-level commanders watching tactical feeds and intervening, subordinates who stop exercising judgement and an organisation whose decision cycle is bounded by the bandwidth of one headquarters. Free Play and Honest Feedback: The Intervention That Reliably Works If this field has one robust empirical finding, it is this: instrumented, free-play, force-on-force training against a competent adversary, followed by non-attributional debriefing based on objective data, produces large improvements in combat effectiveness. Everything else is contested. This is not. The canonical cases are all from the 1960s and 1970s. The Ault Report of 1968 examined the United States Navy’s poor air-to-air performance over Vietnam and led to the Navy Fighter Weapons School, where crews flew against dissimilar aircraft flown by pilots trained in adversary tactics and then debriefed at length without rank protection. The Air Force’s Project Red Baron analysis produced the finding that a crew’s first ten combat missions carried disproportionate risk, which led directly to Red Flag in 1975 and the logic of providing those first ten missions in Nevada rather than over hostile territory. The National Training Center at Fort Irwin applied the same architecture to ground manoeuvre from 1980: a permanent, expert opposing force, laser instrumentation that made engagement outcomes objective rather than arguable and an after action review conducted by observer-controllers in which a colonel could be shown to have lost. The mechanism matters more than the institutions. Free play supplies an adversary who is not required to lose, which is the only way an exercise generates information rather than confirmation. Instrumentation supplies ground truth, which removes the negotiation over what happened. Non-attributional debriefing supplies the psychological conditions under which people will state what they actually did, which is where Amy Edmondson’s work on psychological safety in The Fearless Organization becomes load-bearing rather than decorative. Chris Argyris adds the depth dimension with the distinction between single-loop and double-loop learning: a debrief that corrects the execution of a plan is single-loop, while a debrief that questions whether the concept itself was sound is double-loop and organisations overwhelmingly stop at the first because the second implicates people who are present in the room. Peter Perla’s The Art of Wargaming carries the same logic into the analytical space and makes a claim that is routinely ignored. Wargames do not predict. Their output is not the result but the experience of the participants and the questions the game exposes. Williamson Murray and Allan Millett’s Military Innovation in the Interwar Period demonstrates the value of taking that seriously: the forces that innovated successfully between the wars, in carrier aviation, amphibious operations and combined arms mechanisation, were those that ran honest experiments and accepted unwelcome findings, while those that used exercises to validate existing commitments learned nothing they did not already believe. This is the historical argument that the quality of a force’s training feedback loop predicts its performance a decade later. The Pathologies: Scripted Exercises, Captured Metrics and the Hollowing of Realism Given that free play with honest feedback works, the interesting question is why it is so rare. The field’s answer is that every institutional incentive pushes against it and that the resulting failure modes are stable and recognisable. The first is scripting. An exercise designed to demonstrate a capability cannot also test it. Millennium Challenge 2002 remains the clearest documented instance: Paul Van Riper, commanding the opposing force, used motorcycle couriers and coded signals to defeat network-centric surveillance and a swarming small-boat attack to inflict heavy notional losses on the blue fleet, after which sunk ships were refloated and constraints imposed on the red commander so the intended concept could be exercised. The exercise had a purpose other than learning and that purpose won. The pattern generalises to any event whose outcome is reported upward as evidence of readiness. The second is metric capture. Readiness reporting systems require quantified statements about training status and the quantity that can be counted is repetitions completed against a task list. Goodhart’s observation applies with full force: once completion of the task list becomes the target, units optimise for completion. The most damaging version is the certification event that a unit cannot afford to fail, which reliably produces a rehearsed performance rather than a test. In Evaluating Training Programs - The Four Levels, Donald Kirkpatrick’s four levels of training evaluation, reaction, learning, behaviour and results, offer an uncomfortable diagnostic here, because most military training evaluation never rises above the second and frequently rests at the first. Whether trainees enjoyed the course and whether they passed the end-of-course test tell you almost nothing about behaviour under fire. The third is the slow hollowing of realism by risk aversion. Charles Perrow’s Normal Accidents and Diane Vaughan’s The Challenger Launch Decision describe how organisations under pressure gradually redefine what counts as acceptable, with Vaughan’s normalisation of deviance naming the process by which an anomaly becomes routine. The military training version runs in the opposite direction with the same mechanism: each safety restriction is reasonable in isolation, each removes a small amount of realism and the cumulative effect after two decades is an exercise environment in which nothing dangerous, confusing or unscripted can occur. The cost is invisible because the counterfactual is never observed until a war begins. The fourth is cultural and is the subject of Norman Dixon’s On the Psychology of Military Incompetence, which despite its age remains the sharpest treatment of why hierarchical, authoritarian institutions select against the traits that learning requires: intellectual curiosity, willingness to be wrong in public and tolerance of subordinates who contradict superiors. Eliot Cohen and John Gooch’s Military Misfortunes provides the analytical frame that ties these together. Catastrophic failure is rarely a single error. It is the aggregation of a failure to learn from the past, a failure to anticipate the future and a failure to adapt in the present and their point is that these failures are systemic rather than personal, which means removing the commander does not fix them. The Ironies of Automation: What Machines Do to the People Who Supervise Them Every claim made about human-machine teaming in the last five years was anticipated by Lisanne Bainbridge in a four-page paper published in 1983. Ironies of Automation observes that the designer who automates a system removes the tasks that can be automated and leaves the human with those that cannot, which are by definition the hard ones. The human is then asked to monitor a process that is almost always correct, a task at which humans are known to be poor and to take over competently at the moment it fails, despite having lost through disuse the manual skill that takeover requires. Automation thus creates a requirement for more training, not less and for training in precisely the rare degraded modes that the automation makes rare. The military instance that should be on every syllabus is the Patriot air defence fratricides of 2003, in which the system engaged and destroyed a Royal Air Force Tornado and a United States Navy F/A-18. John Hawley’s analysis for the Army Research Laboratory and his subsequent Patriot Wars report attribute the failures not to software error alone but to a training and operational culture that produced operators conditioned to trust system classifications, with little basis for challenging them, inside engagement timelines that allowed no meaningful deliberation. The crews were trained to operate the system. They were not trained to disbelieve it. John Lee and Katrina See’s Trust in Automation: Designing for Appropriate Reliance supplies the concept that resolves this. The training objective is not trust and it is not scepticism. It is calibration, the alignment of the operator’s confidence with the system’s actual competence across specific conditions. Calibration requires the operator to have seen the system fail, repeatedly, in the specific ways it fails, which means training must deliberately expose people to machine error rather than presenting the system at its best. David Woods and Erik Hollnagel’s Joint Cognitive Systems generalises this into the notion of automation surprises and brittleness: automated systems perform well inside their design envelope and degrade abruptly at its edge and the human is left holding the consequences of a boundary they cannot see. Sidney Dekker’s Field Guide to Understanding Human Error completes the argument by showing that labelling the resulting outcome operator error terminates the investigation at the point where the useful learning begins. For machine learning systems the problem sharpens. A classifier’s failure envelope is not documented, is not stable across updates and may be adversarially manipulable. The operator cannot be taught the boundary because nobody knows it. What can be trained is the habit of maintaining an independent estimate of the situation and noticing divergence, together with the organisational permission to act on that divergence when the machine is confident and you are not. Whether such permission exists is a matter of command climate and promotion policy, not of curriculum. Desirable Difficulties: Why Effective Training Feels Like Failing Training The most counterintuitive body of knowledge in this field comes from cognitive psychology and contradicts the intuitions of nearly every instructor. Robert Bjork and Elizabeth Bjork’s work on desirable difficulties, summarised in Making Things Hard on Yourself, But in a Good Way, establishes that performance during training and learning are different things and are frequently in tension. Conditions that slow acquisition and increase errors during practice, including spacing repetitions apart in time, interleaving different tasks rather than blocking them, varying the conditions of practice and requiring retrieval from memory rather than re-presentation, consistently produce better long-term retention and better transfer to novel situations. Conditions that make practice look successful, massed blocked repetition under constant conditions with the instructor’s example available, produce rapid apparent mastery that decays. The implications are severe because training organisations are evaluated on the visible output of the training event. An instructor whose students perform smoothly at the end of the week is rewarded. An instructor whose students struggle all week and perform better six months later in an unfamiliar setting is not observed. This bias is amplified by simulation. Synthetic environments make it easy to deliver high-repetition, low-variability, well-lit practice that generates excellent in-simulator performance and poor transfer. Which raises the fidelity fallacy. The instinct is that a better simulator is a more physically accurate one. The team training research, particularly the body of work by Eduardo Salas and Janis Cannon-Bowers, points elsewhere. What matters is cognitive and psychological fidelity, whether the simulation elicits the same decisions, the same uncertainty, the same coordination demands and the same stress as the real task. A tabletop exercise with a hostile umpire can have higher cognitive fidelity than a photorealistic simulator with a scripted enemy. Salas’s evaluation of Crew Resource Management, the most widely adopted team training intervention in history, also identifies what actually transfers in team settings: shared mental models, explicit cross-checking, assertion by junior members and closed-loop communication. These are the behaviours that determine whether a distributed team of humans and machines degrades gracefully and they are trainable. Stress is the final variable. James Driskell and Joan Johnston’s work on stress exposure training establishes a three-stage approach, information about the stressors and their effects, skill acquisition under moderate stress and practice under conditions approximating the real thing, with the finding that exposure must be graduated to be useful because overwhelming stress early degrades learning rather than inoculating against it. Dave Grossman’s On Combat popularised parts of this and is widely read in military and police circles, but its physiological claims and its heart-rate zone model are contested and it should be treated as illustrative of a training culture rather than as evidence. Adaptation Rate: The Capability That Actually Matters Now Everything above converges on a single organisational variable. Williamson Murray’s Military Adaptation in War argues that adaptation during a conflict matters more to its outcome than peacetime innovation does, because peacetime assumptions are reliably wrong and the question is how fast they get corrected. Adaptation rate depends on three things: the speed with which information about what is failing reaches people who can change something, the speed with which candidate solutions can be generated and tested and the speed with which validated solutions propagate to everyone who needs them. Training touches all three. It is the propagation mechanism and it is also the test bed. The Ukrainian and Russian experience since 2022 is the working illustration and the interesting feature is not the drones but the cycle time. Countermeasure and counter-countermeasure exchanges in electronic warfare and first-person-view drone employment have repeatedly run on timescales of weeks. A force whose training system requires nine months to update a programme of instruction is structurally incapable of competing in that exchange, regardless of the quality of its instruction. What has emerged instead, documented in RUSI work by Jack Watling and Nick Reynolds and in the analyses of Michael Kofman and Franz-Stefan Gady, is a bifurcated system: unit-level and volunteer-run schools that iterate in days and a formal institutional system that lags badly, alongside a persistent basic training crisis driven by the need to push large numbers of mobilised personnel through quickly. T. X. Hammes had anticipated part of the underlying economics in Technologies Converge and Power Diffuses, arguing that small, smart and cheap systems would shift advantage toward whoever could field and iterate in quantity. What his argument implies for training is less often noticed: when the cost of a repetition collapses, as it has for small drone piloting, the binding constraint moves from the cost of practice to the speed of curriculum revision and the throughput of qualified instructors. Red teaming is the deliberate manufacture of adaptation pressure and Micah Zenko’s Red Team is the best treatment of both its power and its corruption. Zenko’s finding is that red teams are effective in proportion to their independence and are systematically defanged: given restricted scope, denied access, subordinated to the organisation they are meant to test or staffed by people whose careers depend on the tested organisation’s approval. The Millennium Challenge case is the canonical military example of the same disease. A red team that cannot win is a ceremony. Finally, personnel policy is a training variable and is usually excluded from the conversation. Stephen Rosen’s Winning the Next War argues that military innovation succeeds when senior officers create viable career paths in the new speciality, so that talented juniors can become experts without sacrificing promotion. In cyber, electronic warfare and machine learning the relevant expertise takes years to build, is largely tacit, is acquired through unstructured exploration and is highly portable to industry. Tim Kane’s Bleeding Talent documents how rigid rotation, up-or-out promotion and seniority-based assignment destroy exactly this kind of accumulation. An organisation that rotates its best cyber operator into a staff job after two years has made a training decision, whatever it calls it. Buchanan’s account of how operational cyber capability depends on deep familiarity with specific targets makes the cost concrete: the expertise was never in the curriculum, it was in the person and the person has left. The Unified Picture The single most important insight in this field is that combat effectiveness is a function of an organisation’s learning rate and that learning rate is bounded by the quality of the feedback loops the organisation is willing to tolerate. Every load-bearing idea in this essay is an expression of that proposition. Kahneman and Klein state the precondition: learning requires an environment where feedback is informative and the adversary’s central project is to make it uninformative. Klein’s recognition-primed decision making identifies what the loop deposits, a library of patterns that enables fast action and that expires. Boyd’s orientation names the same deposit from the inside and adds the requirement to destroy and rebuild it. Ericsson specifies the conditions under which any individual loop functions and therefore why most military training does not qualify. Biddle establishes that the thing being learned, the modern system of force employment, is difficult enough that only sustained high-quality loops can install it, which is why it predicts outcomes better than equipment does. Van Creveld and King show that the loop operates at unit level rather than individual level and that cohesion and tactical competence are the same product viewed from different angles. Shamir and Bungay explain why decentralised execution is the only structure in which local feedback can be acted on locally and why pervasive surveillance tempts organisations to destroy that property. The free-play and after action review architecture of Top Gun, Red Flag and the National Training Center is simply the highest-quality feedback loop anyone has built, which is why it is the field’s only robust empirical result. Bjork explains why good loops feel bad, since error during practice is the signal being harvested rather than a defect to be suppressed. Bainbridge, Lee and See, Hawley and Woods describe what happens when a machine is inserted into the loop and absorbs the feedback that the human needed. Zenko, Edmondson, Argyris, Dixon, Vaughan and Cohen and Gooch catalogue the ways institutions sever their own loops to protect careers, concepts and programmes. Murray and the Ukrainian case reduce the whole thing to a measurable quantity, which is cycle time. The twenty percent worth carrying is therefore small and tightly connected. Environmental validity determines what can be learned at all. Feedback quality determines the rate. Free play with an adversary permitted to win and objective instrumentation with non-attributional debriefing is the known mechanism for maximising that rate. Force employment quality, not equipment, is the dependent variable that decides battles. Decentralisation is the structural precondition for local adaptation. Automation relocates the feedback and thereby creates new failure modes in the human. And institutional incentives, including metrics, safety, careers and programme advocacy, are the standing threat to all of it. Those six propositions explain most of what the literature contains and most of what happens in practice. The Limits The field’s confidence exceeds its evidence in several specific places and the mode of thinking has failure modes of its own. The empirical base is narrow and selected. The canonical demonstrations of free-play training come from a small set of wealthy Western professional forces during a particular period and the strongest single claim, that Top Gun transformed Navy air-to-air exchange ratios, is confounded. Rules of engagement changed, the Combat Tree identification system and the Teaball control facility changed, the adversary’s posture changed and the Air Force, which did not have an equivalent school at the time, faced different conditions. The training intervention is probably real and it is probably not as large as the folklore holds. Building an entire doctrine of learning on a handful of such cases is exactly the kind of inference the field criticises elsewhere. The learning-organisation literature verges on unfalsifiability. Psychological safety, honest feedback culture and adaptive capacity are attributed after the fact to whoever won. Any outcome can be explained by them and therefore they predict little on their own. When Edmondson’s construct is used to explain a defeat in retrospect it is doing rhetorical rather than analytical work. Excellence in training cannot compensate for a wrong theory of victory and the field systematically understates this because Biddle deliberately brackets strategy out of his model. The German army of 1939 to 1945 is the standing refutation, tactically and operationally superb by almost any measure of force employment and comprehensively defeated by a strategy that could not have succeeded. Murray’s own work on German military effectiveness makes the point. A force that invests exclusively in learning how to fight better can accelerate itself toward a destination it should not have chosen. The twenty percent logic itself creates a blind spot. It privileges what recurs, what is measurable and what can be practised, which means it systematically undervalues rare-event competence. Nuclear command and control, mass casualty medical response and the first forty-eight hours of an unexpected theatre are precisely the situations where experience cannot accumulate and where the pattern library is empty by construction. The field has little to say about them beyond generic exhortations to educate rather than train. There is a real and underexamined trade between depth and throughput. Almost the entire literature was produced by and for small professional volunteer forces, for which quality per soldier is the natural optimisation target. An attritional war with mobilisation demands the opposite: adequate competence delivered to very large numbers very fast, accepting a lower ceiling. Ukraine’s basic training difficulties are not a failure to read the training literature. They are a situation the training literature was not written about and importing depth-optimised assumptions into a throughput problem produces bad advice. Finally, simulation and synthetic training carry a specific hazard that the fidelity debate does not fully capture. Trainees learn the simulation, including its physics artifacts, its scenario logic and the exploitable behaviour of its scripted or machine-learned opponents. Where the synthetic adversary is itself an optimising system, there is a live risk of teaching a force to defeat a model rather than an enemy, with a confident and well-drilled result that fails on contact. The more sophisticated the synthetic environment, the more plausible its errors become. What This Demands of You Start by asking, of any training activity you observe or design, what the feedback loop is and how long its cycle takes. If you cannot identify the moment at which a participant learns that they were wrong and cannot state how long after the error that moment arrives, no learning is being produced regardless of how the activity is labelled or how much it costs. Ask whether the task you are trying to train sits in a high-validity or a low-validity environment. Do the cues reliably predict the outcomes and how fast is an adversary changing that mapping. Train the high-validity components hard and to real mastery. For the low-validity components stop trying to build intuition and start building the machinery that detects when an intuition has expired. Ask of every exercise who is permitted to lose. If the opposing force cannot win, if the unit cannot fail the certification, if the outcome has already been reported upward, you are watching a demonstration. Exercises that produce no unwelcome information have produced no information. Separate performance from learning in your own judgement. When a training session looks smooth, treat that as weak evidence of teaching and possible evidence of a session made too easy to generate transfer. When participants struggle, vary conditions, get lost and are forced to retrieve rather than recognise, expect better retention and worse reviews. Be prepared to defend the second against people who saw the first. With any automated or algorithmic system, ask what the operator will be doing at the moment it fails, whether they have ever seen it fail in that way and whether they retain the manual skill their takeover assumes. Then ask the harder question: if the operator disagrees with the machine and is wrong, what happens to their career and if they disagree and are right, does anyone find out. Ask where expertise is stored in your organisation. If it lives only in individuals, calculate its half-life against your posting cycle. If it lives only in documents, ask when a document last changed a behaviour. If it lives in units, ask what happens when the unit is reconstituted. Ask what your metrics are doing to behaviour rather than what they are measuring. Every readiness statistic is an instruction to somebody and you should assume it is being followed literally. Ask what the adversary learns by observing you train, because your exercises are now visible and ask what you would learn about yourself if you were watching. And treat cycle time as your default measure of capability. Not the number of trained personnel, not the hours delivered, not the platforms fielded, but the interval between the detection of a problem in the field and the arrival of a validated corrective in the hands of the last unit that needs it. That number is the thing this entire field is about and almost nobody measures it. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

2. 

   23 Jul 2026

   On Revolution

   There is a mode of thinking that treats a revolution not as an explosion of accumulated misery but as a coordination problem that was suddenly and often invisibly solved. Most people read a revolt backward. They see a regime fall, assume the anger that toppled it had been rising toward a threshold and conclude that suffering plus injustice eventually produces revolt the way heat plus pressure produces steam. This is the volcano model and it is close to worthless. It cannot explain why the same grievance sits quiet for decades and then ignites in a week, why the poorest and most oppressed are rarely the first to rise, or why almost no expert forecast the fall of the Shah, the collapse of the Soviet bloc or the Arab uprisings of 2011. The person who has internalised this field sees what the volcano model hides - grievance is nearly constant and what actually varies is the capacity to act together, the belief each person holds about what everyone else secretly believes and the willingness of the state’s own agents to keep obeying. The single insight that unlocks most of the field is this. A revolution is a withdrawal of cooperation, cooperation runs on beliefs about other people’s beliefs and those beliefs are hidden until the moment they are not. Understand that and the field’s central puzzles, from the free rider problem to the sudden unpredictability of collapse, stop being separate mysteries and become one. The Grammar of the Field: Why the Question Is Action, Not Anger The foundational move, the one without which nothing else in the field parses, is to stop asking why people are angry and start asking why they act. Discontent is everywhere and always. Charles Tilly built his career on demolishing the assumption that grievance explains anything, arguing across works from The Rebellious Century to From Mobilization to Revolution that collective action is produced by organisation, resources, opportunity and repertoires of contention rather than by the intensity of suffering. His concept of the repertoire is load-bearing. People do not invent protest from raw emotion. They draw on a limited, historically specific menu of learned forms, the strike, the petition, the barricade, the occupation and that menu changes slowly across centuries. The Tilly programme, extended with Sidney Tarrow and Doug McAdam into the political process synthesis of Dynamics of Contention, replaced the psychology of frustration with a sociology of capacity. The reason capacity is the hard part and not anger was established by Mancur Olson in The Logic of Collective Action. Olson showed that a shared grievance does not produce shared action, because the benefits of a successful revolt, freedom, a new regime, lower taxes, are public goods that flow to everyone whether or not they personally took the risk. The rational individual therefore free rides, letting others face the bullets while reserving the reward. If everyone reasons this way no one moves and the regime survives on a discontent so universal it is politically inert. This is the collective action problem and it is the true engine room of the field. Every serious theory of revolution is, at bottom, an account of how this problem gets solved - through selective incentives, through dense social networks, through organisation, through emotion or through a shift in what people believe others will do. Grievance is the precondition that explains almost nothing. Solved coordination is the variable that explains almost everything. The State Is the Real Variable: Structure Over Will Once the question becomes capacity rather than anger, attention moves to the entity whose capacity matters most, the state. Theda Skocpol’s States and Social Revolutions is the pivot of the modern field and its central claim is deliberately deflating. Revolutions are not made, they come. The great social revolutions in France, Russia and China were not caused by revolutionary vanguards willing them into being but by the breakdown of state administrative and coercive power under the combined pressure of international military competition and the resistance of landed elites to reform. When a state loses a war, goes bankrupt paying for one or fractures between its rulers and its own upper class, the coercive apparatus that normally makes revolt suicidal stops functioning and the peasantry, organised in autonomous village communities, pours through the gap. The revolutionaries mostly arrive to administer a collapse they did not cause. This structural logic runs backward through Barrington Moore’s Social Origins of Dictatorship and Democracy, whose comparative account of how landlord and peasant class configurations produced democratic, fascist or communist paths yielded the compressed dictum “no bourgeoisie, no democracy,” and forward through Jack Goldstone’s Revolution and Rebellion in the Early Modern World, which added the demographic engine. Goldstone showed that sustained population growth in early modern agrarian states outran their fixed fiscal and administrative capacity, producing state fiscal crisis, intra-elite competition for a stagnant pool of offices and rising popular grievance at once, a conjunction that recurred across England, France, the Ottoman Empire and China. The unifying claim of the structural school is that the timing and success of revolution track the condition of the state and its elite far more tightly than the condition of the masses. Alexis de Tocqueville had already seen the sharpest edge of this in The Old Regime and the Revolution. The French monarchy fell not when oppression was worst but when it was easing, because reform reveals a regime’s weakness, raises expectations faster than it can meet them and turns endurable conditions into intolerable ones. The most dangerous moment for a bad government is when it begins to reform. The Deprivation Fallacy: The Seductive Wrong Theory The field’s most instructive dead end is the theory that feels most obviously true. Ted Robert Gurr’s Why Men Rebel gave rigorous form to the intuition that revolt springs from frustration, building on Davies’s J-curve hypothesis that revolutions occur when a long period of rising prosperity is followed by a sharp reversal, opening an intolerable gap between what people have come to expect and what they now get. This is relative deprivation and its appeal is that it matches how injustice feels from the inside. It failed as an explanation for a precise reason. Deprivation, relative or absolute, is so widespread that it cannot discriminate between the societies that revolt and the many more that endure identical or worse conditions in silence. Tilly’s critique was decisive. Frustration is a constant, action is a variable and you cannot explain a variable with a constant. The mismatch does not tell you when people will mobilise, who will lead, whether they can overcome the free rider problem or whether the state will crack. Relative deprivation survives as a description of the emotional texture of grievance, not as a theory of revolution and its long dominance is the field’s standing warning that the most psychologically satisfying account is often the least predictive. The same warning applies to the oldest psychological tradition, the crowd theory of Gustave Le Bon’s The Crowd, which portrayed the revolutionary mass as a single irrational organism in which the individual dissolves, loses judgement and is swept by contagion into barbarism. Le Bon’s model flattered elites by pathologising dissent and it has been dismantled by empirical crowd research showing that riots and revolts are patterned, socially structured and governed by shared norms and grievances rather than by mindless contagion. Its residue lingers wherever a commentator calls a protest a mob and thinks the word is an explanation. Elias Canetti’s Crowds and Power is the richer literary descendant of this tradition and even it is better read as phenomenology than as causal theory. Preference Falsification: Why the Volcano Was Always Hollow Here the field turns counterintuitive and here sits the concept that does more explanatory work than any other. Timur Kuran’s Private Truths, Public Lies introduced preference falsification, the act of misrepresenting one’s true preferences under perceived social or political pressure. In a repressive regime almost everyone conceals their real opposition, professing loyalty they do not feel, because the cost of speaking out is high while the effect of one voice is nil. The consequence is a society whose private discontent is enormous while its public opinion appears solidly loyal and, crucially, no one can see the discontent, not the citizens, not the rulers, not the analysts. Everyone underestimates how many others share their views. This solves the central paradox the volcano model could not touch. It explains why stable dictatorships and impending collapse look identical from the outside right up until the end. The mechanism of collapse is the revolutionary cascade. Kuran, drawing on the same logic Mark Granovetter formalised in his threshold models of collective behaviour, showed that each person has a private threshold, the number of others they must see resisting before their own fear is outweighed and they too go public. Because thresholds are hidden and distributed, a small trigger that moves a few low-threshold individuals into the open can lower the perceived risk for the next tier, whose defection lowers it for the next, in a chain that can flip a society from apparent unanimity of loyalty to open revolt within days. The 1989 collapse of Eastern Europe, which Kuran analysed under the title Now Out of Never is the paradigm case. This yields the field’s hardest truth, argued directly by Charles Kurzman in The Unthinkable Revolution in Iran. Revolutions are structurally unpredictable, not merely hard to predict. When everyone is watching everyone else to decide whether to move, the outcome depends on a web of mutually referential expectations that can tip on an accident and the retrospective sense that the fall was inevitable is a cognitive illusion, hindsight reconstructing a cascade that could easily have fizzled. The feeling of inevitability after the fact is itself a symptom to distrust. Hidden Transcripts: Dissent Where It Cannot Be Seen If preference falsification means public loyalty conceals private opposition, the question becomes where that concealed opposition lives and James C. Scott mapped its terrain. Preference falsification means people in a repressive setting hide what they really think and put on a show of loyalty they do not feel. So a natural question follows. If the opposition is real but hidden, where does it actually go? What does it look like day to day? James C. Scott is the person who went and found it. In his book Weapons of the Weak he lived among poor Malaysian peasants and watched how people who are far too weak to openly rebel still resist. They do it in small quiet ways that never become a protest or an uprising. They work slowly on purpose. They steal little things. They pretend not to understand an order. They quietly break or spoil tools and crops. They spread gossip that damages the powerful. They act incompetent so they can avoid doing what they are told. None of this is dramatic and none of it gets them arrested, but all of it is refusal. It is resistance by people who cannot afford to confront power to its face yet will not fully obey it either. In his later book Domination and the Arts of Resistance Scott turns this into one clean idea with two parts. The public transcript is the polite, obedient performance the powerless put on when the powerful are watching. The hidden transcript is what those same people actually say when they are safely among themselves with no boss or official present, where they mock, criticise and reject the people who rule them. The deference on the surface and the contempt underneath are both real, but only one of them is allowed to be spoken out loud. The last sentence is the payoff. When you look at an oppressed group from above and see them going along quietly, it is tempting to conclude they accept their situation, even that they have swallowed the rulers’ worldview, which is what the word hegemony implies. Scott says do not believe it. That calm surface is usually a mask people wear because they are being watched, a survival tactic, not proof of what they believe. Apparent consent is often just concealed refusal that has not yet found a safe moment to come into the open. A mask worn under surveillance and the apparent quiescence of the oppressed is frequently a tactical performance rather than genuine belief. This reframes dissent itself. Open revolt is the rare visible tip of a vast submerged mass of concealed refusal and a revolution is often the moment the hidden transcript is spoken aloud in public for the first time. Vaclav Havel’s essay The Power of the Powerless captured the same structure from inside a communist state through the parable of the greengrocer who places an official slogan in his shop window, not because he believes it but because refusing to would mark him. The whole edifice of the regime, Havel argued, rests on millions of such small daily acts of ritual compliance by people who no longer believe, so that the regime is far more brittle than it looks and the revolutionary act is simply living in truth, withdrawing the compliance the system needs to sustain its own appearance of legitimacy. Scott and Havel between them show that the line between quiescence and revolt is not a line between contentment and anger but between concealed refusal and public refusal, which returns the whole analysis to the withdrawal of cooperation. Power Rests on Obedience: The Dependency Theory of Rule The insight latent in Havel becomes a full theory of power in Gene Sharp’s The Politics of Nonviolent Action. Sharp’s foundational claim is that political power is not a possession the ruler owns but a relationship the ruled continually grant. No ruler can personally coerce a whole population. Every regime depends on the cooperation of soldiers who fire, clerks who file, workers who produce, police who arrest and citizens who obey and this cooperation flows through identifiable pillars of support, the military, the bureaucracy, the business class, the media, the religious authorities. Withdraw enough of it and the most heavily armed state becomes powerless, because the guns require hands to hold them. Revolution on this account is the strategic, coordinated subtraction of obedience and repression is not the opposite of this process but a step within it, because ordering soldiers to shoot compatriots is precisely the demand most likely to fracture a pillar of support. Erica Chenoweth and Maria Stephan gave this argument its empirical spine in Why Civil Resistance Works, the most consequential quantitative finding in the modern field. Analysing hundreds of campaigns from 1900 to 2006, they found that nonviolent resistance campaigns succeeded roughly twice as often as violent ones and that the mechanism was participation. Nonviolent movements draw far broader and more diverse participation, because the barriers to joining a march are far lower than the barriers to joining a guerrilla cell and mass participation is what shifts the loyalties of the security forces and the other pillars. From this came the widely cited observation that no campaign in their data failed after mobilising the active participation of 3.5 percent of the population. The related mechanism is backfire, or the repression paradox, in which visible violence against disciplined nonviolent protesters delegitimises the regime, alienates its own agents and swells the movement, the dynamic Sharp called political jiu-jitsu. The strategic core of the field’s applied wing follows directly. To assess a regime’s vulnerability, do not measure the population’s anger, map its pillars of support and ask which will hold when ordered to fire. The Emotional and Moral Engine: What Actually Moves People to Risk Structure and strategy explain opportunity, but something has to move a specific person to accept mortal risk on a specific morning and the field’s emotional turn addresses it without relapsing into Le Bon. James Jasper’s The Art of Moral Protest reintroduced emotion as rational and structured rather than irrational and contagious and his concept of the moral shock is central, the moment a specific event, a massacre, a stolen election, an image of a burned child, violates a person’s deepest moral intuitions so sharply that it overrides the ordinary calculus of the free rider problem and propels them toward action. Emotions here are not the enemy of reason but the currency in which grievance becomes motivation. Elisabeth Jean Wood’s study of insurgent collective action in El Salvador identified what she called the pleasure of agency, the intrinsic satisfaction of asserting dignity and authorship over one’s own life by defying the powerful, a benefit that accrues only to those who actually participate and therefore partly dissolves Olson’s free rider problem from within. Eric Hoffer’s The True Believer sits here as the enduring, if impressionistic, portrait of the psychology that mass movements recruit and reward, the frustrated self seeking to lose itself in a collective cause, the interchangeability of fanaticisms and the appeal of a holy cause to those who have made a mess of their own lives. Hoffer is a moralist rather than a social scientist and his generalisations outrun his evidence, but he named something real about the fusion of personal emptiness and collective zeal that more rigorous work has struggled to replace. Alongside him, Albert Hirschman’s body of work supplies the calmer economics of engagement. Exit, Voice, and Loyalty framed dissent, voice, as one of two responses to organisational decline, the alternative to simply leaving and showed that loyalty is what keeps voice alive when exit is available. Shifting Involvements explained the oscillation of whole societies between phases of private consumption and phases of public passion, so that the disappointments of private life periodically reload the reservoir of people available for collective action. Framing and Imagination: Turning Condition into Cause A grievance does not interpret itself and the final foundational element is the work of interpretation. David Snow and Robert Benford’s framing theory holds that movements succeed only when they construct collective action frames that perform three tasks, diagnosing a condition as an injustice rather than a misfortune, attributing it to identifiable culpable agents rather than fate or nature and proposing that collective action can remedy it. A drought is a tragedy, a famine caused by a hoarding elite is an outrage and only the second framing mobilises. This connects to what McAdam, in Political Process and the Development of Black Insurgency, called cognitive liberation, the shift in consciousness by which a population that has accepted its condition as legitimate and unchangeable comes to see it as unjust and alterable, the psychological precondition without which political opportunity goes unexploited. Benedict Anderson’s Imagined Communities supplies the largest frame of all, showing how print capitalism let strangers imagine themselves members of a single bounded community willing to die for one another, the nation, which became the dominant vessel into which modern revolutionary and anti-colonial energy has been poured. Frantz Fanon’s The Wretched of the Earth pushed the interpretive act to its most radical edge, arguing that for the colonised, participation in revolutionary violence was itself the process through which a people stripped of dignity reconstituted itself as an agent, a claim about the psychology of liberation that remains as contested as it is influential. The Unified Picture The single most important insight in the field is that a revolution is a coordinated withdrawal of the cooperation on which all rule depends and that cooperation is governed not by how much people suffer but by what each person believes others will do, under conditions where those beliefs are deliberately hidden. Every load-bearing concept in this essay is an expression or an implication of that one idea. Olson’s collective action problem states the difficulty of coordinating withdrawal at all. Kuran’s preference falsification, with Granovetter’s thresholds, explains why the beliefs that would permit coordination stay hidden and how they cascade into the open, which is why Kurzman is right that the timing is unpredictable. Skocpol, Goldstone, Moore and Tocqueville identify the structural conditions, state breakdown, elite fracture, fiscal and demographic crisis, reform under weakness, under which the cost of the first defector falls far enough for a cascade to start. Sharp and Chenoweth name what is actually being withdrawn, the obedience flowing through the pillars of support and show that broad participation, not violence, is what pulls those pillars down. Scott and Havel locate the concealed opposition that the cascade brings into the open, so that revolt is the public speaking of the hidden transcript. Jasper, Wood and the framing theorists explain what moves the individual first mover across the threshold, the moral shock, the pleasure of agency and the frame that turns a condition into a remediable injustice. The 20 percent that produces 80 percent of the field’s power is a short chain. Grievance is close to constant, so it cannot be the variable. What varies is coordinated action, which is blocked by the free rider problem and by hidden preferences and unblocked by a shift in the perceived condition of the state and in beliefs about what others will do. Rule depends on obedience, revolt is the withdrawal of it and the withdrawal cascades because each defection lowers the risk to the next defector. Hold those five moves together, the constancy of grievance, the primacy of coordination, the concealment of true preferences, the dependency of power on obedience and the cascade dynamic of withdrawal and you can explain the field’s central facts, above all the twin observations that revolt does not track suffering and that its timing cannot be forecast even in principle. The Limits The mode of thinking has its own failure modes and refusing to name them would repeat the deprivation theorists’ error of mistaking an elegant model for the world. The structural school buys its explanatory power by discarding prediction. Skocpol can tell you which conditions accompany revolution after the fact but cannot tell you when a fractured state will actually fall and her comparative method, drawing law-like conclusions from a handful of great cases selected because they were revolutions, is exposed to selection on the dependent variable, the same survivorship bias that afflicts most study of successful revolts and ignores the far larger number of state crises that produced no revolution at all. Preference falsification and the cascade model have the opposite defect. They are almost too powerful, explaining any outcome after the event while forbidding prediction before it, which makes the theory difficult to falsify and tempts its users into unfalsifiable just-so stories about hidden thresholds no one can measure. Kurzman is admirably candid that his anti-explanation explains why we cannot explain, which is honest but thin comfort to anyone who wants to know what happens next. The Chenoweth and Stephan finding, the empirical crown of the applied wing, has been pressed hard since publication. The 3.5 percent figure is a description of a historical dataset rather than a causal law, campaigns that reach such participation are probably already winning for other reasons, so mobilisation and success may be jointly caused rather than one driving the other and the post-2010 record of nonviolent movements that stalled or were crushed, from the Arab uprisings to Hong Kong to Belarus, warns against reading the twentieth-century average as a guarantee. More deeply, the whole field carries a state-centric and often Western bias, treating the capture or collapse of a central state as the definitional event and undertheorising the content of revolutionary ideology, what people actually want, as opposed to the mechanics of how they mobilise. It is far better at explaining the withdrawal of cooperation than at predicting what the victors will build and its recurring blind spot is the aftermath, the frequency with which the cascade that topples a regime installs something as bad or worse, a pattern Crane Brinton’s older The Anatomy of Revolution described as the movement from moderate rule through radical terror to a Thermidorian reaction and which the mechanics of mobilisation say almost nothing about. A framework that predicts the fall but is silent on the morning after is powerful and incomplete at once. What This Demands of You Internalising this field requires you to break a reflex. When you see unrest, or its absence, stop asking how angry people are, because that question has almost no predictive content and start asking a different set. Ask what the state’s coercive apparatus believes about its own future, because the loyalty of the people who hold the guns is the hinge on which everything turns. Ask what each person thinks everyone else believes and treat public unanimity in a repressive setting as evidence of nothing, because preference falsification means the surface tells you the cost of speaking, not the distribution of thought beneath it. Ask where the hidden transcript is being spoken and what it says, rather than reading ritual compliance as consent. When a regime begins to reform, raise rather than lower your estimate of its danger, because Tocqueville’s lesson is that easing is the perilous moment. Map the pillars of support and name the specific institution whose defection would collapse the rest, because that, not the size of the crowd, is where a movement wins or dies. You must also police your own hindsight. When a revolution has happened and feels as though it had to, distrust that feeling completely, because the sense of inevitability is the cascade dynamic disguising the accidents that could have stopped it and the same illusion will make you overconfident about the next case. Hold the free rider problem in front of every claim that a shared grievance will produce action and demand to know what solves it here, the network, the incentive, the moral shock, the emotion, the frame, or nothing. Separate the question of why a regime falls from the question of what replaces it and never let a compelling account of the first substitute for an answer to the second. The habit this field installs, if you let it, is to look past the noise of anger toward the silent architecture of belief, obedience and coordination that decides whether anger ever becomes anything at all. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

3. 

   19 Jul 2026

   Chokepoints: How Chips, Code and Coalition Decide Who Wins the AI Century

   In July 2026 a private company in San Francisco decided who in the United States government would get early access to a model capable of finding and exploiting software vulnerabilities faster than any human team. Not the Pentagon. Not Congress. Anthropic. Within weeks the same administration that had promised the most permissive AI regulatory regime in the world found itself running something close to a licensing system it had never voted for and did not control. Six thousand miles away, Beijing watched the same capability arrive at its own labs and prepared to route it through an agency that had been quietly building the machinery to handle exactly this moment since the year after ChatGPT launched. Same technology. Two governments. Two entirely different responses, shaped less by strategy than by which coalitions each government happened to have already built. That gap is the subject of this essay. It sits underneath a supply chain for advanced chips concentrated in a handful of firms that could, if severed, freeze a rival’s military and AI development overnight. It sits underneath a decade of cyber weapons that blurred the line between war and peace so thoroughly that a piece of malware aimed at Ukraine ended up costing global companies billions. It sits underneath a race between the United States and China that most people still describe as a contest over who builds the smartest model, when the real contest, the one with the shorter fuse, is over who controls the chokepoints - the chips, the export lists, the access rosters, the agency with the standing authority to say yes or no before a model that can hack anything reaches the open internet. The books referred in this piece here follow were written to explain how power actually moves through complex systems long before anyone had heard of Claude Mythos. What happened this summer is simply the first time you can watch every one of their arguments play out in real time, in public, with both superpowers running the experiment at once. The World Runs on Silicon and Whoever Controls It Controls Everything Else Begin with the physical layer because everything else is downstream of it. Chris Miller’s history of the semiconductor industry is the most important strategic document written in the last decade dressed up as business history. The central argument is simple and devastating. Every meaningful capability in the modern world, military precision, artificial intelligence, financial infrastructure, communications, surveillance, logistics and medicine, runs on chips. And the supply chain that produces those chips is among the most geographically concentrated and technically fragile systems ever built. The design of leading-edge chips is dominated by a handful of American firms and a Taiwanese company called TSMC. The manufacturing equipment required to make them is controlled largely by a Dutch company called ASML, which makes machines that no other organisation on earth knows how to replicate. The rare earth materials flow primarily through China. The assembly happens across East and Southeast Asia. This is not a resilient supply chain. It is a single point of failure wearing the costume of a global network. The policy implication is that semiconductor access is now a primary instrument of national power. When the United States restricted ASML from selling its most advanced machines to China, it was not an export control. It was a strategic weapon aimed at freezing Chinese military and AI development at a specific technological moment. The response from China has been a multi-hundred-billion dollar programme to build domestic chip capability from scratch. Whether they succeed or not will shape the balance of power for the next thirty years more decisively than most military acquisitions. The lesson here is that technology policy and defence policy are no longer separate domains. The country that controls semiconductor manufacturing controls the ceiling of every other country’s military and economic capability. Decisions about domestic chip investment, export controls, alliance structures around technology and the terms of trade with China are all, at their core, national security decisions even when they are framed as economic ones. Cyber Weapons Changed the Rules of War and Most Governments Have Not Caught Up The second foundational layer is understanding what happened to warfare when code became capable of physical destruction. Stuxnet was the moment everything changed. A piece of software, almost certainly built jointly by the United States and Israel, physically destroyed Iranian centrifuges by making them spin themselves apart while reporting normal operations to their operators. No bombs. No soldiers. No declaration of war. A nation’s nuclear programme damaged by a USB drive and a few thousand lines of code. Zetter’s account of this is the essential text because she traces not just the what but the why and the what next. Once Stuxnet was discovered and reverse engineered, the blueprint was public. Every sophisticated state actor and eventually many non-state actors now understood that physical infrastructure could be attacked through software. What followed is documented by Greenberg in Sandworm. Russia’s military intelligence unit systematically used cyber operations against Ukraine as a testing ground with steadily escalating ambition. They took down the Ukrainian power grid twice. They interfered with the election infrastructure. And then they deployed NotPetya, ostensibly a ransomware attack against Ukraine, which escaped its intended boundaries and tore through the global economy. Maersk lost almost its entire global shipping infrastructure. Merck lost production capability. FedEx sustained massive damage. Total global losses approached $10 billion from a weapon that was not even aimed at most of its victims. This is the critical concept that most policy frameworks have not yet fully absorbed - in cyber warfare, collateral damage is not an accident or an acceptable side effect. It is a structural feature of the domain. Code does not respect borders. A weapon deployed against one target propagates through every connected system that shares its vulnerability. The interconnectedness that makes the global economy function is simultaneously the attack surface that makes it catastrophically vulnerable. Perlroth adds the market layer that makes this worse. Governments discovered that they could purchase unknown software vulnerabilities, zero-days, from a grey market of researchers and brokers and use them as weapons before the software manufacturers could patch them. The NSA stockpiled hundreds of these vulnerabilities. When a Shadow Brokers breach released many of them publicly, those same government weapons were immediately weaponised by criminal and state actors against the entire world. The policy failure here is profound. A government agency sat on vulnerabilities in software used by hospitals, power plants and banks because those vulnerabilities were more valuable as weapons than the public was protected by their disclosure. EternalBlue, an NSA-developed exploit, became the engine of both WannaCry and NotPetya. American taxpayers funded the development of a weapon that then attacked American hospitals. The combined lesson of these three books is this. Cyber is not a domain that exists alongside conventional military, economic and diplomatic power. It is the connective tissue running through all of them. Offensive cyber capability can achieve strategic military objectives without triggering the international response that a kinetic attack would. It can destroy economic value at scale with plausible deniability. It can undermine democratic processes without a single soldier crossing a border. And the defences against it require a fundamentally different architecture than the defences against conventional threats because the attack surface is every device, network and human operator connected to the system you are trying to protect. Clarke and Knake, writing from direct experience building US cyber policy, crystallise this into a framework that should inform every national cyber strategy. The goal of cyber security policy cannot be to prevent all attacks. That is not achievable in a connected system. The goal is resilience. The ability to absorb attacks, degrade their impact, recover rapidly and deny adversaries the clean strategic victories they are seeking. This reframes the entire policy question. Instead of asking how do we stop attacks, you ask how do we ensure attacks cannot achieve strategic objectives. That requires investment in redundancy, segmentation, rapid response capability and an honest assessment of which systems are so critical that they must be air-gapped from the broader network regardless of the operational inconvenience. The public-private tension Clarke and Knake identify is perhaps the most practically difficult policy problem in this space. Critical infrastructure in most democracies is privately owned. Power grids, water systems, financial networks and telecommunications are run by companies whose primary obligation is to shareholders. The security investments required to protect these systems against nation-state level cyber threats are expensive, reduce short-term profitability and protect externalities that the individual company does not fully capture. Market incentives alone will not produce the level of investment that national security requires. Some form of regulatory mandate, liability framework or direct government investment is necessary. Designing that intervention without creating compliance-theatre, regulatory capture or innovation stagnation is the central challenge. The AI Race Is Not What Most People Think It Is The AI competition between the United States and China is consistently misunderstood in public debate because it is framed primarily as a race to build the most powerful models. Kai-Fu Lee, who has operated at the highest levels of AI development on both sides of this competition, offers a more accurate and more disturbing picture. The United States leads on fundamental AI research. It has the best universities, the most capable AI researchers and the companies producing the most sophisticated models. China has a different set of advantages that may matter more in the medium term. It has vastly more data, partly because its population is larger and partly because its regulatory and cultural environment places far fewer constraints on data collection. It has a government that treats AI development as a national strategic priority and coordinates investment across state and private actors in ways that democratic systems struggle to replicate. And it has an enormous pool of capable implementation engineers who can take research breakthroughs and deploy them into real-world applications at a speed and scale that most Western institutions cannot match. Lee’s argument is that the first wave of AI value was created through research breakthroughs. That era produced the underlying models and capabilities. The second and larger wave of value will come from applying those capabilities to specific domains - finance, medicine, logistics, manufacturing, education and defence. In that application race, data and deployment speed matter more than raw research capability. China’s structural advantages become more pronounced not less as the competition shifts into this phase. The policy implication is that export controls on AI chips and model weights, while not irrelevant, are not sufficient as a strategic response. The more important questions are about the structure of talent pipelines, the terms on which AI research produced in US universities is accessible to foreign actors, the regulatory environment that determines how quickly AI can be deployed in domestic industries and the alliance structures that determine whether allied nations are building AI capabilities that interoperate with or compete against American systems. The Mythos Moment: When the Framework Stopped Being Hypothetical Everything above was written to be evergreen, a set of durable lessons drawn from decades of chip fights and cyber wars. It did not stay hypothetical for long. In the summer of 2026, Anthropic’s Claude Mythos and Claude Fable models crossed a capability threshold in offensive cyber operations that the frameworks above predicted but that no government had actually built the machinery to handle. This is worth treating as a live case study, because it is the rare moment where you can watch Miller’s chokepoint logic, Zetter and Greenberg’s weaponised-code logic, Lee’s application-race logic and De Mesquita and Smith’s coalition logic all fire inside the same news cycle. The American response confirmed the institutional-reality thesis before it confirmed anything about the technology itself. Washington had spent years promising a permissive, let-it-rip regulatory posture toward frontier AI. Within days of Mythos’s release the administration had, in the words of one close observer, backed into what amounted to a de facto involuntary licensing regime, precisely the outcome it had pledged to avoid. Access to Fable and Mythos was suspended in mid-June under export-control authority and only restored at the start of July once the underlying controls were lifted. The revealing detail is not the suspension itself. It is who set the agenda. Project Glasswing, the closed pre-release access programme that decided which government bodies and critical-infrastructure operators got early access to the models, was designed and staffed by Anthropic, a private company. No federal agency chose the guest list. The scramble that followed, Commerce, the newly stood-up Center for AI Standards and Innovation, the NSA and the Treasury all reportedly jockeying to establish jurisdiction, is exactly the pattern De Mesquita and Smith describe - institutions defaulted to protecting the interests of whichever coalition was already in the room, because no one had pre-negotiated who was supposed to be in the room at all. China’s channel for the same problem already exists, which is the point analysts most familiar with Chinese AI governance keep making. Since the year after ChatGPT’s release, Chinese model developers have submitted new releases to the Cyberspace Administration of China for a battery of pre-deployment tests, historically focused on content and political-sensitivity filtering rather than cyber risk. That apparatus is now being extended. Regulators have already required at least one major lab to add safeguards against malicious code generation as part of a routine rectification campaign and the expectation among close observers is that cyber-capability tests will simply be bolted onto the existing filing process rather than requiring new legislation. A Chinese frontier lab is very unlikely to skip that heads-up and self-release a Mythos-class model the way a Western company might, because the CAC maintains standing, informal touchpoints with the major labs that make quiet non-disclosure practically unthinkable. The forecasts on timing vary but not by much. A co-founder of Zhipu AI has publicly predicted a Chinese model on par with Mythos before the end of 2026. The independent research group IAPS puts the date closer to February 2027. Either way, the question facing Chinese regulators is not whether they will face this decision but how many months they have to prepare for it and unlike their American counterparts they are not starting from a jurisdictional void. What a Chinese Glasswing would look like follows directly from that asymmetry. Because the American programme was initiated by a company and only reluctantly backstopped by government, access decisions were negotiated ad hoc. Because any Chinese equivalent would almost certainly be initiated by the CAC or a comparable security organ, the access list would be pre-approved before a single account was issued, likely running through central ministries and state-owned enterprises first, then critical but poorly hardened infrastructure operators, then wider commercial release over a period of months. The same underlying technology, filtered through two different institutional starting points, is expected to produce two structurally different rollouts. That is not a story about which country regulates AI better. It is Meadows’s point about leverage: the highest-leverage intervention is not the specific rule but the paradigm and channel through which rules get made and the United States and China are working from different paradigms built years before Mythos existed. The fault line most likely to define the next year runs through open source rather than through the cyber tests themselves. Chinese labs have built genuine global standing on open-weight releases and the leadership of both MiniMax and Zhipu made high-profile, public recommitments to open source within days of each other in mid-2026, one pledging a share of company market value to an open-source foundation, the other circulating an internal memo framing open release as central to the company’s mission. Set against that is reporting that China’s Ministry of Commerce and the National Development and Reform Commission have discussed restricting overseas access to the country’s most advanced models, closed and open alike and floated criminalising the leak of proprietary AI under national security law. That combination, an economic ministry rather than the CAC driving the conversation, tracks the precedent Beijing set when it placed recommendation algorithms on its controlled-export list ahead of the forced TikTok divestiture: not a blanket ban, but a reserved veto over future deals and diffusion. This is Miller’s chokepoint logic migrating one layer up the stack, from the chips that train frontier models to the weights those chips produce. Whether Beijing exercises that veto will depend less on the cyber-risk mathematics, since well-resourced state and criminal actors do not obviously need open weights to do serious damage, than on how much diplomatic and commercial value the government still assigns to its self-positioning as the world’s more open AI power. The financial mechanics converged from a different direction. Around the same moment, OpenAI reportedly offered the US government a five percent equity stake, widely read as an attempt to secure political protection rather than as an ordinary investment. Whatever the precise motive, the effect is to pull the United States toward the same instinct that led China to hold equity in strategic firms for decades, without any of the institutional apparatus China has built to manage state ownership of enterprise. Call it state capitalism with American characteristics: two governments arriving at deeper financial entanglement with their frontier AI companies from opposite starting points, one moving from state-owned enterprise toward more market discipline, the other moving from arm’s-length regulation toward equity and control. Acemoglu and Johnson’s argument, that outcomes from a new technology are set by whoever ends up holding the power to direct it rather than by the technology itself, has rarely had a cleaner natural experiment running in real time on both sides of the Pacific at once. The Political Economy of Technology: Who Technology Actually Serves Zuboff and Acemoglu and Johnson ask the question that techno-optimists consistently avoid. When a new technology is developed and deployed at scale, who captures the value it creates and who bears the costs? The answer across the history of technology, documented by Acemoglu and Johnson with rigorous economic analysis, is that outcomes depend almost entirely on the power structures that govern the technology’s development and deployment. The industrial revolution was not automatically beneficial to workers. For decades it produced immiseration, longer hours, worse health and greater precarity for the majority of people whose labour it used. The value flowed to capital owners. It was only through sustained political organisation, regulatory intervention and institutional change that productivity gains were eventually distributed more broadly. The same pattern repeated with each subsequent wave of technological change. The current wave of AI and automation is reproducing this pattern at higher speed and the equity-stake manoeuvring described above is an early sign of where the value is being contested before the productivity gains have even fully materialised. The productivity gains are real. The distributional mechanisms to share them are weak or absent. And the companies capturing the value have achieved sufficient political influence that the regulatory interventions that would redistribute it face structural resistance. For a national policy maker, this means that the question of what technology to develop is inseparable from the question of what institutional structures will govern its deployment and who will benefit. Technology policy that ignores political economy is not neutral. It is a choice in favour of whoever currently holds economic power. Zuboff’s contribution is to name and theorise a genuinely new economic logic that most regulatory frameworks were not built to address. The major digital platforms do not sell products or services to their users in the conventional sense. They extract behavioural data from users, process it into predictive models of human behaviour and sell the ability to influence that behaviour to advertisers and other buyers. The user is not the customer. The user is the raw material. The product is the modification of behaviour at scale. This is not a metaphor or a polemic. It is a description of the business model as designed and executed. The implications for national policy are serious and underappreciated. If the dominant communications, information and social infrastructure of a democracy is built on a business model whose core function is the manipulation of human behaviour, then the information environment in which democratic deliberation occurs is systematically compromised. Not by foreign interference, though that is a real additional problem, but by the structural incentives of the domestic platforms themselves. Engagement maximisation produces outrage because outrage is more engaging than nuanced information. Outrage produces polarisation. Polarisation produces political environments that are easier to manipulate and harder to govern. This is not a side effect. It is a predictable consequence of the incentive structure. Designing policy that addresses this without destroying the genuine value these platforms create or sliding into state control of information is one of the genuinely hard problems of democratic governance in this period. Systems Fail in Ways That Cannot Be Prevented by Being More Careful Perrow’s contribution to this body of knowledge is the most sobering and the most important for anyone responsible for the resilience of critical infrastructure. His central argument, developed from studying complex industrial accidents, is that in systems characterised by tight coupling and interactive complexity, catastrophic accidents are normal. Not in the sense of being frequent but in the sense of being inevitable given sufficient time. Tight coupling means that when something goes wrong, the components of the system interact so quickly and automatically that there is no opportunity for human intervention before failures cascade. Interactive complexity means that the number of possible failure modes and interactions between components is too large for any operator to fully model or anticipate. Nuclear power plants, air traffic control systems, financial markets, chemical processing facilities and interconnected cyber infrastructure all exhibit both properties. The implications for policy are uncomfortable. You cannot make these systems safe by hiring better operators, writing better procedures or improving individual components. The safety properties of the system are emergent from its architecture. A system that is tightly coupled and interactively complex will eventually produce a catastrophic failure regardless of the competence of the people operating it. The policy response is not to accept catastrophe as inevitable and do nothing. It is to understand that the design of systems must prioritise the ability to fail safely, meaning failures must be contained, reversible and recoverable. This requires accepting operational inefficiency in exchange for resilience margins. It requires building in redundancy that will feel wasteful in normal times. It requires creating genuine authority and resources for operators to shut things down when something feels wrong even before they can articulate exactly what is wrong. And it requires regulating against the commercial pressure that consistently removes safety margins in pursuit of efficiency, because in normal operation the margins look like waste and the catastrophic failure they prevent is hypothetical. Applied to cyber and technology infrastructure, Perrow’s framework suggests that the interconnectedness which makes digital systems powerful is simultaneously what makes them capable of catastrophic cascading failure. NotPetya is the cyber equivalent of a Perrow normal accident. It was not the result of a single point of failure or human error. It was the result of a complex, tightly coupled global system interacting with a malicious input in ways that no single actor could anticipate or contain. The Mythos-class debate over whether a model that can find cyber vulnerabilities is dangerous mainly because it is open-weight or dangerous regardless of release format is, underneath the disagreement, an argument about exactly this - whether the vulnerability lives in the model or in the tightly coupled, poorly segmented systems the model is being pointed at. The policy response is architectural - segment critical systems, enforce genuine air gaps where the stakes justify the cost, build recovery capability not just prevention capability and be deeply sceptical of efficiency arguments that remove the redundancy and friction that resilience requires. Grand Strategy: The Art of Not Being Stupid at Scale Gaddis teaches something that technical and policy experts consistently undervalue. The most common failure mode of intelligent, well-intentioned people in positions of strategic responsibility is the failure of proportion. Gaddis's argument is that smart, capable people in senior roles rarely fail because they picked the wrong goal. They fail because the size, shape or timing of their response does not match the situation. They might be right about what needs to happen and still ruin it by moving too fast, too slow, with too much force, with too little, or by fighting today's problem with a tool built for a different one. Proportion, not correctness, is the scarce skill. They pursue the right objective with means that are mismatched to the environment, the constraints and the time horizon. They optimise for what they can measure rather than what matters. Once someone is accountable for results, they gravitate toward whatever can be counted and reported, because it is defensible. The problem is that the thing most worth protecting in a complex situation is often exactly the thing that resists measurement. Trust, timing, second-order political cost, whether a coalition holds together. Chase the metric and you can hit every target on paper while missing the actual objective. They mistake confidence for clarity. Confidence is a feeling. Clarity is an accurate read of the situation. The two get confused because confidence feels like understanding from the inside, but a confident person can be confidently wrong. Gaddis is warning that certainty is not evidence of correctness. It can just as easily be evidence that someone has stopped questioning their own model of the world. They confuse the map for the territory. This is the classic systems-thinking warning, and Gaddis applies it to strategy specifically. A plan, a briefing, an intelligence assessment: these are all simplified representations of reality, built with necessary omissions. The failure is treating the simplified version as if it were the whole truth, then acting on the map's gaps as if they were the territory's gaps. His study of the great strategists reveals a consistent pattern in the effective ones. They hold their objectives clearly and firmly while remaining radically flexible about the means of achieving them. They understand the limits of their own power and calibrate their ambitions accordingly, not out of timidity but out of the recognition that overreach is the most common cause of strategic failure. They think in multiple time horizons simultaneously, asking not just what achieves the objective today but what position does this action create for the next move and the move after that. And they develop the capacity to tolerate ambiguity, to act under uncertainty without either paralysing themselves with the need for more information or charging forward in denial of the uncertainty. For a cyber and technology policy role specifically, this translates into several concrete disciplines. First, resist the pressure to produce policies that look decisive rather than policies that are effective. In complex technical domains, the politically satisfying response and the strategically correct response frequently diverge and a rushed licensing regime built in the two weeks after a model release is the textbook example of the former crowding out the latter. Second, take seriously the second and third order consequences of any intervention. A policy that solves a near-term problem by creating a more serious medium-term problem is not a policy. It is a debt. Third, understand the difference between problems that yield to direct action and problems that require patient positional work to create conditions in which the problem can eventually be addressed. Most significant technology policy challenges are in the second category. Systems Thinking: The Tool That Separates Policy from Wishful Thinking Meadows is the practical implementation of Gaddis’s strategic wisdom. Where Gaddis teaches you how to think about ends and means at the level of grand strategy, Meadows gives you the specific cognitive tools to trace how interventions actually move through complex systems. The central insight is that most policy failures are not failures of intent or intelligence. They are failures of systems understanding. Well-intentioned interventions produce unexpected and often counterproductive outcomes because the intervener understood the immediate causal chain but not the feedback loops, time delays and emergent behaviours that determine what actually happens at the system level. In technology policy, this is everywhere. Regulating a platform to reduce harmful content creates incentive to move harmful content to less visible but more extreme venues. Mandating data localisation to protect privacy creates infrastructure that authoritarian governments can use to surveil their populations more effectively. Restricting the export of AI technology to adversaries accelerates their domestic development of the same technology without the dependencies that provided influence. Subsidising domestic chip manufacturing without addressing the talent pipeline produces expensive facilities that cannot operate at full capacity. Meadows identifies the highest-leverage intervention points in complex systems. The goals of the system, the rules that govern it and the paradigms that determine what goals and rules are even considered possible. Most policy operates at the lowest-leverage levels - adjusting flows, tweaking parameters, adding regulations. The interventions that actually transform system behaviour operate at the level of the fundamental goals and rules. This is why the most consequential technology policy moments are not the passage of individual regulations but the shifts in the underlying framework. The decision that the internet should be treated as a common carrier, the decision that software patents are valid, the decision about whether AI systems are products subject to product liability and now the decision about whether a single private company gets to design the first international access framework for a dangerous capability before any government does. These framework decisions determine the space within which all subsequent lower-level policy operates. The Institutional Reality: Why Good Ideas Lose De Mesquita and Smith provide the framework that everything else requires to actually function. You can have the correct analysis, the right policy proposal, the technical evidence and the strategic rationale. If you do not understand how decisions actually get made inside institutions, you will watch correct ideas lose to worse ones repeatedly and develop an increasingly mystified frustration that does not help anyone. The core argument is that leaders in any institution, democratic governments, military organisations, international bodies, corporations and regulatory agencies, make decisions that serve the coalition that keeps them in power. This is not corruption in the cynical sense. It is a structural feature of how power works. A leader who consistently ignores the interests of the coalition they depend on does not remain a leader. Therefore the revealed preferences of any institution are not determined by its stated mission but by the interests of the key coalition members whose support is essential. Applied to technology and cyber policy, this explains several persistent patterns. Why critical infrastructure cyber security remains systematically underfunded despite universal acknowledgement that it is a serious national security risk - the companies that own the infrastructure have political influence, the investment required reduces their profitability and the catastrophic attack that justifies the investment has not yet occurred in their specific sector. Why data privacy regulation consistently produces compliance frameworks that large incumbents can satisfy but small competitors cannot - large incumbents participate in drafting the regulation and the compliance cost is manageable for them while it is prohibitive for new entrants who might disrupt them. Why international cyber norms agreements progress slowly despite genuine shared interest in reducing the risk of catastrophic conflict - the military and intelligence agencies of major powers do not want to constrain the offensive capabilities they have spent decades and billions developing. It also explains, with uncomfortable precision, why Washington’s mid-2026 scramble and Beijing’s anticipated response look so different despite facing the identical underlying capability. The American coalition around AI policy was built on a promise of light-touch regulation to a coalition of investors, founders and a White House that had staked political capital on winning the AI race through speed. That coalition had no ready answer when a company-run access programme forced the jurisdiction question into the open, so the fight became about which existing agency could claim the new territory rather than about the underlying risk. The Chinese coalition was built years earlier around a different bargain: continued commercial latitude for AI labs in exchange for compliance with a content and political-risk testing regime run through a single agency the labs already had a relationship with. Extending that regime to cyber risk requires no new coalition, only an addition to an existing one. Neither arrangement is more virtuous than the other. Both are simply the coalition each government could build with the tools and trust it had already accumulated, exactly as De Mesquita and Smith would predict. Grand Unified Picture Pull all of this together and what emerges is a coherent and demanding vision of the strategic environment. The physical foundation of all modern power is semiconductor technology whose supply chain is concentrated, fragile and currently the primary arena of US-China strategic competition. Whatever policies you design in the cyber and technology space will be shaped by and will shape this underlying hardware reality. On top of that physical layer sits a cyber domain in which the distinction between peace and war has been permanently blurred. States conduct continuous operations against each other’s infrastructure, institutions and information environments below the threshold that would trigger conventional military response. The weapons used in these operations are built from vulnerabilities in the shared software and hardware infrastructure that everyone depends on. There is no clean line between civilian and military targets because civilian infrastructure is military-grade critical. Deterrence logic from the nuclear era applies imperfectly and incompletely. Resilience is more achievable than deterrence and more important to invest in. The AI layer is developing faster than governance frameworks can track and as of mid-2026 that is no longer a warning about the future but a description of the present. The countries and institutions that develop the ability to deploy AI systems reliably, evaluate them honestly and govern them coherently will have structural advantages that compound over time. The alignment problem is not solved. Deployed systems fail in ways that are predictable in type and unpredictable in instance. Any procurement or deployment of AI for consequential purposes requires a level of scepticism and evaluation rigour that most institutions are not yet applying and the fact that the world’s most capable AI company rather than any government wrote the first access rules for a frontier model with serious offensive cyber capability is the clearest available evidence that most institutions are, in fact, behind. The political economy of technology determines whose interests these systems serve. The platforms that mediate information in democratic societies are built on business models that systematically compromise the information environment those democracies depend on. The productivity gains from AI and automation are currently flowing primarily to capital while the costs are distributed across labour. Equity stakes, sovereign shares and state-directed access programmes are the opening moves in a much longer argument over who owns the value this technology creates. These are not natural outcomes of technological progress. They are the product of specific choices about power and governance that can be made differently. Complex systems built on tight coupling and interactive complexity will fail catastrophically regardless of the quality of the people operating them. Resilience architecture, the ability to fail safely and recover rapidly, is more achievable and more important than the illusion of prevention. Strategy requires holding clear objectives with flexible means, calibrating ambition to real constraints, thinking across multiple time horizons and developing tolerance for the uncertainty that is the permanent condition of anyone trying to act in a complex environment. Systems thinking requires tracing the feedback loops and emergent behaviours that determine what interventions actually do rather than what they are intended to do. And none of it matters if you cannot navigate the institutional reality of how decisions actually get made, which requires understanding the coalition dynamics, the incentive structures and the distribution of power that determine which ideas become policy and which remain correct but irrelevant. Watch Washington and Beijing over the next six to twelve months and you are watching this entire framework tested in real time, by two governments with different coalitions, different starting infrastructures and, for the first time, the same underlying technological threat. The Demand This Makes of You The synthesis of all these books and of the live case now running alongside them, is not a set of conclusions. It is a set of demands on your cognitive and political capacity. You must hold technical depth and strategic breadth simultaneously because the domain requires both and most of the people you will work with have only one. You must understand the difference between how institutions present their decision-making and how it actually operates. You must develop the ability to trace the second and third order consequences of interventions through complex adaptive systems without either paralysing yourself or pretending the complexity does not exist. You must be sceptical of both techno-optimism and techno-panic as they are both substitutes for analysis rather than forms of it. You must understand that the adversaries you are designing against are not static targets but adaptive systems that will change in response to your interventions. You must also understand that the most important decisions in this domain over the next decade will not be made by technical experts alone. They will be made through political processes in which interests, coalitions, narratives and institutional inertia will be as determinative as evidence and analysis. Your ability to operate effectively in those processes, to build the coalitions that can move the key decisions in the right direction, will determine whether your technical and strategic knowledge actually changes anything. The current window matters more than most, because commentators close to both governments now argue that today’s AI labs hold more direct influence over policy outcomes than they are likely to retain once the public and elected legislators start treating AI as the kind of issue people actually vote on. What gets locked into precedent in the next year or two, who sets access rules, who holds equity, which agency owns the file, tends to persist long after the original justification for it has faded. The books from the previous list, Greene and Cialdini in particular, are not separate from this domain. They are the operating layer underneath it. The technical knowledge above will make you credible. The understanding of how power, influence, perception and coalition dynamics actually work will determine whether that credibility translates into impact. This is not a counsel of cynicism. It is the minimum realism required to actually change anything in a senior policy role rather than simply being very well-informed about why things went wrong. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

4. 

   17 Jul 2026

   On Ridley's Work

   Matt Ridley is a British science writer, journalist and hereditary peer (Viscount Ridley) whose work spans evolutionary biology, genetics, economics, and the history of innovation. Since the early 1990s he has written eight major books and thousands of articles, all united by a single animating conviction: that complex order arises from the bottom up, not the top down, and that this process – whether in genomes, markets, ecosystems or cultures – is the most reliable engine of progress humanity has ever found. He is not a practicing scientist, but he reads the primary literature with unusual rigour and translates it without dumbing it down. The result is a body of work that is simultaneously intellectually serious and accessible to general readers. Ridley trained as a zoologist at Oxford, wrote a doctoral thesis on the mating system of pheasants, then spent years as science editor and American editor of The Economist before turning to books full-time. That background – evolutionary biology crossed with free-market economics crossed with science journalism – explains almost everything about his intellectual DNA. Before surveying each book, it is worth identifying the small cluster of core ideas that recur, mutate and deepen across his entire output. These are not conclusions so much as lenses. Master them and every Ridley book becomes more legible. Spontaneous Order and Bottom-Up Emergence The single most consistent thread in Ridley’s work is a hostility to top-down, intentional design as an explanation for complex beneficial outcomes. He draws heavily on Friedrich Hayek’s insight that market prices encode distributed knowledge no central planner can aggregate and generalises it far beyond economics. Sexual reproduction, genetic regulation, language, common law, the internet, the city, scientific knowledge itself – all are examples of systems that improve through decentralised trial-and-error rather than directed control. For Ridley, the instinct to find an author behind every good outcome is a cognitive error with real costs. It leads us to trust planners, regulators and visionaries over processes that are messier but more adaptive. This idea appears embryonically in The Origins of Virtue (where cooperative norms emerge without a sovereign), matures in The Rational Optimist (where trade and specialisation self-organise into civilisation), and becomes the explicit thesis of The Evolution of Everything (where every chapter argues that a given domain – money, morality, technology, government – evolved rather than was designed). Evolution as Universal Metaphor Ridley is a Darwinian not just about biology but about almost everything. The core Darwinian algorithm – variation, selection, heredity – operates wherever there is replication with error and differential fitness. Ideas mutate and compete. Technologies undergo selection pressure from markets and users. Institutions evolve by retaining practices that work and discarding those that do not. This is not metaphor for Ridley; it is mechanism. He takes seriously the claim that cultural evolution follows rules analogous to, but distinct from, genetic evolution. He is influenced by Richard Dawkins’s meme concept and by the broader tradition of evolutionary epistemology running from Karl Popper through David Deutsch. The payoff is significant. If beneficial change is evolutionary, then it is incremental, combinatorial, unpredictable in detail but directional in aggregate. Blueprints and five-year plans fail not because planners are stupid but because evolution cannot be shortcut. Exchange, Specialisation and the Collective Brain The central argument of The Rational Optimist – arguably Ridley’s most influential book – is that human prosperity is a product of exchange. When people trade, they specialise and when they specialise they get better at what they do. Each trader effectively has access to the skills and knowledge of everyone they trade with. Ridley borrows the anthropologist Henrich’s concept of the “collective brain”. No individual knows how to make a computer mouse or a pencil from scratch, but the market coordinates millions of specialists whose combined knowledge makes such objects trivially cheap. The richer and more connected a population, the larger its collective brain and the faster its technology improves. Conversely, isolated populations – Easter Island, Tasmania after rising seas severed it from Australia – lose technology they once had. The implication is that trade, globalisation and economic integration are not optional extras but are the mechanism of cultural and material progress. Nature Via Nurture, Not Nature Versus Nurture Ridley’s genetics books (Genome and Nature Via Nurture) systematically demolish the false dichotomy between genetic determinism and blank-slate environmentalism. His key insight – stated most sharply in Nature Via Nurture – is that genes are not fixed blueprints but dynamic switches. They respond to environmental signals. A gene for aggression may only be expressed under conditions of stress and poor nutrition. IQ has high heritability in wealthy families (because environment is relatively uniform and genetic variation is what differentiates outcomes) but low heritability in poor families (where environmental variation dominates). Genes cause some people to seek out certain environments, which then further shape them – a process called gene-environment correlation. The upshot is that nature and nurture are not opponents but collaborators. Ridley credits this insight partly to the work of Robert Plomin and the behavioural genetics tradition. Rational Optimism Grounded in Data Ridley’s optimism is not temperamental cheerfulness. It is a claim about empirical trend lines. Global average life expectancy, child mortality, per capita income, literacy rates, violence rates, famine deaths, hours worked per unit of consumption – all have improved dramatically over the past two centuries and most are still improving. He is a fierce critic of what he calls “the doom industry” – the recurring tendency of intellectuals, activists and journalists to extrapolate from short-term setbacks or local trends to global catastrophe. He catalogues a long list of predicted catastrophes that did not occur. The population bomb, resource exhaustion, acid rain’s destruction of forests, the ozone hole triggering mass cancer, peak oil ending civilisation. This does not make him a complacent denier of problems. He takes climate change seriously as a phenomenon (less seriously as an imminent catastrophe). His point is that the same processes of innovation and adaptation that have solved every previous resource constraint will continue to operate, provided markets, property rights and free inquiry remain intact. The Gradualism and Serendipity of Innovation How Innovation Works (2019) is the most focused expression of a theme present throughout Ridley’s career. Genuine innovation is never the product of a lone genius having a eureka moment. It is almost always incremental, builds on prior work, involves multiple independent discoveries (steam power, calculus, natural selection, the telephone – all had near-simultaneous independent inventors) and is driven by practical tinkerers rather than academic theorists. Invention follows demand and availability of enabling technologies; it does not precede them. The steam engine was improved by people who worked with steam engines, not by thermodynamicists. The Wright brothers were bicycle mechanics, not aeronautical theorists. The implication for policy is that innovation cannot be planned or directed but can be enabled by keeping barriers low, allowing failure, protecting intellectual property modestly (too much IP protection stifles recombination) and preserving freedom of inquiry. Book-by-Book Survey The Red Queen: Sex and the Evolution of Human Nature (1993) This book asks the oldest question in evolutionary biology - why does sex exist? Asexual reproduction is twice as efficient (no males needed). Ridley’s answer draws on the Red Queen hypothesis, named for the Lewis Carroll character who must keep running to stay in the same place. Sex exists because parasites exist. Parasites evolve rapidly to exploit host genomes. Sexual reproduction shuffles the host’s genome every generation, making it a moving target. The parasite that cracked last generation’s immune system finds this generation’s immune system has different locks. This arms race – parasite versus host – is the engine that keeps sexual reproduction advantageous despite its cost. Ridley extends the argument to human sexual behaviour. Mate choice, sexual jealousy, promiscuity differentials between sexes, male ornamentation (wealth, status, wit as peacock tails) and the evolution of the human brain itself as partly a product of sexual selection. He draws on Robert Trivers’s parental investment theory – females invest more per offspring and are therefore choosier – and applies it with unusual rigour to human data. The book also covers Geoffrey Miller’s hypothesis that human intelligence, language and creativity are sexually selected traits. Costly signals of genetic quality that attracted mates and thereby spread through the population. Whether or not every claim holds up, the book introduced a generation of readers to the power of sexual selection as an explanatory framework. Key concept: The Red Queen dynamic – co-evolutionary arms races that force continuous change just to maintain fitness – recurs in Ridley’s later work as a metaphor for competitive markets and technological change. The Origins of Virtue: Human Instincts and the Evolution of Cooperation (1996) This is Ridley’s most philosophically ambitious early book. Its central puzzle is the one that obsessed Hobbes, Rousseau and every subsequent social philosopher. Given that individuals pursue self-interest, how does cooperation arise and persist? The standard economic answer (rational self-interest in repeated games, policed by reputation) is part of the story, but Ridley shows it is insufficient. Humans cooperate at scales and with strangers in ways that cannot be explained by kin selection or simple reciprocity. His argument draws on evolutionary game theory, particularly Robert Axelrod’s tournaments showing that tit-for-tat strategies beat pure defection in iterated prisoner’s dilemmas and on the anthropology of trade and gift exchange. He argues that humans have evolved psychological instincts for reciprocity, fairness, punishment of cheaters, and in-group loyalty – instincts shaped by the conditions of the Pleistocene when our social group was small and repeated interaction was the norm. The crucial insight is that virtue – cooperation, fairness, generosity – is not the product of reason or divine command but of natural selection acting on social creatures. Hobbes was wrong: the state of nature was not a war of all against all. But Rousseau was also wrong: pre-civilisation humans were not noble savages. They were cooperative within their band and often brutal toward outsiders. Modern institutions – markets, property rights, the rule of law – harness and extend those evolved instincts to allow cooperation between strangers at civilisational scale. This is the seed of what becomes the full argument of The Rational Optimist. Key concept: The “tragedy of the commons” is not inevitable; human communities have repeatedly solved it through evolved norms, property rights and institutions – without requiring Leviathan to impose order. Genome: The Autobiography of a Species in 23 Chapters (1999) Published the year before the completion of the Human Genome Project, this book takes one gene from each of the 23 pairs of human chromosomes and uses it as a window into a different aspect of human biology and evolution. The structure is deliberately journalistic – each chapter is self-contained – but the cumulative effect is a comprehensive portrait of what genes do and do not do. Several chapters are especially important. The chapter on chromosome 6 introduces the major histocompatibility complex (MHC), the immune system’s genetic fingerprint, and connects it to the parasite-driven evolution of sex – a callback to The Red Queen. The chapter on chromosome 11 addresses addiction and the dopamine reward system, arguing that personality is partly heritable and that drug addiction is partly a matter of genetic susceptibility, not merely moral failure. The chapter on chromosome 17 covers BRCA1 and genetic predisposition to cancer, navigating the ethics of genetic testing with unusual care. Throughout, Ridley insists that genes do not determine outcomes mechanically. They are context-dependent. The same gene can produce different outcomes in different environments. The book was written before GWAS (genome-wide association studies) made it clear that most complex traits are influenced by thousands of genes each of tiny effect, so some of its examples of single-gene causation feel slightly dated – but the core argument about genetic influence being probabilistic rather than deterministic remains sound. Key concept: Genes are not destiny but propensity – probabilistic influences on development shaped by evolutionary history and modulated by environment. Nature Via Nurture: Genes, Experience and What Makes Us Human (2003) This is Ridley’s most technically demanding book and arguably his most important contribution to the nature-nurture debate. Its central argument is that the question “which matters more, genes or environment?” is malformed. Genes work through the environment. They are not alternative explanations but complementary mechanisms. Ridley introduces the concept of reaction norms: the same genotype produces different phenotypes (the set of observable characteristics of an organism resulting from interaction of its genotype with environment) depending on environment. He details the mechanism of gene expression – how transcription factors, methylation and environmental signals switch genes on and off. He uses the example of the serotonin transporter gene (5-HTTLPR): individuals with the short variant are more susceptible to depression, but only if they experience adverse life events. Without the adverse events, the short variant has little effect. This gene-environment interaction was one of the landmark findings of early 21st-century psychiatry (though it has been partly complicated by subsequent larger studies, the principle it illustrates remains robust). He also demolishes two extreme positions: the genetic determinism that says genes fix outcomes and the blank-slate environmentalism of John Watson and his successors that says genes are irrelevant. Twin studies show that virtually every measured psychological trait has significant heritability. But heritability estimates are population-level statistics, not individual predictions. And identical twins reared apart are not identical people – environment contributes substantially, even when measured heritability is high. The deepest point is about development: Organisms are not built from a genetic blueprint the way a house is built from architectural plans. They self-assemble through a process in which genes and environment are in constant dialogue. The genome is less a programme and more a recipe – and the dish – depends on the kitchen. Key concept: Genes are experience-dependent developmental switches. Nature requires nurture to express itself, and nurture operates through nature. Francis Crick: Discoverer of the Genetic Code (2006) This short biography is the outlier in Ridley’s catalogue – it is the only one that is primarily about a person rather than an idea. Ridley wrote it as part of the HarperCollins Eminent Lives series. It traces Crick’s career from his work on the X-ray crystallography of proteins at the Cavendish Laboratory, through the discovery of the double helix with James Watson in 1953, to his later work on consciousness at the Salk Institute. The book is useful for understanding Ridley’s intellectual heroes. Crick exemplifies the virtues Ridley values: willingness to transgress disciplinary boundaries, preference for mechanistic over vitalist explanations, comfort with uncertainty and revision, and the conviction that biology is chemistry – that life does not require special forces beyond physics. Ridley also highlights Crick’s fierce materialism about consciousness, which anticipates Ridley’s own materialist and reductionist leanings. Key concept: Scientific progress depends on individuals willing to ask questions that established disciplines have declared out of bounds. The Rational Optimist: How Prosperity Evolves (2010) This is Ridley’s most commercially successful book and the fullest expression of his economic and historical worldview. Its argument is in three parts. First, a historical claim. By almost every measurable standard of human welfare – longevity, nutrition, physical safety, leisure, access to information and goods – the average human alive today is better off than any human alive at any prior point in history, including kings and emperors. This is not a Western story; global averages have risen. Ridley marshals data from economists Angus Maddison, Julian Simon and others. Second, a causal claim. The mechanism behind this improvement is trade and specialisation. When people exchange, they gain access to the productivity of others. The collective brain grows. Technology accumulates because it is combinatorial – each new tool is a combination of prior tools – and exchange accelerates recombination. Ridley traces this from the first stone tool trade networks 200,000 years ago through the Bronze Age long-distance trade that brought tin and copper together, to the internet connecting billions. Third, a forward-looking claim. There is no reason to expect this process to stop. The ingredients that enabled the Great Enrichment – markets, property rights, relatively open borders for goods and ideas, scientific institutions – remain largely in place. The population bomb did not materialise (fertility rates fall as incomes rise). Peak oil was a myth (new extraction technologies keep finding more). African poverty is retreating (sub-Saharan GDP per capita has risen sharply since 2000). Climate change is a real challenge but likely manageable through adaptation and low-carbon innovation, not through de-industrialisation. The book’s most important analytical tool is the concept of ideas having sex. Ridley argues that the exchange of ideas is the root of innovation. Two ideas meet, recombine, produce a new idea, as genomes do in sexual reproduction. The internet is, on this view, the greatest idea-exchange platform in history and therefore the greatest engine of innovation ever constructed. Key concept: Prosperity is not extracted from nature or accumulated by force – it is generated by exchange and specialisation, and the more connected a society, the more prosperous it becomes. The Evolution of Everything: How New Ideas Emerge (2015) This book is the most philosophically explicit of Ridley’s works, and in some ways the most radical. Its argument is simple and sweeping. Almost nothing important in human history was designed. It evolved. Each of fifteen chapters takes a domain – the universe, morality, life, genes, culture, the economy, technology, the mind, personality, education, population, leadership, government, religion, money – and argues that the conventional narrative (which attributes the domain’s existence and improvement to a designer, a visionary, a lawgiver, a sovereign) is wrong. The real explanation is always some version of evolutionary emergence. Variation, selection, cumulation. The intellectual lineage Ridley claims is striking. He invokes Lucretius (who argued that the universe is entirely material and self-generating), Adam Smith (whose invisible hand is an evolutionary mechanism), Charles Darwin, Herbert Spencer (cautiously), Friedrich Hayek, and the complexity theorists. He is explicit that this is a worldview – a systematically bottom-up epistemology – as opposed to the systematically top-down epistemology he sees dominating mainstream political and intellectual culture. Several chapters are particularly rich. The chapter on morality argues that ethical norms are not the product of divine command or Kantian reason but of evolutionary and cultural selection – norms that allowed cooperation spread, those that undermined it contracted. The chapter on technology extends the argument of The Rational Optimist. Technologies evolve by recombination and selection, not by planned invention. The chapter on government is the most politically charged. States, Ridley argues, have systematically over-claimed credit for improvements that were the product of spontaneous social order and their interventions have as often retarded as advanced human welfare. Key concept: Emergent order is both more common and more productive than designed order. The attempt to replace emergence with design – in economies, ecosystems, cultures – almost always makes things worse. How Innovation Works: And Why It Flourishes in Freedom (2019) This book is narrower in scope than The Evolution of Everything but richer in historical detail. It takes approximately fifty significant innovations – steam power, vaccines, the Green Revolution, the internet, fracking, smartphones, batteries – and analyses how each actually happened. The recurring findings are striking and consistent. Innovation is gradual, not sudden. The steam engine improved incrementally over a century; Watt’s “invention” was one step in a long sequence. Vaccines were developed through trial-and-error empiricism long before the germ theory that would later explain why they worked. GPS emerged from a military project that combined insights from dozens of prior technologies. Innovation is serendipitous, not directed. Penicillin was discovered by accident. CRISPR was found by scientists studying bacterial immune systems with no thought of gene editing. The internet was built for resilience in military communication; its commercial applications were entirely unanticipated. This is not to say lucky accidents cannot be cultivated – Pasteur’s dictum that chance favours the prepared mind holds – but it does mean that innovation cannot be reliably produced by directing research toward predetermined goals. Innovation is collaborative and simultaneous. The independent invention of the same technology by unconnected people – Darwin and Wallace, Bell and Gray filing for the telephone patent on the same day, calculus by Newton and Leibniz – shows that invention is often an idea whose time has come, emerging from a shared pool of enabling knowledge. This has implications for how we should think about intellectual property. Patents reward and encourage innovation but also block the recombination that drives it. Ridley advocates for shorter, narrower patents and broader open-source norms. Innovation is done by practitioners, not theorists. The Wright brothers succeeded where Langley failed partly because they were bicycle mechanics who understood how to build controlled, steerable machines, not aeronautical engineers with formal theories. The early computer engineers were electronics tinkerers. Fracking was developed by independent oil drillers, not by petroleum engineers in major oil companies. The policy implication is consistent with all of Ridley’s earlier work. The role of government in innovation is primarily to stay out of the way, protect intellectual property modestly, allow failure and maintain the open exchange of ideas across borders and disciplines. Key concept: Innovation is an evolutionary process – gradual, combinatorial, driven by practitioners, impossible to predict or direct and dependent on freedom to experiment and fail. Viral: The Search for the Origin of Covid-19 (2021) – co-authored with Alina Chan This is Ridley’s most journalistically urgent book, written as the debate about Covid-19’s origins was at its most politically charged. It is co-authored with Alina Chan, a molecular biologist at the Broad Institute who had published one of the first serious scientific papers questioning the natural spillover hypothesis. The book does not claim to prove that Covid-19 was engineered or that it leaked from a laboratory. Its argument is more modest and more important. The question of origin has been systematically suppressed and the evidence for a laboratory leak from the Wuhan Institute of Virology is substantial enough to warrant serious investigation. Ridley and Chan document - the unusual properties of SARS-CoV-2’s furin cleavage site (absent in related bat coronaviruses and characteristic of gain-of-function research); the WIV’s record of collecting and manipulating bat coronaviruses, including ones closely related to SARS-CoV-2; the failure of early searches for an intermediate animal host; the removal of the WIV’s virus database in September 2019; and the involvement of EcoHealth Alliance in US-funded research at the WIV on chimeric coronaviruses. The book is also a study in how science can be distorted by political and institutional pressures. Ridley documents how the Lancet’s early statement dismissing the lab-leak hypothesis as a conspiracy theory was drafted by Peter Daszak of EcoHealth Alliance, who had an undisclosed conflict of interest. He tracks how scientists who raised questions about natural origin were silenced, and how the mainstream media amplified the consensus without scrutinising its basis. This is consistent with themes in Ridley’s earlier work about the tendency of institutions to protect themselves and suppress inconvenient heterodoxy. Key concept: Scientific consensus can be manufactured by institutional interests. the obligation of science journalism is to follow evidence, not defer to authority. Cross-Cutting Themes and Intellectual Influences Across nine books, a consistent intellectual genealogy is visible. Ridley’s deepest debts are to: Charles Darwin – the master of explaining complex adaptation through unintelligent process. Ridley extends Darwinian thinking far beyond biology, treating it as the universal theory of cumulative beneficial change. Adam Smith – not just the author of The Wealth of Nations but of The Theory of Moral Sentiments, which argued that sympathy, fairness and reciprocity are natural human instincts. Ridley sees Smith as the first great theorist of spontaneous order and cultural evolution. Friedrich Hayek – whose critique of central planning rested on the epistemic impossibility of aggregating distributed knowledge. Ridley explicitly Hayekises every domain he enters: genes, cultures, technologies, institutions. Robert Trivers – whose theories of parental investment, reciprocal altruism and parent-offspring conflict gave evolutionary biology its framework for understanding social behaviour. The Red Queen and The Origins of Virtue are deeply Triversia. Julian Simon – the economist who bet Paul Ehrlich that resource prices would fall despite population growth, and won. Ridley adopts Simon’s insight that human ingenuity is the ultimate resource and that more people means more problem-solvers. Richard Dawkins – whose gene-centred view of evolution and concept of the extended phenotype underpin Ridley’s genetics writing. Ridley is a committed gene-selectionist. David Deutsch – whose epistemology in The Beginning of Infinity – that all problems are soluble in principle given the right knowledge and institutions – reinforces Ridley’s optimism. Ridley reviewed Deutsch’s book enthusiastically and draws on him in The Evolution of Everything. Criticisms and Limitations A fair survey must note where Ridley’s framework strains. His optimism about markets and spontaneous order is sometimes insufficiently attentive to market failures – externalities (notably pollution), public goods (basic research) and the distribution of gains from trade. His Hayekian instinct against central planning can tip into a reflexive anti-statism that his own historical evidence does not always support. Several of his innovation case studies involve significant public funding (GPS, the internet, vaccines). His treatment of climate change has evolved. In The Rational Optimist he significantly downplayed climate risk and overstated the case for adaptation, earning him justified criticism from climate scientists. He has partially walked this back but remains on the sceptical end of the spectrum on policy responses, particularly carbon taxes. His Viral co-authorship made him a controversial figure in science journalism, though the lab-leak hypothesis he defended has since received substantially more serious treatment from intelligence agencies and virologists, vindicating at least his insistence that the question deserved investigation. More broadly, his model of cultural evolution is sometimes too loose: the analogy between genetic selection and cultural selection is powerful but the mechanism of cultural inheritance (imitation, instruction, communication) differs enough from genetic inheritance that conclusions drawn from one do not always transfer cleanly to the other. The Ridley Synthesis Strip away the specific claims and what remains is a coherent philosophical system: the universe is not designed but self-organising; complex beneficial outcomes arise from decentralised processes of trial, error and selection; the appropriate response to this is to protect the conditions under which these processes operate (markets, property rights, free inquiry, political pluralism) rather than to substitute expert planning; and the long-run trajectory of human welfare is upward because the collective brain is growing. This system puts Ridley firmly in a specific intellectual tradition – classical liberal, broadly libertarian, deeply empiricist – and it gives his work both its consistency and its blind spots. He is at his best when demolishing naive narratives of progress-by-design or catastrophe-by-trend-extrapolation, and at his most questionable when the anti-planning instinct tips into a complacency about genuine market failures. What makes him indispensable is not that he is always right but that his framework forces readers to confront a question they would rather avoid. How much of what we attribute to intention, design and leadership is in fact the product of processes that nobody planned and nobody fully understands? The honest answer, across biology, economics and culture, is - more than we think. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

5. 

   11 Jul 2026

   Frontier Talent Paradox - HR Management for Edge Domains

   There is a mode of thinking about talent acquisition and workforce management that, once internalised, reveals why most organisations hiring in cyber security and artificial intelligence are structurally guaranteed to fail at it. The mistake most hiring managers and HR leaders make is treating these domains as variants of ordinary technical recruitment. Write a job description, post it, screen for credentials, run a competency interview and select the person who best matches the specification. What the person who has genuinely internalised this mode of thinking sees instead is that high-skill adversarial and generative domains operate on fundamentally different talent economics than the rest of the technical workforce. Credentials lag capability by years. Interview processes designed to assess competence in stable domains actively select against the dispositions that make someone exceptional in dynamic ones. The job descriptions organisations write for red teamers and AI researchers describe, almost perfectly, the kind of person who will not be exceptional at either role. The central insight that unlocks most of the value in this field is this - in domains where the work itself is to think in ways that haven’t been formalised yet, the standard HR apparatus, built for repeatability and legal defensibility, is not merely imperfect but actively counterproductive. Managing it requires building a parallel talent system that runs on different logic. The Domain Problem: Why High-Skill Technical Roles Break Standard HR Logic Geoffrey Moore’s distinction in Crossing the Chasm between technology enthusiasts and pragmatists maps, with some modification, onto a more fundamental divide in technical hiring - between roles where the work is to execute within a defined system and roles where the work is to operate at or beyond the system’s edge. Cyber security, red teaming and AI research are edge-domain roles. The value they produce comes precisely from the practitioner’s ability to think in ways the organisation has not anticipated. Roger Martin’s The Opposable Mind frames this as the capacity to hold two conflicting models simultaneously and produce a resolution that neither model alone would generate. A red teamer who can only think like a defender, or an AI researcher who can only iterate within the current paradigm, is not producing the output the role exists to generate. The HR function, as codified in David Ulrich’s business partner model and operationalised through competency frameworks like the SHRM Body of Applied Skills and Knowledge, was designed for a different problem. Its logic is one of specification, matching and verification. You define what good looks like, you find candidates who demonstrate it and you build processes that are defensible, consistent and legally compliant. This works when what good looks like is stable and when past performance is a reasonable predictor of future performance in similar conditions. In cyber security and AI, neither assumption holds. The threat landscape in offensive security changes faster than any competency framework can be updated. The frontier of AI capability shifts on a timescale measured in months. A red teamer credentialed three years ago who has not stayed at the edge of their practice is, in operational terms, a different person from the one who was credentialed. The competency framework has no mechanism to capture this. Cal Newport’s So Good They Can’t Ignore You makes the distinction between passion-driven and craftsman-driven career development, arguing that rare and valuable skills are built through deliberate practice at the edge of current ability. The implication for hiring is that the candidates who are most valuable in high-skill technical domains are those who have been running their own deliberate practice regime, often outside formal employment, in communities, in competitions, in personal research and in open-source contribution. The HR process that screens for formal credentials and employment history systematically misses them. The Talent Pool: Where Exceptional Practitioners Actually Come From Edward Luce’s The Retreat of Western Liberalism is not a talent management text but its central observation about the bifurcation of economic returns to skill is directly relevant here. In domains where the distance between median and top-decile performance is large, the returns to finding and retaining the top decile are asymmetric. In cyber security, a single exceptional red teamer who identifies a critical architectural flaw before an adversary does may generate value orders of magnitude larger than a team of adequate practitioners who do not. Daniel Kahneman’s Thinking, Fast and Slow documents the consistent finding that expert performance in complex domains is poorly predicted by the kinds of structured interviews that HR processes favour. The cognitive processes that produce insight in adversarial domains are not the same as those that produce coherent answers to behavioural interview questions. The talent pool for exceptional red teamers and AI researchers is characterised by several features that standard recruitment processes treat as liabilities. First, non-linear career paths. The most capable offensive security practitioners frequently have histories that include periods of self-directed learning, participation in capture-the-flag competitions, independent vulnerability research and sometimes legally ambiguous activity in their earlier careers. The background check and clean-record requirement that HR compliance functions treat as non-negotiable filters cut deeply into this pool. Bruce Schneier’s work on security culture, particularly Secrets and Lies, frames the adversarial security mindset as one that requires deep familiarity with how systems fail, which is not a mindset that develops in environments of pure compliance. Second, community embeddedness. In both offensive security and AI research, the primary epistemic community is not the employing organisation but the broader professional network. DEF CON, Black Hat, the NeurIPS and ICLR conferences and their associated workshops and side channels are where the actual frontier knowledge lives. Practitioners who are not embedded in these communities are not at the frontier, regardless of their seniority within a corporate structure. Third, intrinsic motivation structure. In Drive, Daniel Pink identifies autonomy, mastery and purpose as the three components of intrinsic motivation in complex cognitive work. The employment structures that organisations default to, fixed hours, defined scope, hierarchical approval processes, systematically undermine all three. The implication is that the talent pool for these roles is not defined by geography, by educational pedigree or by formal employment history. It is defined by participation in a set of communities and by evidence of self-directed work at the frontier. Recruiting processes that do not reach into these communities are fishing in the wrong water. The Credential Problem: What Certifications Do and Do Not Signal The certification ecosystem in cyber security is extensive. CISSP, CISM, CEH, OSCP, GPEN, GXPN and the Offensive Security family of certifications represent decades of institutionalised effort to create verifiable signals of practitioner competence. In AI, the equivalent markers are degrees from recognised graduate programmes, publications in peer-reviewed venues and contributions to widely used frameworks. These credentials are not worthless. They signal a baseline of structured learning and, in the case of Offensive Security certifications, demonstrated ability to complete practical tasks under time pressure. But they are systematically misused in hiring, in two directions. The first misuse is treating certifications as proxies for current capability. Ryan Naraine and the team behind the annual Verizon Data Breach Investigations Report have documented consistently that the attack techniques most commonly used in successful breaches are not the most technically sophisticated ones. They are the ones that work against the defensive posture that certified practitioners have been trained to implement. The certification curriculum is always behind the current threat. An OSCP holder from two years ago who has not continued active practice knows how to compromise the environments that existed when they trained, not the current ones. The second misuse is treating the absence of credentials as a signal of incompetence. The most capable offensive security researchers and AI practitioners are frequently the people who were too busy doing the work to sit the exams. Bug bounty programmes, coordinated vulnerability disclosure records and competition results are stronger signals of current offensive capability than most certifications. In AI, a GitHub repository with a meaningful implementation of a recent technique or a thread of technical commentary on arXiv papers that demonstrates genuine comprehension, signals more about frontier capability than an MSc from a programme whose curriculum was designed five years ago. David Epstein’s Range makes the case that in complex, wicked domains, breadth of experience and cross-domain pattern recognition produce more adaptive capability than narrow deep specialisation. The implication for hiring is that the candidate whose background does not fit the job description template may be more capable in the role than the candidate who fits it perfectly. The practical correction is not to abandon credentials but to use them as threshold filters rather than ranking criteria, and to supplement them with domain-specific assessments that test current capability rather than historical learning. The challenge is that building such assessments requires internal expertise that many organisations lack, which generates a dependency. You need people who can assess frontier capability to hire people with frontier capability, and if you do not already have them, you cannot build the assessment. The Assessment Problem: Designing Evaluations That Select for the Right Thing Laszlo Bock’s Work Rules, drawing on Google’s large-scale hiring research, is one of the most rigorous empirical treatments of what actually predicts job performance. Its central finding is that unstructured interviews, which is what most senior technical hiring looks like, have a predictive validity approaching zero. Structured interviews with well-defined scoring rubrics do better. Work sample tests, where candidates actually do a representative version of the job, do best. In high-skill technical domains, this logic points clearly toward practical assessments as the primary evaluation instrument. The challenge is designing them correctly. For offensive security and red teaming roles, the practical assessment needs to reflect the actual structure of the work. Red teaming is not a series of isolated technical challenges. It is a sustained, adaptive campaign against a defended target that requires creativity, patience, operational security discipline and the ability to think from the perspective of an actual adversary with specific objectives. The capture-the-flag competition format, which many organisations use as an assessment, tests a subset of these skills under artificial conditions. A well-designed red team assessment gives candidates a scoped engagement against a realistic environment, assesses their methodology and documentation as well as their technical outcomes and evaluates their ability to communicate findings to a non-technical audience. The last criterion is consistently underweighted in technical hiring and consistently important in practice, because the value of a red team engagement is not the compromise of the target but the change in defensive posture that the report produces. For AI roles, the assessment problem is different but structurally similar. The standard technical interview in AI hiring is a combination of coding challenge, mathematical derivation and paper discussion. These assess a narrow slice of the skills that produce genuine research or engineering impact. The ability to identify a promising research direction, to debug a training run that is behaving unexpectedly, to translate a theoretical understanding into a working implementation under uncertainty and to communicate clearly about the limits of a model’s behaviour are not well assessed by these formats. Stuart Russell and Peter Norvig’s Artificial Intelligence: A Modern Approach provides the canonical framework for reasoning about AI systems as agents operating under uncertainty. The hiring assessment that actually tests for the reasoning this framework describes would look more like a case discussion about a poorly behaved model than a coding screen. The deeper issue, which few organisations confront directly, is that the best assessments for frontier technical roles require the assessors to be at or near the frontier themselves. The standard HR solution of having a trained HR interviewer conduct the first-stage screen and a technical panel conduct the second stage assumes that the technical panel can evaluate the candidate accurately. In domains where the field is moving fast, the panel may not be able to. This is not a failure of individual assessors but a structural consequence of the pace of change. The organisation that is two years behind the frontier in its internal capability will systematically undervalue the frontier candidate. The Job Description Problem: How Specifications Exclude the People You Need The job description is the primary instrument of talent attraction in most organisations. It is also, in high-skill technical domains, a primary instrument of talent exclusion. The typical job description for a senior red teamer or AI researcher is a compilation of credentials, tool familiarity and years of experience that reflects not what the role requires but what the hiring manager and HR function are confident they can verify. The result is a specification that selects for a population whose members are verifiably adequate rather than potentially exceptional. The credential inflation problem in cyber security job descriptions has been documented in workforce studies by the (ISC)2 Global Cybersecurity Workforce Study, which has consistently found a large gap between the number of positions open and the number of qualified candidates available, while simultaneously finding that many of those positions specify requirements that exclude a large fraction of the people capable of doing the work. A red team job description that requires five years of experience with a specific commercial platform is not specifying a capability requirement. It is specifying a history requirement that correlates weakly with the capability and excludes the self-taught practitioner who has equivalent or superior skill without the pedigree. In AI roles, the equivalent pathology is the degree requirement. The contribution of a doctorate from a top programme to actual AI engineering or research capability is real but limited. It signals exposure to rigorous methodology, some depth of expertise and the ability to complete a long-form technical project. It does not signal the ability to move quickly on a practical problem, to collaborate effectively in a fast-moving team or to maintain productive uncertainty about whether the current approach is the right one. The most impactful practitioners in applied AI, as opposed to frontier academic research, have frequently taken non-traditional paths, and organisations that hard-code degree requirements are trading access to that population for the defensibility of the hiring decision. The correction here is not to abandon specifications but to write them around capability signals rather than credential proxies. For a red team role, the capability signals include evidence of successful engagement outcomes, participation in the research community, quality of published work whether formal or informal and performance in a practical assessment. For an AI role, they include the quality of implemented work, the track record of shipping systems that behave as intended and the ability to reason clearly about where a model’s outputs cannot be trusted. Writing specifications around these signals is harder than writing them around credentials, because it requires the organisation to have a view about what capability looks like rather than what the credential market signals. But it is the only way to access the full depth of the talent pool. The Retention Problem: Why the People You Most Want to Keep Are the Most Likely to Leave Voluntary attrition in cyber security and AI is structurally higher than in most other technical domains and the organisations that suffer most from it are those that have failed to understand why. The proximate causes are well documented. Compensation competition from the financial sector and hyperscale technology companies, faster career progression in smaller organisations and the pull of independent consulting and research. But the structural cause runs deeper. Patrick Lencioni’s The Five Dysfunctions of a Team identifies trust as the foundational layer of team performance and the cyber security domain has a specific trust problem that conventional retention strategies do not address. Red teamers and offensive security practitioners are employed to think like adversaries, to find the things the organisation does not want found and to report them accurately even when the findings are politically uncomfortable. This requires an organisational culture that actually rewards the honest delivery of bad news and that does not, under pressure, subordinate accurate threat assessment to the preferences of senior stakeholders. Organisations that say they want a red team capability but respond to red team findings with defensive denial or minimisation are not actually maintaining a red team capability. They are maintaining an appearance of one. The practitioners who can see this distinction clearly are the ones most likely to leave. For AI practitioners, the retention problem has a different character. The intrinsic motivation structure that attracts people to AI research is typically a combination of intellectual curiosity, the desire to work at the frontier of a rapidly developing field and a sense that the work has significant consequences. All three components are sensitive to organisational context. Intellectual curiosity is suppressed by environments that require extensive justification for exploratory work. Frontier engagement requires access to compute, data and time for self-directed learning that many organisations do not provide. The sense of consequence is complicated, in ways that are now extensively discussed, by the emergence of AI safety as a field that identifies serious potential harms from the technology that practitioners may be building. Stuart Russell’s Human Compatible and Nick Bostrom’s Superintelligence have made these concerns broadly legible within the practitioner community. Organisations that refuse to engage with them seriously, as distinct from organisations that engage with them and reach different conclusions, signal to practitioners that they do not take the moral dimensions of the work seriously, which is a powerful retention risk for the fraction of the talent pool that is most reflective. The empirical research on retention in knowledge-intensive roles, reviewed in Jacob Morgan’s The Employee Experience Advantage, points to three factors that dominate. The sense that the work is meaningful, the quality of the immediate manager and the culture of the team. None of these are primarily compensation phenomena, though compensation that feels inequitable amplifies all of them. The retention strategy for high-skill technical talent that focuses primarily on compensation is addressing the symptom rather than the cause and is bidding in an auction that the organisation is structurally likely to lose against better-capitalised competitors. The Management Problem: Leading People Whose Work You Cannot Fully Evaluate The management challenge that is specific to high-skill technical domains and almost never honestly addressed in management literature is this - how do you manage people whose work you cannot independently assess? In most management contexts, the manager has sufficient domain knowledge to evaluate the quality of subordinates’ output, to recognise when someone is drifting from good practice and to distinguish genuine difficulty from poor effort. In frontier cyber security and AI roles, this condition frequently does not hold. The red team manager who has not maintained active offensive practice cannot assess whether the team’s engagement methodology is current. The AI engineering manager whose last hands-on technical work was three years ago cannot evaluate whether the model architecture decisions being made are appropriate. This is not a problem that can be solved by generic management skill. The frameworks that dominate management development, from situational leadership theory through to the OKR system popularised by John Doerr’s Measure What Matters, assume that the manager has or can develop sufficient domain understanding to set meaningful objectives and evaluate progress against them. In fast-moving technical domains, this assumption breaks down at the frontier. You cannot write a meaningful OKR for a red team engagement if you do not understand the current threat landscape. You cannot evaluate whether an AI research team is making good use of its compute budget if you do not understand the current state of the art in the technique they are using. The practical responses to this are limited. One is to maintain technical depth at the management level, which requires either hiring managers who are current practitioners or investing substantially in keeping managers current, both of which have real costs and often conflict with the normal career progression logic that moves good practitioners into management. Andy Grove’s High Output Management argues for the manager as the multiplier of team output and the creator of conditions for high performance rather than as the technical director of individual contributor work. This framing works better for high-skill technical domains than the technical director model, but it requires the manager to have a clear view of what high performance looks like from the outside, even if they cannot evaluate it from the inside. In practice, this means maintaining strong peer review culture within the team, cultivating external networks that provide an independent calibration of the team’s work and being explicit about the limits of the manager’s technical evaluation capacity rather than papering over them. The second response is to separate the management of people from the technical governance of work, creating lead practitioner roles that carry technical authority without full managerial responsibility. This is structurally common in research organisations and is increasingly common in mature cyber security functions. It has the advantage of keeping technical judgement in the hands of those best placed to exercise it, and the disadvantage of creating coordination complexity and sometimes ambiguous accountability. The Culture Problem: Building Organisations That Can Sustain Adversarial Thinking The deepest problem in HR management for red teaming and AI safety roles is not a process problem but a culture problem. Both fields require the organisation to genuinely value the production of uncomfortable findings and to build structures that protect and reward the people who produce them. This is harder than it sounds. Edgar Schein’s Organizational Culture and Leadership identifies three levels of culture viz artefacts, espoused values and basic assumptions. The artefacts of an organisation that says it takes security seriously are visible: a red team function, a security operations centre, an AI ethics committee. The espoused values are articulable: we prioritise security, we are committed to responsible AI development. The basic assumptions are what actually governs behaviour when the findings produced by the red team or the AI ethics committee conflict with commercial priorities. Schein’s framework predicts, and empirical observation confirms, that it is the basic assumptions that determine outcomes, and that artefacts and espoused values that conflict with basic assumptions produce cultural dysfunction rather than the intended outcomes. The red team that is not allowed to test the most critical production systems because the risk is judged too high is not actually a red team. It is a compliance theatre function. The AI ethics review process that has no authority to delay or halt a product launch is not a governance mechanism. It is a documentation mechanism. The practitioners who work in these functions can see the gap between the espoused value and the basic assumption clearly, and the ones who care most about the work are, again, the most likely to leave. Amy Edmondson’s research on psychological safety, documented in The Fearless Organization, shows that teams perform better on complex tasks when members feel safe to raise concerns, report errors and challenge prevailing views without fear of negative consequence. Psychological safety is not a soft cultural amenity. It is a functional requirement for any team whose output depends on honest assessment of a situation that powerful people may not want to hear assessed honestly. Red teams and AI safety functions both require it. Neither can be built in environments where the basic assumption is that the job is to validate decisions already made. The cultural condition for sustaining adversarial thinking is, paradoxically, a leadership culture that is genuinely comfortable with being wrong. James Reason’s Managing the Risks of Organizational Accidents shows, through detailed analysis of high-consequence failures in aviation, nuclear and chemical industries, that organisations with the strongest safety records are those that normalise the reporting of near-misses and minor failures rather than suppressing them. The same logic applies to cyber security - the organisation that treats a red team finding as evidence of a problem with the defenders rather than a valuable input to the defensive programme is on the wrong side of this distinction. The Compensation Problem: Getting the Economics Right Compensation in cyber security and AI has become genuinely complex in a way that standard HR compensation frameworks handle poorly. The market for frontier talent in both domains is global, the supply is constrained and the demand from technology companies, financial services firms and government contractors creates a competitive dynamic that most organisations’ compensation bands were not designed for. The issue is not simply that the market rate is high, though it is. The issue is that the market rate is highly bimodal. The compensation premium for genuinely frontier-capable talent over competent but not frontier talent is large and growing. The HR response of benchmarking to the 75th or 90th percentile of the relevant market category systematically underpays for frontier talent because the frontier practitioners are in the tail of the distribution and the benchmark is a central tendency measure. Edward Lazear’s work on personnel economics, particularly the tournament theory of executive compensation, suggests that in markets where the value of being at the frontier is very high, the compensation differential between positions in the hierarchy will be set by the value of the top prize rather than by traditional incremental frameworks. The implication for red team and AI talent is that the organisation needs to understand which roles are genuinely frontier-critical and be prepared to pay in the tail of the distribution for them, which requires making explicit decisions about which roles are and are not in that category rather than applying a uniform banding approach. The non-monetary dimensions of the compensation package are, in this domain, unusually important. Conference attendance, time for personal research and open-source contribution, access to high-end compute resources, the ability to publish findings and the organisational status of the function all function as compensation in the talent market for people whose primary motivation is mastery and frontier engagement. An organisation that pays at the 85th percentile of salary but does not allow its practitioners to attend DEF CON or publish research is not actually competing well on total compensation. An organisation that pays at the 70th percentile but gives its AI researchers meaningful access to frontier compute and publication rights may be. The compensation strategy needs to be designed with this population’s actual preference structure in mind rather than extrapolated from the broader technical workforce. The Diversity Problem: Structural Barriers in Fields With Non-Standard Talent Pipelines The front end of the talent pipeline for red teamers and AI researchers is the informal community: the capture-the-flag competitions, the open-source communities, the research seminars and the conference networks. Access to these communities is not uniformly distributed. It is correlated with socioeconomic background, with educational institution, with geography and, particularly in offensive security, with the social networks that give people access to the knowledge that these pathways exist. The person from a disadvantaged background who could be exceptional at offensive security is unlikely to have encountered the CTF community as a pathway. The person without access to a well-resourced university environment is less likely to have the compute access and mentorship that produces frontier AI research capability. The organisations that have made genuine progress on diversity in these domains have done so by going upstream: funding and participating in community-level programmes that expand access to the front end of the pipeline. SANS CyberTalent, the UK government’s CyberFirst programme and the Recurse Center in AI are examples of interventions that address pipeline access rather than just pipeline selection. For organisations that are serious about diversity in these roles, HR investment at the front of the pipeline will produce better outcomes than investment in selection process modification at the back, because the selection process cannot diversify a pipeline that is not diverse. The Unified Picture The single most important insight in this field is that in high-skill adversarial and generative technical domains, the standard HR system is not a neutral instrument that can be calibrated for better performance. It is an instrument designed for a different problem, and applying it to this problem produces systematic errors that compound. The credential system is behind the capability frontier. The interview process selects against the dispositions most valuable at the frontier. The job description format excludes non-standard but highly capable candidates. The compensation framework misses the tail where the value is concentrated. The retention strategy addresses proximate rather than structural causes of attrition. The management framework assumes evaluative capacity the manager does not have. The culture defaults to espoused values that conflict with basic assumptions. None of these are independent problems. They are expressions of a single underlying miscalibration - the application of a repeatability-and-compliance logic to a domain that runs on adaptability-and-frontier logic. The 20 percent of concepts that produce 80 percent of the field’s explanatory power are these. First, the talent pool for frontier technical roles is defined by community participation and demonstrated work, not by credentials and employment history and recruitment that does not reach into the relevant communities is not accessing the pool. Second, practical assessment is the only reliable evaluation instrument for roles where the work cannot be fully specified in advance and building assessment capability requires internal frontier expertise. Third, the retention of high-skill technical talent is driven primarily by autonomy, mastery and the sense that the work has genuine consequence and by the alignment or misalignment between the organisation’s espoused values and its basic assumptions about whether uncomfortable findings are welcome. Fourth, the management of people whose work you cannot fully evaluate requires a structural response, not a personal skill response, specifically the separation of people management from technical governance and the cultivation of external calibration mechanisms. Fifth, the culture required to sustain adversarial thinking is a leadership culture that is rewarding the production of uncomfortable findings rather than suppressing them, and this cannot be created by HR policy without leadership behaviour that reinforces it consistently. These five concepts connect directly. A talent pool that is defined by community participation and demonstrated work requires a recruitment process that reaches into those communities and can evaluate demonstrated work, which requires internal frontier expertise, which is a retention and management challenge, which is a culture challenge at its root. The failure to build the right culture makes it impossible to retain the people who would enable you to build the right recruitment and assessment process, which means you cannot access the talent pool, which means the culture problem compounds. The whole system is load-bearing at every joint. The Limits The framework described above has real limits that its practitioners sometimes fail to acknowledge. The first is the problem of verification at scale. The argument for community participation and demonstrated work as hiring signals is compelling for frontier roles. It is much harder to operationalise for the large volume of mid-level roles that constitute the majority of headcount in any mature cyber security function. The organisation that needs to hire 50 security analysts as well as five frontier red teamers cannot build bespoke practical assessments and community-sourcing pipelines for every role. The standard HR apparatus, for all its flaws in frontier hiring, has genuine advantages in consistency, legal defensibility and the ability to process volume. A sophisticated talent strategy needs to segment clearly between the roles where the frontier logic applies and the roles where standard logic is adequate. The second limit is the insularity risk. The emphasis on community embeddedness and frontier participation can, if taken too far, reproduce existing social networks as hiring criteria under a different label. The people who are embedded in DEF CON and NeurIPS communities are not a random sample of the potential talent pool. They are people who had the access, the social capital and the resources to participate in those communities. Using community participation as a hiring signal without attention to who gets access to the community reproduces the diversity problem at a different level. The third limit is the culture solution’s dependency on leadership. The argument that building a culture that rewards adversarial findings requires leadership behaviour that reinforces uncomfortable honesty is correct. But it is also a description of something that is largely outside the control of the HR function. HR can design the processes, write the policies and build the training. It cannot change the basic assumptions of the leadership team. Organisations whose senior leadership is constitutionally uncomfortable with honest bad news will not build the culture required to sustain a genuine red team or AI safety function regardless of how well the HR strategy is designed. The fourth limit is that the field itself is not value-neutral. The capabilities that red teamers and AI researchers develop are dual-use in a profound sense. The person who is excellent at identifying and exploiting vulnerabilities is also the person who could exploit them without authorisation. The person who is excellent at building powerful AI systems is also building the technology whose misuse the AI safety community is most concerned about. The HR function has a role in assessing trustworthiness and ethical orientation that the framework above underweights. Personnel security vetting, ethical reference checking and ongoing monitoring of practitioner behaviour are not separable from talent strategy in these domains. Treating them as purely compliance functions rather than as components of the talent system is a mistake that organisations occasionally pay for at very high cost. What This Demands of You If you manage or advise on hiring in cyber security, red teaming or AI, the mode of thinking this essay describes demands specific changes in how you operate, not general improvements in attitude. You need to stop using credentials as ranking criteria and start using them as threshold filters. This means you need to know what the threshold is, which means you need to have, or have access to, someone who is currently at the frontier of the domain and can tell you what minimum competence looks like right now, not two years ago. You need to build or commission practical assessments that reflect the actual structure of the work. For red team roles, this means scoped engagements with evaluation of methodology and communication, not just technical outcome. For AI roles, this means case-based evaluation of reasoning under uncertainty, not coding screens and paper discussions. If you do not have the internal expertise to design these assessments, your first hire should be someone who does. You need to rewrite your job descriptions around capability signals rather than credential proxies. This is not a minor editing task. It requires a clear view of what the role actually produces and what evidence of the ability to produce it looks like. That view must come from domain experts, not from HR frameworks. You need to audit the gap between your organisation’s espoused values about security and AI safety and its basic assumptions about what happens when findings conflict with commercial priorities. If you find a gap, closing it is a leadership problem, not an HR problem. Your job is to name the gap clearly, not to paper over it with process. You need to manage your pipeline, not just your positions. The talent pool for frontier roles is built at the front end through community investment, education programmes and early career pathways. If you are only present at the point of hire, you are competing for a pool you have not helped to create, against organisations that have. You need to separate technical governance from people management in your senior technical roles and create lead practitioner tracks with real authority and real status. The person who should be assessing the quality of your red team’s methodology should be a current practitioner with credibility in the field, not the person who is responsible for their annual review. And you need to ask, about every significant hiring and management decision in these domains, whether the process you are using was designed for a domain where the work is repeatable or a domain where the work is to do what hasn’t been done before. If it was designed for the former, it is the wrong tool, and using it is not a neutral choice. It is a choice to select systematically for the wrong people. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

6. 

   3 Jul 2026

   On Death

   Death is the only subject that is simultaneously the most universal human experience and the most comprehensively avoided topic in modern culture. Every human being who has ever lived has died. Every human being alive now will die. Every relationship you have will end either in your death or in the death of the person you love. The loss is total, the timeline is uncertain and the fact is absolute. And yet the dominant culture of the modern developed world has organised itself with remarkable thoroughness around the project of not thinking about this. The books referred in this piece collectively constitute the most serious available confrontation with what death is, what it does to the living, what it reveals about the lives being lived and what a genuinely mature relationship with mortality makes possible that its avoidance permanently forecloses. This is not comfortable reading. It is, in the precise sense of the word, the most important reading available to any person who takes their life seriously. The Foundation: Death Is Running Your Life Whether You Know It Or Not Becker’s Denial of Death is the starting point because it performs the most important single reorientation in this entire body of work. It does not tell you how to die well or how to grieve well or how to make peace with mortality. It tells you something more fundamental and more disturbing: that the terror of death is already running your life, has been running it since you were old enough to understand that you would die and will continue running it until you engage with it directly enough to choose your response rather than having your response chosen for you by the avoidance mechanism. His central argument, which he developed from the work of Otto Rank and which Solomon, Greenberg and Pyszczynski subsequently confirmed with thirty years of empirical research, is that human beings occupy a unique and uniquely terrible position in the animal kingdom. We share with every other animal the biological drive for self-preservation. Unlike every other animal, we have the cognitive capacity to know that self-preservation will ultimately fail. We know we will die. We know it with certainty. We know it could happen at any moment. And we know that nothing we do will prevent it. This knowledge, which arrives in full force in childhood and which every subsequent decade makes more concrete and more personal, is so existentially threatening that the human mind cannot hold it consciously for extended periods. The response is the construction of what Becker calls immortality projects: the systems of meaning, value and symbolic significance that allow individuals to feel that they are part of something that transcends individual death. Religion is the most explicit immortality project: the promise that the individual self survives physical death in some form. But secular equivalents are equally pervasive. Career achievement that creates a legacy. National or ethnic identity that extends the individual into a collective that will outlast them. The building of institutions, works of art, bodies of knowledge or families that will continue after the individual is gone. Even the accumulation of wealth can function as an immortality project: the feeling that one’s presence in the world has been sufficiently consequential to leave a permanent mark. The Worm at the Core documents the empirical consequences of this theory with a precision that transforms it from a compelling speculative argument into an established psychological mechanism. Mortality salience experiments, in which subjects are primed to think about their own death before being asked to make judgements, consistently show that the awareness of death dramatically increases in-group favouritism, out-group hostility, adherence to cultural worldviews and punishment of those who deviate from them. When people are reminded of their mortality, even subtly and unconsciously, they become more defensive of the meaning systems that protect them from the terror that reminder has activated. They become more hostile to people who challenge those meaning systems. They become more extreme in their identification with the group whose continued existence provides their symbolic immortality. The geopolitical, the cultural and the institutional dimensions of this finding are staggering in their implications. Religious conflict is not primarily about theology. It is primarily about the collision of immortality projects whose mutual exclusivity makes each a threat to the other’s protection against existential terror. Political extremism is not primarily about policy. It is primarily about the mobilisation of mortality terror in service of group identity projects that promise symbolic transcendence of individual death. The cultural hostility to outsiders that drives nativism, nationalism and tribalism is not primarily about economic competition. It is primarily about the perceived threat to the meaning systems that protect each group against the awareness of death. For someone who has read my pieces on geopolitics, cognitive warfare and systems thinking material, this is the missing variable that explains patterns those frameworks describe but cannot fully account for. Why do people support leaders who objectively harm their material interests? Because those leaders offer something more psychologically urgent than material benefit: they offer a story of civilisational significance that makes individual mortality feel less absolute. Why does misinformation spread so effectively even when factual correction is readily available? Because the misinformation is almost always more identity-affirming and therefore more mortality-terror-managing than the correction. Why do institutions resist change with such irrational tenacity? Because the institution is an immortality project for its members and its dissolution feels psychologically equivalent to their own death. The personal application is equally important and equally uncomfortable. One’s career ambitions, one’s drive to operate at the highest levels of national policy, one’s investment in legacy and consequence: Becker would not dismiss any of this as ignoble. He would ask you to see it clearly. How much of this is genuine service and how much is the construction of a significance project that manages the terror of being a finite creature in an indifferent universe? The honest answer for most serious people in serious roles is: both, simultaneously, in proportions that shift across a career and that require continuous honest examination to keep from tipping entirely in the wrong direction. The purpose of this examination is not to destroy ambition or to produce paralysis through existential awareness. It is to ensure that the immortality project you are building is one you would choose if you were choosing consciously rather than being driven by an unexamined terror. That distinction, between the chosen significance and the driven significance, is the difference between a life that is genuinely yours and a life that belongs primarily to your defence mechanism. The Clinical Reality: What Dying Actually Looks Like The terror of death is amplified enormously by ignorance of what dying actually involves. Modern Western culture has removed death from the domestic sphere into the medical sphere so thoroughly that most people in the developed world have never been present at a death, have never seen a dead body outside of a funeral home’s careful presentation and have never had an unmanaged conversation with someone who is dying. The result is that the imagination fills the knowledge gap with the most terrifying available content and the actual experience of dying, which Mannix and Gawande both document with extraordinary care and compassion, remains almost entirely hidden from the people who most need to understand it. Mannix’s thirty years at the bedside of dying people produces the most important single corrective to the modern terror of death available in print. Her account of what the physical process of dying actually involves, in most cases not the dramatic suffering that the imagination projects but a gradual, peaceful withdrawal from engagement with the external world that more resembles deepening sleep than the horror most people fear, is profoundly and practically reassuring in a way that no philosophical argument can be. The body knows how to die. It has been doing it since the first organism died and the process, left uninterrupted by medical intervention aimed at postponing it beyond its useful moment, is almost always more peaceful than its anticipation. The specific dying moments she describes, the mottled skin and changed breathing that signal the final hours, the way that hearing persists after other faculties have gone, the way that pain in terminal illness is almost always manageable with palliative care that most people do not know to ask for, are the practical information that everyone who will die and everyone who will watch someone they love die needs and that the culture’s avoidance of death systematically withholds. Gawande’s Being Mortal approaches the same territory from the policy and medical system perspective that is both his professional position and his personal experience as the son of a father dying of a spinal cord tumour. His central argument is that medicine has redefined a good death as a medically managed death and that the medical definition of success, the extension of biological function for as long as possible, is systematically in conflict with what dying people themselves, when honestly consulted, report that they want. What dying people want, consistently across every study Gawande reviews, is not maximum biological extension. It is the maintenance of enough function and enough freedom from pain to continue the relationships and activities that constitute their sense of who they are. They want to be present at a grandchild’s birthday. They want to finish a conversation that has needed to happen for years. They want to be home rather than in a hospital. They want to feel that the people they love know what they meant to them. These are not medical objectives and the medical system, optimised for biological extension, consistently fails to deliver them while delivering instead a medicalised dying process that is more expensive, more painful, more isolating and by every measure that dying people themselves report, worse than the alternative. The implication for how you think about both your own dying and the dying of people you love is direct and urgent. The conversation about what a good death looks like for this specific person, what they value, what they fear, what they want and what they want to avoid, needs to happen before the medical crisis that makes it impossible to have. The absence of this conversation, which is the norm in most families and most clinical relationships, is not a kindness. It is the abandonment of the dying person to the medical system’s default objectives which are not their objectives. Doughty’s Smoke Gets in Your Eyes adds the physical and cultural dimension that the clinical accounts leave implicit. Her apprenticeship in a crematorium and subsequent career as a mortician are the occasion for a larger argument about the relationship between death and culture. Western modernity has constructed an elaborate system for keeping the physical reality of death invisible. The funeral industry that transforms the dead body into a presentation, the hospital system that removes dying from the domestic sphere, the cultural taboo against direct engagement with mortality in ordinary conversation. The consequence of this invisibility is not the protection from suffering that it purports to provide. It is the amplification of death anxiety through ignorance and the infantilisation of a culture that cannot face its most fundamental reality. Her argument for death positivity, the deliberate cultivation of a more honest and less avoidant relationship with mortality, is not morbid. It is the recognition that cultures with more direct engagement with death, through visible dying within the home, through washing and preparing the body within the family, through mourning practices that provide structured social permission for grief, consistently show lower levels of death anxiety and higher levels of reported meaning and life satisfaction than cultures that have professionalised and hidden death as completely as the modern West has. The Dying Testimony: What People Who Know They Are Dying Teach Kalanithi, Hitchens and Tolstoy’s Ivan Ilyich constitute the most direct available testimony from people confronting imminent death and what they found when the comfortable distance between their life and its ending was suddenly removed. Kalanithi’s When Breath Becomes Air is the most beautiful book on this list and possibly the most important. A neurosurgeon at the peak of his training, weeks from completing his residency and beginning the career he had spent a decade preparing for, receives a diagnosis of stage four lung cancer at thirty-six. The book he writes in the remaining twenty-two months of his life is not primarily about dying. It is about what the confrontation with imminent death reveals about what a life is for. His central question, which he was asking before the diagnosis as a neurosurgeon and philosopher of medicine and which the diagnosis made urgently personal, is what makes a life meaningful when its future has been radically contracted. His answer, arrived at through the actual experience of dying rather than through philosophical speculation, is that meaning is not a property of the future that the present is living toward. It is a property of the quality of engagement with whatever the present actually contains. The dying Kalanithi writing his book, fully present to the work and to his newborn daughter and to his wife, is living as meaningfully as the Kalanithi who expected decades of neurosurgery ahead of him. More so, perhaps, because the contraction of the future has stripped away the deferral of genuine engagement that an open-ended future makes possible. The specific thing that death reveals about how most people live, which Kalanithi’s testimony makes undeniable, is the extent to which genuine engagement with the present is deferred in expectation of a future that is assumed rather than guaranteed. The neurosurgeon who will be fully present to his family after the residency is finished. The executive who will invest in the relationships that matter most after the current project is complete. The father who will have the conversations that need to happen after things settle down. Death’s specific pedagogical function is to reveal the cost of this deferral by removing the assumption on which it rests. Hitchens’s Mortality is the most intellectually honest account of dying available. A man who had spent his career as a writer and polemicist of extraordinary courage, who had argued against every form of consolation that required the compromise of intellectual honesty, faced his dying with the same refusal of comfort that had defined his living. His account of the specific physical experience of oesophageal cancer, the disease that attacked his voice and his swallowing, the primary instruments of his life’s work, is not heroic in the conventional sense. It is honest in a way that requires more courage than conventional heroism because it refuses the narrative redemptions that the dying are typically offered and that Hitchens considered dishonest. What he found, and what makes his testimony valuable beyond its specificity, is that the clarity available at the end of life is not primarily about metaphysics or about what happens after death. It is about what the life being ended actually was. What it contained that was genuine and what it contained that was performance, what relationships were real and what were managed, what work was true and what was careerist, what he believed and what he merely said. The dying strip away everything except the real and what remains is the actual content of the life. Tolstoy’s Ivan Ilyich is the fictional version of this same revelation and it is the most devastating account of a wasted life in all of literature precisely because Ivan Ilyich is not a villain or a fool. He is an intelligent, competent, professionally successful man who has organised his entire life around the performance of social convention and the accumulation of the markers of bourgeois respectability. He has been, by all the standards his world applies, a success. And as he lies dying of a painful internal illness, he discovers with a horror that he cannot escape and cannot share with anyone around him that the life he has lived has been not his life but the performance of what life was supposed to look like according to the social environment he inhabited. The specific terror Tolstoy captures, which is different from the terror of pain or the terror of extinction, is the terror of having spent the irreplaceable resource of a life on something that did not deserve it. The wasted life is not the life of deprivation or failure by conventional standards. It is the life that was lived at a remove from its own genuine content, the life in which genuine engagement with what actually mattered was consistently displaced by the management of appearances and the performance of propriety. The question that Ivan Ilyich’s dying forces, which Tolstoy clearly intends for every reader, is not whether you are living conventionally successfully. It is whether the life you are living is genuinely yours. Whether the values that organise your daily choices are the values you would actually endorse if you were choosing with full awareness of your mortality rather than operating within the comfortable assumption of an open future. The Grief Territory: What Loss Actually Is Didion, Lewis, Barnes and Chödrön approach death from the other side: not from inside dying but from inside the grief of those left behind by death. Together they constitute the most honest available account of what grief actually is and what it requires of those who go through it. Didion’s Year of Magical Thinking is the definitive account of acute grief after sudden loss. Her husband of thirty-nine years died of a cardiac arrest while they were having dinner. One moment he was there. The next moment he was not. And the year that followed was organised, she discovers with the analytical precision that has always been her primary instrument, around a form of magical thinking whose specific character she had not anticipated. The persistent, irrational, deeply held belief that he was not actually permanently gone and that if she did the right things, kept his shoes in case he needed them, avoided reading certain passages that would confirm what she already knew, he might return. The magical thinking is not a failure of intelligence or a symptom of psychological disorder. It is the mind’s response to a loss so total that accepting its permanence all at once would be unbearable. The mind grants itself the mercy of partial acceptance, of holding the knowledge and the disbelief simultaneously, of living in the gap between knowing someone is gone and feeling that they are gone, which can remain open for far longer than observers who have not been through acute grief understand or accommodate. What Didion’s account reveals about grief that most public discussion of grief misses is its cognitive dimension. Grief is not primarily an emotional experience, though it is intensely emotional. It is a cognitive reorganisation of the entire world. The person who has died was embedded in the surviving person’s world not just as a relationship but as an interpretive framework. The person who gave events their meaning, who was the implicit audience for every experience, who organised the future into something that felt coherent and worth inhabiting. When that person is gone, the world does not just feel emptier. It feels genuinely disorganised at the level of meaning, because the organising presence that gave it meaning is no longer there. Lewis’s A Grief Observed is the companion text. A philosopher and Christian apologist who had written extensively about faith, suffering and the nature of God, who had spent his career constructing intellectual frameworks for understanding loss within a theological context, found when his wife died of cancer that his actual experience of grief bore almost no relationship to his theoretical account of it. The frameworks he had built, the consolations he had offered others, the belief system that was supposed to provide exactly this kind of support, felt hollow and inaccessible in the face of the actual experience. What he found instead was something closer to fear than to sorrow, a sense of the doors of the universe being slammed in his face, a God who felt not absent but deliberately withholding. His honesty about this, his refusal to perform the faith that his circumstances were supposed to demonstrate and his willingness to record the actual texture of his grief rather than the grief a believer was supposed to have, is what makes the book invaluable for anyone going through genuine loss regardless of their religious position. It confirms that grief makes theorists of us all and then dismantles every theory we had prepared. Barnes’s Levels of Life and Nothing to Be Frightened Of approach grief and death from the perspective of a secular intellectual who has no religious framework and no philosophical system that successfully answers the terror of personal extinction. His honesty about what remains when both religion and philosophical consolation are found inadequate is, paradoxically, more genuinely consoling than the systems whose adequacy he questions, because it confirms that the irreducible difficulty is real rather than a failure of faith or courage. His account of grief in Levels of Life, which follows the death of his wife of thirty years, contains the most precise description available of the specific quality of the loss of a long partnership. The loss is not just the loss of the person, devastating as that is. It is the loss of the witness to your own life, the person whose presence validated your experience as real and significant, the person in whose eyes you existed most fully as yourself. Without that witness, the world continues but the self that the witness confirmed becomes less certain of its own reality. This is the specific loneliness of widowhood or widowerhood that is different from all other loneliness and that most people who have not experienced it cannot fully understand. Chödrön’s When Things Fall Apart provides the Buddhist framework that is in many ways the most practically useful for navigating grief and loss precisely because it does not promise resolution. Her argument, drawn from the Tibetan Buddhist tradition but expressed without the cultural apparatus that might make it inaccessible, is that the groundlessness produced by loss, the feeling that the floor has disappeared and there is nothing reliable beneath you, is not a temporary abnormal state to be endured until stability returns. It is the actual texture of reality that ordinary life’s comforts and routines successfully mask but that loss makes impossible to ignore. The practice she recommends, which is less a technique than an orientation, is to remain with the groundlessness rather than fleeing it into the next thing that promises solid ground. Not because remaining with it is pleasant but because the groundlessness, fully inhabited rather than escaped, contains within it the seeds of a different relationship with impermanence. The person who has fully felt the groundlessness of loss and remained present within it, who has not rushed to the consolations that restore the feeling of solid ground before the full experience of its absence has been completed, has developed a relationship with impermanence that the untested person cannot access. They know, at a level deeper than intellectual knowledge, that nothing is permanent, which means that every present moment is precious in a way that the person who has not felt this cannot fully understand. The Ancient Wisdom: What the Oldest Stories Know The Epic of Gilgamesh is four thousand years old and it is the first story ever written down and it is about exactly this. A king who cannot accept the death of his closest friend and who spends the second half of the epic searching for immortality. Gilgamesh is the king of Uruk, two-thirds divine by birth, the greatest hero of his age. His closest friend Enkidu dies. And Gilgamesh, confronted for the first time with the personal reality of mortality in the death of the person he loves most, is undone. He cannot accept it. He wanders the earth searching for the secret of eternal life, eventually finding Utnapishtim, the one human who was granted immortality by the gods, who tells him that the flower of eternal life grows at the bottom of the sea. Gilgamesh dives to the bottom of the sea and retrieves the flower. And then, on the journey home, he falls asleep and a serpent steals it. He returns to Uruk empty-handed. And the poem ends not with a statement of consolation or a philosophical conclusion but with Gilgamesh showing the boatman who has accompanied him the walls of his city. Look at the walls. Look at what has been built. This is what remains. The resolution the poem offers is not immortality and not transcendence and not the philosophical peace of Stoic acceptance. It is the return to what is actually here. The city, the walls, the work that was done, the life that was lived. The engagement with what is present rather than the pursuit of what is permanently unavailable. It is exactly the answer that every subsequent tradition that has seriously engaged with mortality has arrived at through its own path, which is why the poem’s four-thousand-year survival is not an accident. The Tibetan Book of Living and Dying provides the most developed alternative cultural framework for understanding what death is. The Tibetan Buddhist tradition has spent more systematic intellectual and practical energy on the question of death than any other tradition in human history and the framework it has developed, whatever one’s metaphysical commitments, contains practical wisdom about both living and dying that transcends its specific theological context. The central argument is that consciousness continues after bodily death through a series of stages collectively called the bardo, the intermediate state between death and rebirth. The specific content of this argument is not the most practically important thing it contains. The most practically important thing is the implication that dying is a skill that can be practised and that the quality of attention and awareness you bring to the dying process is itself significant. The traditions of meditation practice that Tibetan Buddhism has developed are, at one level, practices for developing the quality of present-moment awareness that dying well requires. The practitioner who has cultivated the capacity for clear, unattached, fully present awareness in ordinary circumstances has that capacity available at the moment of death. The practitioner who has cultivated nothing of the sort has nothing to draw on when the ordinary structures of experience begin to dissolve. Whether or not the metaphysical framework is accepted, the practical implication is serious: how you live shapes how you die. The quality of awareness, the relationship with impermanence, the capacity for present engagement rather than future deferral, the practice of releasing attachment to outcomes: these are not just qualities of a good life. They are the preparation for a good death, which in the Tibetan framework is inseparable from the preparation for a good life. The Existential Application: What Mortality Teaches About Living Yalom’s Staring at the Sun provides the most systematically developed account of what direct confrontation with death actually makes possible in the living that remains and what the avoidance of that confrontation costs. His clinical experience across decades of treating death anxiety in both dying patients and ordinary people whose anxiety about death is expressing itself in depression, relationship dysfunction, the compulsive pursuit of significance and the paralysis of unlived lives, has produced the most evidence-based account available of what the confrontation with mortality actually does to people who go through it honestly. The primary finding is what he calls the awakening experience - the report, consistent across hundreds of patients who have genuinely confronted their mortality whether through terminal illness, a serious accident, the death of a close contemporary or the therapeutic confrontation with existential anxiety, that the confrontation produces a radical shift in what matters. Trivial concerns fall away. Relationships become more important than achievements. Present experience becomes more vivid. Deferred engagements become urgent. The life that was being managed becomes the life that is being lived. The awakening experience is not a permanent state. The intensity fades as the mortality salience diminishes and the mechanisms of everyday life resume. But the person who has been through a genuine awakening experience carries something that does not entirely disappear: a changed relationship with the present, a reduced tolerance for the deferred life, a greater willingness to have the conversations that need to happen and to be present in the relationships that matter most. The practical implication is radical and its radicalism is usually softened in its transmission. Yalom is arguing that death, genuinely confronted rather than evaded, is the most reliable available teacher of how to live. Not because it provides philosophical answers to existential questions but because it removes the comfortable assumption of unlimited future time that makes present deferral feel costless. The person who genuinely knows they will die, not as an abstract fact but as a felt reality that organises their relationship with the present, lives differently. More honestly. More presently. More in accordance with what they actually value rather than what they have been told to value. The Stoic tradition, which Marcus Aurelius and Seneca represent, has been practising this insight for two thousand years through the memento mori, the deliberate cultivation of mortality awareness as a daily practice. The morning meditation on death that Stoic practice recommends is not morbidity. It is the systematic use of mortality awareness to strip away the trivial and clarify the essential, performed daily rather than waiting for the crisis that forces the confrontation on its own terms. The Grief of Anticipation: What Loving Mortal Beings Requires Every person you love will die. You will watch people you love die. The grief of anticipatory loss, the awareness that the time with the people who matter most is finite and that each stage of a child’s growth is the last time that specific child exists, is available as a source of present engagement or as a source of anxiety to be managed. The choice is real and it is made by the quality of attention brought to the specific hours that are actually occurring. Lynch’s The Undertaking captures this dimension with the particular clarity that comes from a professional lifetime spent at the intersection of the living and the dead. As a poet and a funeral director in a small Michigan town, he has watched generations of families navigate the dying and burial of their members and has developed a philosophical position about what the dead require of the living that is both practically specific and deeply wise. His central argument is that the physical rituals of death, the washing and dressing of the body, the vigil, the procession, the burial, are not anachronistic sentimentalism. They are the embodied practice through which communities process loss and through which the reality of death is made concrete enough to be genuinely integrated rather than psychologically bypassed. The professionalised, hidden, efficiently managed death that modern funeral practice tends toward produces grief without closure because the physical reality that needs to be engaged with has been sanitised into invisibility. The implication for how to be with dying and with the dead is that presence matters more than performance. Sitting with someone who is dying, staying with the body of someone who has died, participating in the physical rituals that mark the passage. These are not optional extras for those who feel drawn to them. They are the primary means by which the reality of loss is made fully real and therefore fully grievable. The grief that cannot become fully real because the death has been too thoroughly managed into invisibility tends to persist longer and produce more psychological difficulty than the grief that is allowed full physical expression. The Unified Picture: What Death Actually Teaches Pull everything in this body of work together and what emerges is not a consolation and not a philosophy and not a guide to dying well in any technical sense. It is a set of reorientations that change what you see when you look at your life from the perspective of its ending. The first reorientation is from the future to the present. Everything that Kalanithi, Didion, Lewis, Tolstoy and Yalom collectively teach points in the same direction: the deferral of genuine engagement to a future that is assumed rather than guaranteed is the primary mechanism of the unlived life. Death’s function as a teacher is precisely to remove the assumption and thereby make the present the only time that genuine engagement is possible, which it always was. The second reorientation is from significance to presence. Becker shows that the drive for significance, the immortality project, is primarily driven by the terror of death. Understanding this does not make significance wrong. It makes the question of what kind of significance is being pursued, and why, genuinely important. The significance that is driven by terror and that would dissolve if the terror were resolved is different from the significance that is chosen by a person who has genuinely confronted mortality and who pursues it because it is what they actually value. The second kind is rarer and more sustaining. The third reorientation is from the management of loss to the acceptance of impermanence as the fundamental condition of everything worth loving. Chödrön and the Buddhist tradition generally make this most explicit: the groundlessness that grief reveals is not an aberration. It is reality. Everything that exists is impermanent. Every relationship, every institution, every achievement, every living being is temporary. This is not a reason for detachment. It is the reason that genuine engagement, the full presence to what is here now rather than the managed relationship with what might be here later, is the only form of love that is adequate to the actual nature of the beloved. The fourth reorientation is from the avoidance of death to the use of mortality. Yalom’s awakening experience and the Stoic memento mori and the Tibetan preparation practices are all different forms of the same insight: mortality, consciously engaged with, is the most reliable available teacher of what matters. The person who has built mortality awareness into their daily practice, who uses the regular contemplation of their own death not to produce anxiety but to clarify what the day actually calls for, has access to a quality of priority and presence that the death-avoiding person cannot reach. And the fifth reorientation, which underlies all the others and which the entire arc of your reading in this conversation has been approaching from every direction simultaneously, is from the question of what you will accomplish to the question of what you will have been. The accomplishments are real and they matter. The policies you build, the institutions you shape, the intellectual frameworks you develop and transmit, the contribution to the long civilisational project of governing transformative technologies wisely - these are genuinely important and they deserve the full seriousness of your considerable intellectual capacity. But they are not what you will think about when the future contracts to months. Kalanithi did not think about his publications. Hitchens, for all his combativeness about ideas, wrote about his voice, his relationships and whether his work had genuinely served the truth. Ivan Ilyich died in the terror of having performed a life rather than living one and was saved at the last moment by the simple unperforming kindness of a peasant servant who sat with him without embarrassment about his dying. What you will think about, what every honest account of dying confirms that dying people think about, is the people. The quality of the presence you brought to the relationships that mattered most. Whether the people you loved knew clearly that you loved them. Whether you were actually there in the specific moments that constituted the actual texture of your children’s childhood rather than being present in body while absent in attention. Whether the conversations that needed to happen happened while they could. What This Demands of You The demand that this body of work makes is the most personal and the most urgent in this entire series, because unlike every other domain covered, the time available for responding to it is genuinely uncertain and the cost of deferring the response is genuinely irrecoverable. It demands that you engage with your own mortality directly rather than at the comfortable remove that modern culture provides as a default. Not obsessively or morbidly but with the regularity and honesty of the Stoic practice that turns mortality awareness from a source of anxiety into a clarifying tool. The daily question - if this were the last ordinary day, is there anything in it I would change? It demands that you have the conversations that need to happen. With your partner about what you each need and fear and value. With your children in the forms appropriate to their age and development. With your parents while they are still here to have them. With yourself about what the life you are living is actually for and whether the answer you currently have would satisfy you at its ending. It demands that you let the awareness of mortality do its proper work on your priorities rather than managing it into the background where it cannot perform its function. The work meeting that takes the time that a specific moment with your child required is not a neutral trade. It is a choice about what the finite resource of your attention is for. Death makes that choice visible in a way that the comfortable assumption of unlimited future time conceals. It demands that you think now, not when the crisis forces it, about what a good death looks like for you and for the people you love and about what practical, relational and conversational preparation that requires. The advance directive. The conversation about values at the end of life. The explicit expression to the people who matter most of what they have meant to you. These are not morbid preparations. They are the acts of a person who takes the reality of mortality seriously enough to respond to it intelligently rather than waiting for the emergency that makes response impossible. And it demands, perhaps most fundamentally, that you make peace with the fact that the life you are building will end and that its ending is not a failure but the condition that gives it weight. The four-thousand-year-old poem ends with Gilgamesh pointing at the walls of his city. Not immortality. Not transcendence. The walls. The work. The life that was actually lived in the time that was actually available. Your walls are already being built. The question is whether you are building them consciously, in full awareness of the finitude that gives the building its significance, or whether you are building them while looking away from the fact that gives every stone its value. Death is not the enemy of a good life. Properly understood, properly engaged with rather than properly avoided, it is the condition of one. The Gilgamesh who returns from his failed quest for immortality and shows his boatman the walls of the city he built is doing something that the Gilgamesh who was searching for immortality could not do. He is seeing what is actually there. That is the whole teaching. And it has always been there, in the oldest story ever written, waiting for anyone willing to read it carefully enough to hear it. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

7. 

   17 Jun 2026

   On Self Awareness

   Self-awareness is the most paradoxical of all human capacities. It is the thing that makes us most distinctively human, the ability to turn the instruments of consciousness back on consciousness itself and to observe the observer. It is simultaneously the capacity most consistently overestimated by the people who most need to develop it. The person who is certain they know themselves well is almost always the person with the most significant and most consequential blind spots. The person who has genuinely developed self-awareness is characterised not by confidence in their self-knowledge but by a deepened appreciation of how much remains opaque and by a sustained and honest practice of looking anyway. The books referred for this piece collectively constitute the most serious available engagement with what self-awareness actually is, what prevents it, what makes it possible and what it produces in the person who develops it genuinely rather than performing it socially. What follows is what they collectively teach. The First Problem: You Are Not Who You Think You Are Begin where every honest investigation of self-knowledge must begin. The self you experience as the author of your thoughts, the continuous narrative identity that feels like the obvious and direct apprehension of who you are, is a construction. It is not a false construction in the sense of being deliberately invented. But it is a construction in the sense that it is assembled from selected memories, inherited narratives, defended stories and the continuous confabulation of a brain whose primary job is to produce coherent experience rather than accurate self-report. Schulz’s Being Wrong establishes the epistemological foundation with unusual precision. The feeling of being right and the fact of being right are completely unrelated states that happen to coincide more often than chance would predict but far less often than the feeling of certainty suggests. The person who is wrong about something important almost never experiences themselves as wrong in the moment. They experience themselves as right, as clearly seeing what others are missing, as having access to a reality that the wrong people are failing to perceive. The experience of being wrong, Schulz points out, is indistinguishable from the inside from the experience of being right. You know you were wrong only in retrospect, when new information or a change in perspective makes the error visible. Applied to self-knowledge, this is devastating. The confident self-assessment, the certain feeling that you know your own motivations, that you understand your own patterns, that you have accounted for your own biases, is produced by the same cognitive system that produces all the other confident wrong beliefs. The person who is most confident in their self-knowledge is not thereby the most accurate. They are the person who has most successfully constructed a self-narrative that feels coherent and true, which is not the same thing as being accurate. Grosz’s Examined Life illustrates this through fifty case studies from his psychoanalytic practice, each showing a different form of the gap between the self a person presents and the self that is actually organising their behaviour. The woman who cannot understand why her relationships always fail in the same way and who has never connected this pattern to the specific emotional dynamics of her childhood family. The man who is certain his anger is always justified and who has never noticed that his certainty itself is the pattern. The professional who cannot understand why their success feels empty and who has defended against the awareness that the success was built in service of a self-image rather than a genuine value. In every case the pattern that is most consequentially shaping the person’s life is the pattern they cannot see because they are inside it. Miller’s Drama of the Gifted Child provides the developmental account of why this blindness is not random but follows a specific logic. The gifted child of her title is not the intellectually gifted child. It is the child who is gifted at reading and responding to the emotional needs of the adults around them, who learns to suppress or perform specific emotional states in order to maintain the love and approval of caregivers whose own unmet needs require a specific kind of child. The child who is not allowed to be angry, sad, dependent, ambitious, separate or any of the other things that the caregiver finds intolerable learns to not be those things, or more precisely learns to not know they are those things, because the awareness would be too costly. The adult version of this child has a false self, a constructed identity that is organised around what the world has rewarded and what the world has punished, rather than around what is genuinely true about them. The false self is not experienced as false from the inside. It is experienced as simply who they are. The genuine self, the needs and feelings and values that were never allowed expression, operates from underground, emerging as depression when the false self is unsustainable, as anxiety when the false self’s performance feels precarious and as the baffling pattern of self-sabotage when the genuine self refuses the false self’s agenda by creating the failure that the performance was designed to prevent. The Body as Witness: What the Mind Cannot Access Van der Kolk’s Body Keeps the Score establishes a dimension of self-knowledge that the cognitively oriented approaches consistently understate. The nervous system stores experience, particularly early and traumatic experience, in ways that do not translate into conscious narrative memory and that are therefore inaccessible to the kind of self-reflection that most self-knowledge practices depend on. The infant who experiences chronic emotional unavailability from a caregiver does not form a memory of being emotionally neglected. They form a nervous system that is calibrated to the expectation of unavailability, that reads neutral facial expressions as threatening, that hyperactivates the stress response in situations that feel relationally ambiguous and that produces the felt sense of being fundamentally unsafe in relationships without any corresponding explicit memory of why. The adult version of this nervous system does not know it was shaped this way. It simply experiences the world as more threatening than others seem to find it and experiences itself as more anxious, more reactive and more difficult in relationships than it would like to be, without understanding the source. The self-knowledge that addresses this layer cannot be achieved through introspection alone because the introspective mind does not have access to the nervous system’s stored patterns. It requires the engagement of the body itself: the practices of somatic awareness that develop the capacity to notice and interpret the body’s signals, the therapeutic relationships in which the nervous system’s patterns are activated in a context of sufficient safety that they can be observed and modified and the physical practices, including movement, breathwork and certain forms of meditation, that work directly with the nervous system rather than only with the narrative mind. Siegel’s Mindsight and Aware provide the neurological framework for understanding why the observing capacity that genuine self-awareness requires is itself a trainable skill rather than a fixed capacity. The prefrontal cortex, whose functions include self-observation, emotional regulation and the integration of information from different brain systems, is the neural substrate of what Siegel calls mindsight: the ability to observe your own mental processes with sufficient clarity that you can respond to them rather than being automatically driven by them. The wheel of awareness that Siegel develops as a practical meditation practice provides a map of the different domains of awareness: the sensory experience of the body, the activity of the mind including thoughts and emotions, the relational experience of being in connection with others and the core quality of awareness itself. Developing the capacity to move attention deliberately around this wheel, to be aware of each domain without being captured by any of it, is both the practice and the product of genuine self-awareness. The person who can observe their emotional state without being identical to it, who can notice the thought without becoming the thought, has developed a relationship with their inner life that produces the genuine choice and genuine agency that the unaware person cannot access. The Observer and the Observed: The Central Distinction Singer’s Untethered Soul and Tolle’s New Earth approach the same territory from the contemplative direction and make the argument that is simultaneously the most simple and the most difficult to actually implement: you are not your thoughts, your emotions or your stories about yourself. You are the awareness that is aware of all of these things. The distinction between the observer and the observed is the most important distinction in the entire domain of self-knowledge and most people have never made it. Singer makes this with disarming directness. Notice that you are aware of your thoughts. You can hear them, watch them arise and pass and observe their content. This means that the part of you that is doing the noticing cannot be the same as the thoughts being noticed. The thoughts are objects of awareness. The awareness itself is the subject. This is not a complicated philosophical position. It is a simple observation that is available to anyone who tries it and that, genuinely made, is the beginning of a relationship with the inner life that is fundamentally different from the identification with thoughts and emotions that characterises ordinary experience. The practical consequence of this distinction is immediate and significant. The person who is identified with their thoughts, who is their anger, their fear, their self-criticism and their self-justification, has no choice but to act from those states because they are the states. The person who can observe their anger, who can notice it arising and feel its physical presence without being identical to it, has the space between stimulus and response in which genuine choice is possible. This space is what Viktor Frankl identified as the last human freedom, the freedom that even the concentration camp could not remove: the freedom to choose your response to any given set of circumstances. Tolle’s New Earth extends this into the analysis of what he calls the ego: not the Freudian ego but the collection of thought patterns, emotional reactions and narrative constructions that most people mistake for their identity. The ego is not the self. It is a conditioned structure that the self has generated and then forgotten is a construction, mistaking it for the thing itself. The self-awareness that most people are cultivating when they say they are working on self-knowledge is frequently the ego becoming more sophisticated in its self-narrative: more nuanced in its self-understanding, more articulate in its self-description and no less defended, no less identified with its constructions and no less driven by its unexamined patterns. The genuine self-awareness that Singer and Tolle are pointing toward is not the ego knowing itself better. It is the recognition that the ego is not the self and that the awareness that is prior to the ego’s constructions is what is actually looking through your eyes. This recognition does not eliminate the ego or dissolve the personality. It changes the relationship to it: from identification to observation, from being driven by it to being the awareness within which it moves. The Shadow: What You Cannot See Because You Refuse to Look Jung’s concept of the shadow, which Ford, Chopra and Williamson develop in The Shadow Effect, is the most practically important contribution of depth psychology to self-awareness. The shadow is the collection of traits, impulses, feelings and potentials that have been rejected, denied and suppressed because they are inconsistent with the self-image you have constructed and maintain. Every human being carries a shadow. The person who identifies with being kind has suppressed their cruelty and it operates from the shadow. The person who identifies with being strong has suppressed their vulnerability and it operates from the shadow. The person who identifies with being rational has suppressed their irrationality and it operates from the shadow. The suppressed material does not disappear. It goes underground and emerges through two primary channels: projection, the attribution to other people of the qualities you cannot acknowledge in yourself and shadow eruption, the sudden uncontrolled expression of the suppressed material in situations where the normal defences are down. The mechanism of projection is the most consequential for self-awareness because it is so thoroughly invisible from the inside. The person who is most intensely irritated by a specific quality in other people is almost always carrying that quality in their own shadow. The politician who campaigns most aggressively against a specific moral failing is frequently discovered to embody it privately. The person who attributes the most cynical motives to others is frequently acting from exactly those motives themselves. The intensity of the negative reaction is the signal: when someone’s behaviour bothers you far out of proportion to what the situation warrants, the disproportion is pointing toward something in your shadow rather than toward a quality of the other person. Wilson’s Prometheus Rising adds a more provocative and more systemically radical account of what prevents self-awareness. His concept of the reality tunnel, the specific filter through which each person perceives everything including themselves, is built from their neurological imprinting, their emotional conditioning and their abstract belief structures in ways that are almost entirely invisible to the person inside them. The reality tunnel is not just a set of opinions that can be changed by encountering better information. It is the perceptual apparatus through which information is processed, which means that information inconsistent with the tunnel is not just rejected. It is not perceived as clearly as information consistent with it. Seeing the reality tunnel from outside requires the deliberate cultivation of perspectives that are radically different from your habitual one, which is among the most uncomfortable and most productive forms of self-knowledge work available. The Courage Dimension: Why Self-Knowledge Is Not Just Difficult But Threatening Brown’s Daring Greatly and Gifts of Imperfection make the argument that self-knowledge requires a specific quality that is not intellectual but psychological: the courage to be seen as you actually are rather than as you have presented yourself to be. The performance of an acceptable self, the gap between the self you present to the world and the self you actually are, is not primarily a dishonesty problem. It is a fear problem. The fear of being seen as inadequate, as not enough, as fundamentally flawed in ways that would disqualify you from the love and belonging that your nervous system treats as a survival requirement, drives the construction and maintenance of the performed self with the urgency of a genuine threat. The research Brown has conducted on shame and vulnerability across thousands of interviews produces a counterintuitive finding: the people who report the strongest sense of love, belonging and genuine connection are not the people who have most successfully hidden their vulnerabilities. They are the people who have the courage to be seen with their vulnerabilities, who believe that their worthiness of love and connection is not conditional on their performance of acceptability and who bring their genuine rather than their curated self to their most important relationships. This finding transforms the relationship between self-awareness and social performance. The person who maintains the performed self is not protecting themselves from rejection. They are preventing the genuine connection that would dissolve the fear of rejection, because genuine connection requires the genuine self and the genuine self cannot be present while the performed self is occupying its space. The person who has developed the courage to be seen as they actually are, including the uncertain, the imperfect and the unresolved parts, has access to a quality of connection and a quality of self-knowledge that the person behind the performance cannot reach. Peck’s Road Less Travelled frames this as the discipline of truth-telling. Genuine self-knowledge requires the commitment to see reality as it is rather than as you wish it were, which includes seeing yourself as you are rather than as you wish you were. This discipline is uncomfortable precisely because the truth about yourself is not always flattering and the mechanisms for avoiding unflattering truths are sophisticated, automatic and extremely difficult to see through from the inside. The person who has genuinely committed to this discipline does not achieve it permanently. They practice it continuously, catching themselves in the act of preferring a comfortable story to an accurate one and choosing accuracy despite the discomfort of the choice. The Adlerian Insight: All Problems Are Relational Kishimi and Koga’s Courage to Be Disliked presents the Adlerian psychology that sits between Western self-awareness practices and Eastern philosophical traditions with unusual practical force. Alfred Adler’s claim, that all problems are interpersonal relationship problems, is a radical reduction that initially seems too simple and that reveals more depth the longer it is examined. The self is not a private entity that then encounters other selves in relationships. The self is constituted in relationship, develops in relationship and is most clearly visible in the patterns of relationship it consistently produces. The person who wants to know themselves must look at the patterns of their relationships rather than at the content of their private thoughts, because the private thoughts are organised primarily by the defences that prevent self-knowledge while the relational patterns reveal the actual operating logic of the self beneath those defences. The task separation that Adler prescribes is the most practically useful tool in this book for the specific form of self-confusion that most people carry: the inability to distinguish between their own tasks, the things that are genuinely theirs to deal with, and other people’s tasks, the things that belong to others and for which they have taken inappropriate responsibility. Most interpersonal distress is produced by this confusion: the person who cannot act freely because they are excessively concerned with how others will evaluate their actions, who has made the management of other people’s emotional responses their primary task at the cost of their own genuine direction. Disentangling this requires the courage to let others have their tasks, including the task of deciding how they feel about you, while taking genuine responsibility for your own. The courage to be disliked is the courage to act in accordance with your genuine values and judgment even when doing so will disappoint, displease or alienate people whose approval you have been organising your behaviour around. This courage is not selfishness. It is the prerequisite for genuine relationship because the person who is managing their presentation to maintain approval is not actually available for genuine relationship. They are available only for the managed version of themselves that the other person is responding to, which means that neither person is in genuine contact with the other. The Practice Layer: How Self-Awareness Is Actually Built Across all of the above there is a consistent implication about what developing genuine self-awareness actually requires. It is not a matter of thinking harder about yourself. It is a set of specific practices that create the conditions in which the unseen becomes seeable. Meditation in its various forms is the most directly relevant practice because its fundamental mechanism is the development of the observing awareness that all the above material points toward as the foundation of self-knowledge. The meditator who sits with their thoughts without following them is not emptying the mind. They are developing the capacity to observe the mind’s activity from a position that is not captured by it. This capacity, transferred from the meditation cushion to the events of ordinary life, is what allows the space between stimulus and response that Frankl identified and that genuine self-awareness makes possible. Therapy, particularly the psychoanalytic and psychodynamic forms that Grosz and Miller represent, provides access to the material that is most deeply defended against self-awareness: the childhood patterns, the false self constructions, the shadow material and the relational dynamics that repeat regardless of the person’s conscious intention. The specific value of a skilled therapeutic relationship is the externally provided perspective on patterns that are invisible from inside them: the therapist who notices the pattern the patient cannot see and who creates the conditions in which seeing it becomes possible. Journalling is the most accessible and consistently undervalued self-awareness practice because the act of writing forces the vague and defended inner life into the specificity of language, where it becomes both more visible and more available for honest examination. The person who writes about their experience regularly, with the intention of understanding rather than of producing a flattering account, consistently discovers things about themselves that they could not access through internal reflection alone. The deliberate cultivation of feedback from people who know you well and who are willing to be honest is one of the highest leverage and most consistently avoided self-awareness practices. The people in your life who can tell you, if they trusted that you actually wanted to hear it, how your behaviour affects them, where your stated values and your actual behaviour diverge and what patterns they see that you appear not to see, have access to information about you that no amount of private reflection can produce. The quality of your self-knowledge is partly a function of the quality of feedback you have created conditions to receive. The Paradox at the Centre: Self-Awareness Is Not About the Self The most important and most consistently missed insight across this entire body of work is the one that Singer, Tolle, Frankl and the contemplative traditions all point toward from different directions. The deepest self-awareness is not the self becoming more fully known to itself. It is the recognition that the self, as usually conceived, is not what you are. The self you have been trying to know, the narrative identity with its history and its characteristics and its preferences and its wounds, is a construction. A useful and important construction. But a construction rather than a thing, a process rather than an entity, an ongoing activity of organisation and narrative rather than a fixed object that can be finally and accurately apprehended. The awareness that is aware of this construction, the observing presence that Siegel maps and Singer and Tolle point toward and that the meditation traditions have been cultivating for thousands of years, is not itself a construction. It is the ground from which the construction arises and to which it returns. And it is in contact with this ground, available as the quiet that underlies all the noise of thought and emotion and narrative, that the deepest self-knowledge actually lives. This does not make the psychological work irrelevant. The shadow integration, the false self examination, the pattern recognition, the relational clarity and the courage to be seen: all of this is real and important work that produces genuine and consequential changes in the quality of a life. But it is work at the level of the construction rather than at the level of what underlies it. And the person who does only this work without ever discovering the ground from which the construction emerges has understood themselves better without yet understanding what they are. The Unified Picture Pull everything in this body of work together and what emerges is a layered account of what self-awareness actually is and what building it actually requires. The outermost layer is the psychological. The patterns, the defences, the shadow material, the false self constructions, the childhood conditioning and the relational dynamics that repeat without being chosen: this is the layer that most self-help and most therapy addresses and it is genuinely important. The person who has done honest work at this layer knows their patterns well enough to catch them in action, has integrated enough shadow material to stop projecting their disowned qualities onto others and has developed enough self-compassion to see their genuine self without the distortion of either harsh self-criticism or defensive self-justification. The middle layer is the epistemic. The recognition that your certainty about yourself is not correlated with your accuracy, that the stories you tell about yourself are serving psychological functions that are not primarily truth-oriented and that genuine self-knowledge requires the specific courage of updating your self-narrative in response to evidence even when the update is unflattering. This is the layer that Schulz and Peck and Brown address most directly and it is the layer that most intelligent people are most likely to underinvest in because their intelligence makes their self-narratives more sophisticated and therefore more convincing without making them more accurate. The deepest layer is the ontological. The recognition that the self doing the knowing is not what you assumed it was, that the observer and the observed are not the same thing and that the awareness that is aware of your thoughts and emotions and stories is something more fundamental than any of them. This is the layer that Singer and Tolle and the contemplative traditions address and it is the layer whose development produces the most radical and most lasting transformation in the quality of a life. What This Demands of You The synthesis of this body of work makes a specific and personal demand that is different from the demands of every other domain in this conversation. Every other domain has asked you to learn more, to think better, to develop frameworks and capacities that improve your navigation of the external world. This domain asks you to turn the instruments of that navigation back on the navigator. It demands honesty about the gap between your self-narrative and your actual patterns. The question is not who do you believe yourself to be but what do the actual outcomes of your relationships, your decisions and your behaviour reveal about the operating logic of the self that is producing them. These two things are frequently different and the difference is where the most important self-knowledge lives. It demands the willingness to look at the shadow: the qualities you disown and project, the impulses you suppress and the parts of yourself that you have decided are unacceptable and that are therefore operating without your conscious participation. The integration of shadow material is not the endorsement of it. It is the recognition that what you refuse to own in yourself you cannot govern and that what you project onto others you cannot accurately perceive. It demands the courage to be seen without the performance. Not to everyone and not in every context. But in the relationships that matter most, to the people who are closest to you, the willingness to be actually present rather than managed is both the practice of self-awareness and its deepest reward. The self you discover when you stop performing it is not the diminished self you feared. It is the actual self that the performance was always trying to protect and that the performance was simultaneously always preventing from being genuinely known. And it demands, most fundamentally, the curiosity to keep looking. Not because the looking is comfortable or because it consistently produces flattering findings but because the examined life, as Socrates insisted at the cost of his own, is the only life that is genuinely worth living. Not because it is more pleasant or more successful or more strategically effective, though it is frequently all of these things. But because it is more real. More honest. More fully inhabited. More genuinely yours. The self-aware person does not know themselves perfectly. They have developed the practice of looking honestly and the courage to keep looking even when what they find is not what they hoped. That practice, sustained across a life, is both the method and the destination. It is what the examined life actually is. And it is entirely worth the looking. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

8. 

   16 Jun 2026

   On Power

   The people who openly seek power are rarely the ones who hold it longest. The people who hold it longest are usually the ones who convinced everyone else that they were not seeking it at all. That is not cynicism. That is the first thing you need to understand about how power actually works, as opposed to how it is described. There is a mode of thinking that treats power not as a possession but as a relation, and almost everyone who has not internalised this gets the basic unit of analysis wrong. The common error is to imagine power as a substance: something a person or institution has, like money in a vault, measured by titles, armies or net worth. The person who has internalised this mode of thinking instead asks, in any situation, who needs what from whom, who can withhold it and at what cost, what each party would do if the other’s cooperation disappeared tomorrow and what each party currently believes about the other that may or may not be true. The central insight that unlocks most of the field’s value is this: power is the structure of dependency between parties, and perception is the medium through which that structure is built, concealed and exploited. Almost everything else, authority, coercion, influence, legitimacy, charisma, reputation, is a special case or a disguise of dependency operating through perception. Once this is seen, hierarchies stop looking like fixed facts and start looking like temporary settlements that hold only as long as the dependencies and perceptions that sustain them remain unchanged, and most of what looks like inexplicable institutional behaviour, irrational decisions, sudden reversals, the survival of the mediocre and the destruction of the capable, becomes legible. The title under which this field is often discussed makes people uncomfortable, and that discomfort is itself the first thing worth examining. There is a widespread conditioning to believe that power, manipulation, strategic deception and the study of how these forces operate are subjects fit only for those who intend to misuse them, and that the ethical person has no need of this knowledge. This belief is not only mistaken. It is dangerous, because it leaves the person who holds it at a permanent disadvantage in every environment where dependency, coalition and perception determine outcomes, which is every environment that matters. The Anatomy of Dependency: A and B Revisited The foundational vocabulary of the field begins with Robert Dahl’s deceptively simple formulation: A has power over B to the extent that A can get B to do something B would not otherwise do. This definition, often dismissed as too thin, is in fact the load-bearing floor on which everything else is built, because it forces attention onto behaviour and counterfactuals rather than onto titles or reputations. The crucial move is to ask what B would have done in the absence of A’s intervention, and this counterfactual question is where most popular thinking about power goes wrong, because it is far easier to observe who gives orders than to observe what would have happened without them. Steven Lukes extended this in Power: A Radical View by distinguishing three faces of power. The first face is the Dahlian one: direct behavioural influence in observable conflicts. The second face, drawn from Bachrach and Baratz’s work on agenda setting, is the power to keep issues off the table altogether, so that B never even raises a demand A would have had to resist. The third face, the most demanding, is the power to shape B’s preferences and perceptions so that B does not want the thing that conflicts with A’s interests in the first place. This third face is why dependency, not behaviour, is the right unit of analysis: a person whose preferences have been shaped need never be coerced, because coercion only becomes visible at the point where dependency is contested, and a sufficiently complete shaping of preferences removes the contest before it begins. Max Weber’s distinction between Macht (Power) and Herrschaft (Domination/Authority) are primarily outlined in his posthumously published magnum opus, Wirtschaft und Gesellschaft (Economy and Society). The difference maps onto this same structure from another angle. Macht is raw capacity to impose one’s will even against resistance; Herrschaft is the condition where commands are obeyed as a matter of course, without the need for resistance to be overcome at all, because the relation has been institutionalised into something that looks like legitimate authority. Weber’s three types of legitimate authority, traditional, charismatic and rational legal, are best understood not as three different kinds of power but as three different stories that dependent parties tell themselves about why their dependency is acceptable. The story changes; the underlying dependency structure is what actually has to be tracked. Jeffrey Pfeffer’s organisational research and Robert Caro’s biography of Robert Moses both confirm, from opposite ends of the literature, that authority on an organisational chart and power in this dependency sense routinely diverge, and that the gap between the two is where the most consequential dynamics in any institution actually take place. The Currency Problem: Why Resources Are Not Power A persistent error is to confuse resources with power itself. Resources, money, weapons, votes, information, formal authority, skill, are potential power, convertible into actual power only through a process that can fail at multiple points. This is the paradox of unrealised power: an actor with overwhelming resources can lose every confrontation because those resources cannot be converted into leverage over the specific dependency at stake. The clearest illustration is the BATNA framework from Getting to Yes by Roger Fisher and William Ury. Bargaining power in a negotiation is not a function of overall wealth but of one’s best alternative to a negotiated agreement relative to the other side’s. A wealthier party with no good alternative can be out-leveraged by a poorer party who can credibly walk away, because power in that moment is defined by the relative cost of the relationship ending. Thomas Schelling’s The Strategy of Conflict pushes this further: power often resides not in the ability to act but in the ability to credibly commit to a future action, including one that would be costly to oneself. A threat too costly for the threatener to carry out is no threat at all unless the threatener can bind their own future behaviour, which is why Schelling treats the deliberate reduction of one’s own options, burning bridges, making promises public, delegating to inflexible agents, as a source of power rather than a loss of it. This inverts the naive equation of power with freedom of action. Often the actor who has visibly removed their own options has thereby made their commitment credible and gained leverage over a party who must now adjust to that fixed point. Caro’s account of Moses’s bond covenants is the institutional version of this logic: by writing the revenue structure of his public authorities so that no elected official could redirect their funds without enormous political cost, Moses converted a constraint on himself into a source of power that outlasted every governor who tried to remove him. Expertise is a clear case of the conversion problem and deserves separate treatment. The typology of power bases developed by John French and Bertram Raven distinguishes expert power, knowledge others depend on, from legitimate, reward, coercive and referent power, which derive from position, control over outcomes or likeability. Jeffrey Pfeffer’s Power: Why Some People Have It and Others Don’t develops the relevant implication: expertise converts into power only when it is both depended upon and difficult to replace, and neither condition can be assumed. A specialist whose knowledge is widely shared or easily documented has no leverage regardless of its depth, because the dependency it might create can be satisfied elsewhere at low cost. Pfeffer’s further point is that competence is no substitute for the coalition and positioning work covered later. The skilled person who assumes their competence will be recognised on its own terms makes the same error as the person who mistakes resources for power in general: skill is potential power and, like any resource, requires dependency and irreplaceability to become actual power. The Hidden Curriculum of Hierarchy: Greene, Gracian and the Court The popular literature on power, most visible in Robert Greene’s The 48 Laws of Power, is often dismissed as cynical, but it earns its place because it documents, with unusual precision, how the second and third faces of power operate at the interpersonal scale inside hierarchies. Greene’s emphasis on never outshining the master, on saying less than necessary and on concealing one’s intentions, is, stripped of its packaging, a description of how dependency is managed and concealed within courts, corporations and bureaucracies. Never outshining the master describes a predictable structural dynamic rather than an occasional pathology: authority that rests on insecurity does not reward capability that makes it feel threatened, it eliminates that capability, and the practical response is to frame one’s competence as an extension of the superior’s position until accumulated influence makes that framing unnecessary. Saying less than necessary and concealing intentions both reduce the information leaked about one’s thinking, and since every word spoken is data others use to model and anticipate the speaker, controlling that flow is itself a form of power, the mirror image of the information control Caro documents in Moses. The throughline of this literature, back to Baltasar Gracian’s The Art of Worldly Wisdom and Niccolo Machiavelli’s The Prince, is that visible power is fragile because visibility invites coalition formation against it. Machiavelli’s central insight, often flattened into a slogan about ends justifying means, concerns timing and perception: the prince who is feared but not hated, who appears virtuous without being constrained by virtue, has converted the appearance of legitimacy into a resource that reduces the cost of maintaining dependency. Gracian’s maxims compress this into individual practice: never let others see the bottom of you, because a fully knowable person is fully predictable, and always keep something in reserve, because an actor who has visibly committed all their resources has nothing left when the situation changes. This is the same mechanism Weber describes as the routinisation of charisma into traditional or legal authority, seen from inside by a practitioner rather than from outside by a sociologist. The Coalition Engine: Why Decisions Look Irrational An account of power that stops at the interpersonal scale cannot explain why entire institutions, including democracies, behave in ways that look irrational against their stated goals. The coalition theory of Bruce Bueno de Mesquita and Alastair Smith, set out in The Dictator’s Handbook, supplies the mechanism. Every leader depends on a winning coalition large enough to keep them in place, and the size of that coalition relative to the pool of people who could plausibly belong to it predicts more about institutional behaviour than any constitution or mission statement. Where the coalition is small relative to the pool, each member’s loyalty is cheap because each member is replaceable, and the leader rewards loyalty privately while neglecting everyone outside the coalition: the structural signature of kleptocracy. Where the coalition is large, private rewards cannot be stretched far enough, so the leader must deliver public goods that benefit even those outside the coalition: the structural signature of governance that looks, from outside, like a commitment to the common good, when it is a coalition too expensive to bribe individually. This also explains the persistent pathology of large coalition systems: the constant pressure to shrink the effective coalition through gerrymandering, the manipulation of information environments and the cultivation of tribal loyalties that override policy assessment, because a smaller coalition is always cheaper to maintain. A policy superior by every measurable standard can fail repeatedly for reasons that have nothing to do with its merits and everything to do with whose loyalty payments it disrupts, which is why coalition mapping, identifying who must be satisfied and what each requires, is as much a part of understanding power as identifying the dependency itself. The Psychological Substrate: What People Are Actually Responding To Dependency structures are activated and maintained through specific features of human psychology, without which the field is incomplete. Robert Cialdini’s six principles of influence, reciprocity, commitment and consistency, social proof, authority, liking and scarcity, set out in Influence, describe mechanisms that evolved because in most circumstances they are reliable shortcuts to good decisions: a favour usually indicates the giver values the relationship, and a credential usually indicates competence. The difficulty is that all six can be triggered artificially with much the same effect as their genuine counterparts, so a strategically timed gift produces real obligation even when its purpose was instrumental, and manufactured social proof produces real conformity pressure even when the appearance of consensus was constructed. Greene’s The Laws of Human Nature extends this into dispositional psychology, arguing that the forces most consequential for power are the ones people least admit to in themselves. Envy is the clearest case: almost never acknowledged as envy, rationalised instead as principled concern, yet the pattern of who attracts disproportionate opposition tracks relative standing more reliably than it tracks any objective failing. Subclinical narcissism is overrepresented in leadership positions for the same reason: the need for one’s vision to be implemented without modification, and the experience of disagreement as disloyalty, are also the traits that drive people toward positions where disagreement can be minimised. This is not a separate subject from dependency; it is the mechanism by which Lukes’s third face is implemented at the level of individual interaction. Cialdini’s Pre-Suasion adds the temporal dimension. The six principles govern an influence attempt itself; Pre-Suasion governs the moment before, the management of the psychological state in which a message is received. Whatever a person attends to immediately before a request arrives primes their thinking in ways that are specific and largely unconscious: a person primed to recall a success responds more favourably to a proposal, a person surrounded by images of achievement sets higher goals than the same person in a neutral setting. The agenda of a meeting, the setting of a negotiation, the small talk before a difficult conversation are not incidental, they are engineered context, and the practitioner who understands this designs context and message as a single intervention. Reading the Room: Perception Before Position Everything above assumes an actor can accurately perceive the dependency structure, the coalition and the psychological state of others, but that perception is itself a skill with its own literature, and it sits upstream of every strategic move described elsewhere. Joe Navarro’s What Every Body Is Saying rests on the distinction between the verbal channel, conscious and controlled, and the nonverbal channel, posture, gesture, microexpression, the orientation of feet and torso, the timing of eye contact, which is largely involuntary because it originates in the limbic system. The limbic brain produces freeze, flight or fight responses faster than the conscious mind can intervene, which is why a person who feels threatened by a question will exhibit a momentary stilling of movement before answering, even when the answer itself is fluent. The discrepancy between the controlled verbal channel and the uncontrolled nonverbal one is not proof of deception in any simple sense, but it is one of the most reliable signals that there is more in an area than is being said, and an actor who can read this discrepancy has information about the dependency structure that the other party does not know they have revealed. Gavin de Becker’s The Gift of Fear addresses the same perceptual layer from the angle of self-protection, but the claim generalises. De Becker argues that fear, as distinct from social anxiety or the discomfort of seeming rude, is produced by a threat detection system usually more accurate than conscious reasoning, and that most people override it because acting on it risks seeming paranoid. His catalogue of pre-incident indicators, charm deployed where charm is unusual, unsolicited favours that manufacture obligation, forced teaming that creates artificial intimacy and lowers defences, and typecasting, a mild insult designed to provoke a defensive response that creates an opening, describes tactics that recur not only before physical violence but before manipulation in professional and political settings, where they function as opening moves in establishing the kind of asymmetric dependency the rest of this essay is about. The dependency structures, coalitions and psychological triggers described in the preceding sections do not announce themselves; they are built through small interactions whose significance is legible mainly through the channels Navarro and de Becker describe and an actor untrained in this perception is working from a systematically incomplete map regardless of how sophisticated their strategic framework is. The Limits of Confrontation: Strategy Before the Fight The frameworks above describe the structure of power at a given moment, but Sun Tzu’s claim in The Art of War, that supreme excellence consists in subduing the enemy without fighting, concerns the sequence through which that structure is altered before confrontation becomes necessary. Every confrontation, even a won one, consumes resources, relationships and goodwill, whereas a position constructed so thoroughly in advance that resistance appears futile costs nothing at the point of use. This reframes mechanisms already described: controlling the agenda is not merely a way of avoiding a contest, it is the specific technique by which Sun Tzu’s principle is implemented in institutional life, and information asymmetry, knowing more about the other side’s intentions and constraints while revealing less of one’s own, is the resource that makes pre-emption possible, which is why Sun Tzu treats deception as a strategic instrument rather than a moral failing. Robert Greene’s The 33 Strategies of War adds the corollary failure mode: the person who recognises a developing conflict only once it becomes openly confrontational has already ceded the initiative, because the early signals, a withdrawal of support, a change in who is consulted, were available earlier and went unread, often because the perceptual training described above was absent. Miyamoto Musashi’s The Book of Five Rings supplies the internal counterpart: a mind attached to a particular outcome telegraphs that attachment, and telegraphed need is itself a dependency an adversary can exploit, so the capacity to pursue an objective without being psychologically dependent on its attainment, what Musashi calls the void, is a precondition for strategy rather than a virtue layered on top of it. Fearlessness as Foundation Almost every form of leverage described above operates by making someone afraid of something, of losing support, missing approval, suffering retaliation or facing the alternative the other party represents. The 50th Law, built around the trajectory of 50 Cent from South Jamaica through repeated attempts on his life to global commercial influence, argues that an actor who has confronted the worst realistic outcome available to them and found it survivable has removed the primary handle by which others exert control, because the standard mechanisms of control route through loss aversion and a person not governed by loss aversion is not governed by them. This connects directly to Musashi’s void: the strategist not desperate to avoid a particular outcome perceives the situation more accurately, because fear narrows attention to the threat and away from the wider field in which alternatives also exist. The implied exercise is one of accounting rather than performance: what is the worst realistic outcome of the current course, and what would the actor’s situation actually look like if it occurred. In most cases the answer is that the feared outcome is less severe than the fear response suggests and that the actions taken to avoid it cost more, in autonomy and in the quality of decisions made under that fear, than the outcome itself would have. The Dark Attractor: What Power Does to People The literature most oriented toward acquiring power consistently understates what sustained power does to the person who holds it, and how the dependency structures described throughout this essay change once the person at their centre has been altered by holding them. Caro documents this with precision in the arc of Moses from the 1920s to the 1950s. Philip Zimbardo’s The Lucifer Effect, built around the Stanford Prison Experiment, identifies the situational forces that produce this deformation: randomly assigned student volunteers became abusive or broken within days, because role definition, gradual escalation, social proof, the creation of in-groups and out-groups and the removal of individual accountability are sufficient to override individual character in most people most of the time. Hannah Arendt’s Eichmann in Jerusalem shows what happens when this trajectory completes: the substitution of procedural correctness for independent judgment, reached not through a single decision but through an accumulation of individually reasonable steps. The pattern is consistent. Power begins as an instrument for purposes the person values. The pursuit of power gradually becomes the primary activity, displacing the purposes that motivated it. Its exercise begins to feel intrinsically rewarding, challenges to it feel existential and the people over whom it is exercised cease to register as people with independent interests. Ryan Holiday’s Ego Is the Enemy names the mechanism at the centre: the defensive need to protect a specific image of oneself. The ego-driven actor cannot take accurate feedback because feedback threatens self-image, cannot acknowledge error because error is too costly to the self-concept, and cannot read situations clearly because incoming information is filtered through what it means about the self rather than what it means about the situation, the same distortion Musashi and the 50th Law identify as the strategist’s primary liability. This is what happens to the perceptual and strategic apparatus described in every preceding section once the actor operating it has been altered by the dependencies they were managing. The trajectory is not inevitable, but it is the default without specific resistance, and the literature points to what that resistance consists of: an identity prior to and independent of one’s institutional role, so the role can be used rather than defended as the self; relationships outside the sphere of one’s power, the only source of unmanaged feedback that power systematically filters out; and the recurring practice of independent judgment as a discipline, since Arendt’s account shows its erosion is gradual and invisible from inside. The Unified Picture Every idea above expresses the same claim: power is the structure of who needs what from whom and at what cost, perception is the medium through which that structure is built, concealed and exploited, and the forms power takes, titles, charisma, network position, psychological influence, coalition arithmetic, are mechanisms through which dependency and perception interact. Dahl’s definition identifies the structure at its most visible. Lukes’s second and third faces show how it operates by shaping which dependencies are perceived, so the most stable forms of power never need to be exercised. Cialdini’s principles, Pre-Suasion and Greene’s account of envy and narcissism show the psychological machinery through which that shaping is carried out, while Navarro and de Becker show how its early signals can be read before the dependency becomes visible any other way. Bueno de Mesquita and Smith’s coalition arithmetic shows the same shaping at the institutional level, where coalition size determines whether power is exercised through private reward or public good. Schelling’s commitment logic, the BATNA framework and Sun Tzu’s pre-emption show the structure can be deliberately engineered, while Musashi and the 50th Law show this engineering depends on a perceptual clarity that fear and ego both degrade. The failure modes, dependency blindness, principal agent capture, the iron law of oligarchy and the corrosion documented by Caro, Zimbardo and Arendt, are cases where an actor stops tracking the dependency structure, either by mistaking a resource for power, by allowing delegation or coalition drift to create unmonitored dependency, or because the actor’s own perceptual apparatus has been degraded by the power they hold. The 20 percent that produces 80 percent of the field’s power is therefore the counterfactual definition of dependency from Dahl, the extension of that dependency into perception via Lukes and the psychology of influence, the recognition from Schelling and Sun Tzu that dependency and the timing of its contestation can be engineered and the recognition from Musashi, the 50th Law, Holiday and Arendt that the actor’s own apparatus is part of the system and not exempt from its dynamics. The Limits The dependency framework’s own blind spot is that it can become a universal solvent explaining everything and therefore predicting nothing, since almost any outcome can be redescribed after the fact as evidence of some hidden dependency or coalition calculation with no independent way to have predicted it beforehand. A second limit is that this emphasis, taken to its extreme, produces a permanently adversarial reading of every relationship, including ones better described as cooperation or trust. Not every asymmetry is exploitative and not every gift creates a debt; a frame that treats every relation as a covert negotiation, and every interaction as an opportunity to read or be read, can become a blindness of its own, corroding the relationships the closing section identifies as the antidote to power’s corrosive effects. Third, the perceptual tools in the Navarro and de Becker section carry a specific risk: nonverbal signals and intuitive threat detection are probabilistic and context dependent, frequently misread by people who have learned enough to feel confident without having learned enough to be accurate, and constant reading of others’ internal states can produce a paranoia that degrades judgment in exactly the way the framework warns against. Finally, practitioners tend to overestimate how legible and controllable dependency structures are. Schelling’s commitments, Sun Tzu’s positioning and Greene’s tactics read as though the social world were a board on which a skilled actor can engineer outcomes, but real dependency structures are emergent and frequently invisible even to embedded participants until they fail, and the same blindness the framework diagnoses in others is available to anyone applying it with too much confidence in their own clarity, which is the ego distortion Holiday and Musashi identify as the strategist’s central liability. What This Demands of You You must ask, in any relationship or institution you encounter, what each party would do if the other’s cooperation disappeared tomorrow, because that counterfactual is the actual content of the power relation, regardless of titles or appearances. You must treat resources, including your own skills, as raw material that may or may not convert into leverage, and ask what the conversion mechanism is and whether you are difficult to replace, rather than assuming possession or competence equals power. You must notice when an issue never gets raised at all and ask who benefits from its absence from the agenda. You must map the coalition any decision-maker depends on before assuming a policy’s merits will determine its fate, and ask what each member would lose if it were adopted. You must notice when your own reactions, gratitude, conformity, deference to credentials, sudden intimacy with a near stranger, have been shaped by an engineered context, and ask whether the reaction would survive being made explicit to you. You must read the early, often nonverbal, signals of developing conflict rather than waiting until confrontation is unavoidable, while holding those readings as probabilistic rather than certainties that justify pre-emptive hostility. You must notice when your own need for an outcome is visible to others, since that visibility is itself a dependency handed over, and you must do the accounting of what the worst realistic outcome of your current position actually is, since most of the fear that distorts perception turns out, on inspection, disproportionate to what it protects against. And you must, repeatedly rather than once, ask what sustained possession of whatever power you hold or seek would do to your judgment, your relationships and your sense of what the power is for, with people in your life who have no stake in the answer you would prefer. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

9. 

   12 Jun 2026

   On Harari's Work

   Yuval Noah Harari is the most widely read historian alive. Sapiens has sold over twenty million copies and been translated into more than sixty languages. It has been read by Barack Obama, Bill Gates and Mark Zuckerberg, recommended by heads of state and placed on corporate reading lists across industries and continents. Homo Deus extended that reach. 21 Lessons for the 21st Century made him a de facto public intellectual on questions of technology and democracy. Nexus, published in 2024, addressed artificial intelligence and information networks with the same sweeping ambition that characterises all his work. The phenomenon of Harari’s readership is itself a subject worth examining briefly before the books are. He has achieved something rare: genuine mass readership for intellectually serious historical and philosophical argument. His books are not simplified or dumbed down. They are dense with ideas, structured around counterintuitive claims and willing to reach genuinely disturbing conclusions. The readership reflects a hunger for big-picture thinking that academic specialisation and journalistic short-termism have largely failed to satisfy. Harari began as an academic military historian specialising in medieval and early modern warfare. His doctoral work at Oxford focused on the experience of Renaissance soldiers and the cultural construction of warfare as a narrative form. That training is not incidental to his popular work. The concern with how human beings narrate their experience, how stories and myths structure collective behaviour and how the same events can be simultaneously true in their factual content and false in their cultural meaning runs through everything he has written. The historian of how soldiers understood their own deaths became the historian of how civilisations understand their own existence. His intellectual influences are eclectic and often implicit. Jared Diamond’s biogeographical determinism, Michel Foucault’s analysis of power and discourse, Francis Fukuyama’s teleological liberalism (which Harari accepts and then interrogates), Daniel Dennett’s philosophy of mind and consciousness, Peter Singer’s utilitarian ethics and the Buddhist philosophy that Harari practises as a meditator all leave visible marks. The synthesis is original even when the components are borrowed. Unlike Taleb, Harari is not primarily a technical thinker. Unlike Deutsch, he is not a scientist advancing a theory. He is a historian and philosopher of the big picture: someone who reads across biology, anthropology, economics, philosophy and technology and produces a narrative account of how the human animal got here and where it might be going. The appropriate standard for his work is not mathematical rigour or scientific precision but the quality of the synthesis, the originality of the framing and the intellectual honesty with which uncomfortable conclusions are confronted. Sapiens: A Brief History of Humankind (2011/2014) Sapiens covers 70,000 years of human history in approximately 400 pages. This is, on its face, an absurd enterprise. Any such project will involve simplification, generalisation and the suppression of complexity that would qualify or overturn specific claims. Harari is aware of this and takes it as a feature rather than a bug. His explicit ambition is not to produce a comprehensive history but to identify the most important patterns in human development and to offer a framework for understanding how we arrived at the present. The book is organised around three revolutions: the Cognitive Revolution approximately 70,000 years ago, the Agricultural Revolution approximately 10,000 years ago and the Scientific Revolution approximately 500 years ago. These are not arbitrary divisions. Each revolution represents a fundamental shift in the relationship between Homo sapiens and the rest of the biosphere, a shift in the sources of human power and a shift in the cognitive and social structures through which humans organise their lives. The Cognitive Revolution: The Power of Fiction The Cognitive Revolution is Harari’s most original and most influential conceptual contribution. The central question is: why did Homo sapiens outcompete and eventually eliminate every other human species? Neanderthals had larger brains. Homo erectus had a longer evolutionary track record. Other species of genus Homo inhabited the same environments and faced the same selection pressures. What was different about sapiens? Harari’s answer is language, but not language in the obvious sense of the ability to communicate information about the physical environment. Other animals do that. What is distinctive about the language of Homo sapiens is the ability to communicate about things that do not exist: to create and share fiction. Not fiction in the pejorative sense of lies or delusions, but fiction in the sense of intersubjective reality: entities and structures that exist not in the physical world but in the shared beliefs of a sufficiently large community. Money is the paradigm case. A banknote is, as a physical object, a piece of printed paper with negligible intrinsic value. Its value as money exists entirely in the shared belief of a community that it can be exchanged for goods and services. The moment that belief collapses, the banknote becomes physically unchanged but monetarily worthless. Nations, corporations, legal systems, human rights, religions and political ideologies are all fictions in this precise sense: they exist only insofar as sufficient numbers of people collectively believe in them and act as if they exist. This is not a debunking argument. Harari is not claiming that money, human rights or nations are unreal or unimportant. He is claiming the opposite: that these intersubjective fictions are the most powerful forces in human history, more powerful than any physical fact about the environment or any genetic fact about the individual. The ability to create and sustain shared fictions is what enabled Homo sapiens to cooperate in groups of thousands and millions, far beyond the scale accessible to other social animals. Chimpanzees cannot cooperate with strangers. They lack the cognitive machinery to extend trust beyond the small groups in which individual relationships can be maintained. Sapiens can cooperate with millions of strangers because they share common fictions: the same currency, the same legal system, the same national identity or the same religious beliefs. The evolutionary timing of the Cognitive Revolution is important. Around 70,000 years ago, something changed in the structure of the sapiens brain. The fossil and archaeological record shows a sudden explosion of symbolic behaviour: cave paintings, carved figurines, long-distance trade in non-utilitarian objects and, inferentially, the elaborate social rituals that require shared symbolic understanding. Harari argues this reflects the emergence of something like the modern human mind: capable of fiction, capable of narrative and therefore capable of the large-scale flexible cooperation that distinguishes sapiens from every other species. The consequence of this analysis is that human history is fundamentally the history of intersubjective fictions: which ones come to dominate, how they spread, how they interact, how they collapse and what replaces them. The Agricultural Revolution, the rise of empires, the spread of world religions, the Scientific Revolution, the emergence of capitalism and the liberal democratic order are all, in Harari’s framework, the history of changing sets of shared fictions that structure how vast numbers of people understand themselves and coordinate their behaviour. The Agricultural Revolution: History’s Biggest Fraud One of Sapiens’s most provocative chapters concerns the Agricultural Revolution, which Harari describes as history’s biggest fraud. The conventional story is one of progress: the transition from nomadic hunter-gathering to settled agriculture enabled population growth, specialisation of labour, the accumulation of surplus, the development of writing and mathematics and eventually the entire edifice of civilisation. This is true. But Harari asks a different question: was the Agricultural Revolution good for the individual human being? The answer, supported by a substantial body of physical anthropological evidence, is probably no. Hunter-gatherer skeletons show greater average height, fewer signs of nutritional deficiency and less evidence of the repetitive stress injuries characteristic of agricultural labour than early agricultural skeletons. Hunter-gatherers worked fewer hours per day, ate a more varied diet and were exposed to a smaller disease load because they did not live in dense settlements in close proximity to domesticated animals. The transition to agriculture brought population growth, but it also brought famine, epidemic disease, social hierarchy, slavery and warfare at a scale and intensity that hunter-gatherer societies rarely achieved. More precisely: the Agricultural Revolution was not good for individual humans but it was spectacularly good for the genes of the species of plants and animals that were domesticated. Wheat, rice, maize, cattle, pigs and chickens are among the most numerous organisms on earth precisely because their domestication by humans served their genetic propagation regardless of its effects on the humans doing the domesticating. The farmer who cleared a forest to plant wheat and then spent backbreaking days weeding, irrigating and harvesting was, in genetic terms, working for the wheat. The story of agricultural progress, narrated from the perspective of individual human wellbeing, looks rather different from the standard narrative of civilisational advancement. The deeper point is about the gap between what is good for the collective and what is good for the individual. The Agricultural Revolution enabled civilisations but may have made the average life worse. This gap recurs throughout Harari’s analysis of subsequent revolutions. Progress at the level of the system does not automatically translate into progress at the level of the individual organism. Empires, Religions and Money: The Unifiers of Humanity The middle section of Sapiens traces the mechanisms through which isolated agricultural communities became the globalised world. Three forces are identified as the primary engines of this unification: money, empires and religion. Each operates as a different kind of shared fiction. Money creates a universal medium of exchange that allows strangers to cooperate economically without prior trust. Empires create political structures that impose common laws, languages and cultures across diverse populations. Religions provide cosmic justifications for the existing social order and create communities of belief that transcend ethnic and political boundaries. Harari’s treatment of money is particularly sharp. He argues that money is the most universally successful story ever told because it is the only shared fiction that has successfully united virtually all human communities. People who share no language, no religion, no cultural values and no political allegiance can still do business if they agree on a common medium of exchange. The Roman denarius, the Arab dinar, the Spanish dollar and the American greenback have each, in their time, been accepted by populations with nothing else in common. The universality of money is not a natural fact. It is an extraordinary collective achievement, maintained by the ongoing shared belief of billions of people. The treatment of religion is analytically respectful even when it is theologically agnostic. Harari does not argue that religious beliefs are false, though he clearly does not hold them himself. He argues that religions serve specific social functions: they provide cosmic legitimation for social hierarchies, they create communities of shared practice and narrative that generate cooperation and social trust and they provide answers to questions about meaning and suffering that purely materialist accounts struggle to address. The great universalist religions, Buddhism, Christianity and Islam, were not merely spiritual movements. They were, in Harari’s analysis, the ideological infrastructure of the first genuinely global civilisations. Empires are treated with similar analytical detachment. The standard liberal narrative is that empires are oppressive structures imposed by conquerors on unwilling subjects and that the end of colonialism represents straightforward moral progress. Harari accepts the moral critique but complicates the historical narrative. Empires have been the primary vehicle of cultural unification, economic integration and the spread of technologies across the history of civilisation. The Roman Empire imposed Latin, Roman law and Roman infrastructure across a vast diverse population. The consequences for the peoples conquered were often brutal. The long-run consequences for the development of European civilisation were enabling. The tension between these two truths cannot be resolved by either a simple celebration or a simple condemnation of imperial history. The Scientific Revolution: Ignorance as the Engine of Progress Harari’s treatment of the Scientific Revolution identifies its most distinctive feature not as the discovery of specific truths but as the institutionalisation of ignorance: the collective admission that we do not know and the establishment of methods for finding out. Pre-modern knowledge systems, including the great religious traditions, operated on the assumption that all important knowledge had already been revealed, discovered or transmitted. The role of the scholar was to preserve, interpret and transmit that knowledge, not to generate new knowledge by challenging established doctrines. The Scientific Revolution established the opposite assumption: that current knowledge is incomplete, that the most important truths are yet to be discovered and that the method of investigation, controlled observation and experiment combined with mathematical reasoning, is more reliable than any authority or tradition. This epistemological revolution is, for Harari as for Deutsch, the central event of the modern period. But where Deutsch analyses its internal logic, Harari analyses its political and economic entanglements. Science did not develop in isolation. It developed in alliance with empire and capital. European powers funded geographic exploration, natural history, cartography and eventually physics and chemistry because scientific knowledge was strategically and economically valuable. The map was the companion of the cannon. The chemist was the companion of the industrialist. The relationship between knowledge and power in the Scientific Revolution was not incidental but structural: the same institutions that sought to understand the world sought to control it and the two projects reinforced each other. This produces one of Sapiens’s most important observations: the ideology of progress is not merely intellectual but deeply economic. Capitalism, Harari argues, is built on a specific kind of fiction: the belief in future growth. Pre-modern economic systems operated on the assumption that the total amount of wealth was roughly fixed. One person’s gain was another’s loss. Lending at interest was therefore either exploitation or risk, since the total pie could not grow. Capitalism created a new fiction: that investment today will produce more wealth tomorrow, that the economic pie can expand indefinitely. This belief, once sufficiently widespread, becomes self-fulfilling. Banks lend on the expectation of future growth. The loans fund investment. The investment produces growth. The growth validates the loans. The system sustains itself through the collective belief in its own expansion. Happiness: The Question Sapiens Cannot Answer Sapiens closes with what is perhaps its most philosophically honest chapter: the question of whether any of this progress has made human beings happier. Harari surveys the evidence and reaches the uncomfortable conclusion that we do not know. The measurement of subjective wellbeing across historical periods and cultures is methodologically extremely difficult. The subjective experience of a medieval peasant and a contemporary software engineer cannot be directly compared. The peasant may have experienced genuine contentment within the framework of his culture and religion. The engineer may experience chronic dissatisfaction despite material abundance. There is suggestive evidence from studies of subjective wellbeing that beyond a moderate income threshold, increases in wealth do not reliably increase happiness. There is evidence that social connection, meaningful activity and a sense of purpose are more strongly correlated with reported wellbeing than material conditions. There is biochemical evidence, which Harari takes seriously, that happiness is largely a matter of neurochemistry, that the hedonic baseline of individuals is relatively stable across varying life circumstances and that the pursuit of external achievements is to a significant degree a cognitive error. The Buddhist analysis, which Harari discusses as a meditator rather than merely as an intellectual, holds that suffering arises from craving and that the reduction of craving, rather than the fulfilment of it, is the path to genuine wellbeing. If this is correct, then the entire edifice of modern economic life, which is structured around the stimulation and satisfaction of desire, is not merely insufficient for happiness but positively counterproductive. Harari does not fully endorse this position but presents it as a serious challenge to the narrative of progress that the rest of the book has been assembling. Homo Deus: A Brief History of Tomorrow (2015/2016) Agenda: From What Happened to What Comes Next Homo Deus picks up where Sapiens leaves off and turns the historical analysis toward the future. The book begins with the observation that for the first time in human history the great traditional enemies of humanity, famine, plague and war, have been substantially reduced. Famine still exists but is now largely a political failure rather than an ecological inevitability. Epidemic disease still kills millions but the global death rate from infectious disease has been declining for a century and the proportion of human deaths caused by interpersonal violence is at a historical low. These are genuine achievements, worthy of acknowledgement. But what replaces famine, plague and war as the defining preoccupations of human civilisation? Harari’s answer is the pursuit of immortality, happiness and divinity: the three projects of what he calls Homo Deus, the upgraded human being who will use biotechnology and artificial intelligence to transcend the biological limitations of Homo sapiens. This is not science fiction speculation. It is, Harari argues, the logical extrapolation of existing trends. The Scientific Revolution established the principle that all problems are technical problems susceptible to technical solutions. Medicine has progressively pushed back the boundaries of disease and death. Pharmacology has developed progressively more precise interventions in the neurochemical basis of subjective experience. Artificial intelligence is beginning to outperform humans in cognitive tasks previously considered uniquely human. The trajectory of these developments points, within a timeframe of decades to centuries, toward the possibility of radical life extension, pharmacological control of mood and cognitive function and the creation of non-biological intelligence that surpasses human capability across all relevant domains. Humanism and Its Variants The philosophical core of Homo Deus is an analysis of humanism as the dominant religion of modernity. Harari uses “religion” in a sociological sense: a system of beliefs about what has ultimate value, what norms should govern behaviour and what narrative gives meaning to individual and collective life. By this definition, humanism is a religion: it holds that human beings have inherent value, that human experience is the ultimate source of meaning and authority and that the expansion of human freedom, flourishing and self-determination is the highest human goal. Humanism replaced the theistic religions of the pre-modern world not by defeating them in direct argument but by offering a more compelling account of where authority and meaning reside. The theistic account holds that authority derives from God’s revealed will and that meaning derives from one’s relationship to a divine order. The humanist account holds that authority derives from human experience and reason and that meaning derives from the development and expression of individual human potential. The Scientific Revolution and the Industrial Revolution made the humanist account far more practically compelling: science and technology, premised on human reason and human observation rather than divine revelation, demonstrably worked. Within humanism, Harari identifies three major variants. Liberal humanism holds that the individual human experience is the ultimate source of authority and that individual freedom, including freedom of thought, expression, economic activity and lifestyle choice, is the highest value. Socialist humanism holds that the collective human experience, especially that of the oppressed classes, is the ultimate source of authority and that equality, solidarity and collective wellbeing take precedence over individual freedom. Evolutionary humanism, whose most extreme form was National Socialism, holds that the human species rather than the individual or the class is the ultimate subject of history and that evolution toward a higher form of humanity is the supreme value. The twentieth century, in Harari’s reading, was the conflict between these three humanist variants. Liberal humanism won. The fall of the Soviet Union and the apparent universalisation of liberal democratic capitalism seemed to confirm Fukuyama’s thesis that liberal humanism was the final form of human political organisation: not because it solved all human problems but because it offered the best available framework for managing them. Harari accepts this assessment of the late twentieth century but argues that it understates the instability of liberal humanism’s dominance and that the technologies of the twenty-first century will fundamentally undermine its philosophical foundations. The Undermining of Liberal Humanism Liberal humanism is built on a set of assumptions about the human individual that modern science is in the process of dismantling. The core assumption is that there is a unified, autonomous, experiencing self: a subject whose preferences, values and choices are the ultimate source of authority. This is what “trust your feelings,” “be yourself” and “your life, your choice” all presuppose. The inner self is real, its experiences are authoritative and respecting individual freedom means respecting the sovereignty of that inner self. Harari argues that contemporary neuroscience and cognitive science are dismantling this picture. The unified autonomous self appears to be a fiction. What we call the “self” is a narrative produced by the brain’s interpretive systems after decisions have already been made, a post-hoc story that gives the illusion of unified agency to a collection of competing neural processes. Daniel Kahneman’s distinction between the experiencing self and the narrating self, split-brain experiments showing that the brain’s narrative systems confabulate explanations for actions initiated by processes they did not control and the accumulating evidence that unconscious processes drive most behaviour: all of this challenges the humanist picture of the sovereign individual. If the unified self is a fiction, what becomes of the humanist insistence on individual freedom and self-expression? If your preferences are the product of genetic programming, neurochemical dynamics and cultural conditioning rather than a unified authentic self, what does it mean to respect those preferences as sovereign? And more pressingly: if algorithms can predict your behaviour, model your desires and understand your neurochemical responses better than you can, does not the authority that humanism locates in the individual self actually reside more accurately in the algorithm? This is the pivot from humanism to what Harari calls Dataism: the emerging ideology that treats the universe as a flow of data and defines the value of any entity by its capacity to process information. In the Dataist framework, organisms are algorithms. Natural selection has been producing increasingly sophisticated information processing systems over billions of years. Human beings are among the most sophisticated. But there is no principled reason why the algorithms running on biological neural hardware are the most sophisticated possible or why biological hardware is the appropriate substrate for the most valuable forms of information processing. The Algorithm Problem and the End of Human Agency Homo Deus’s most disturbing argument concerns the economic and political consequences of algorithms that can perform cognitive tasks previously requiring human intelligence. Throughout history, human beings have been economically and politically valuable because they possess the twin capabilities of intelligence and consciousness. Intelligence, the ability to process information and make decisions, has been uniquely human. Consciousness, the ability to have subjective experience, remains uniquely human or at least uniquely biological. The development of artificial intelligence threatens to decouple these two capabilities and in doing so to create an unprecedented economic and political crisis. Algorithms are developing intelligence, defined functionally as the ability to perform cognitive tasks effectively, at a rate that is beginning to match or exceed human performance across an expanding range of domains. But they are not developing consciousness. They are not having experiences. They are processing information without the subjective dimension that characterises biological cognition. The economic problem is that most humans are valuable to the economy primarily for their intelligence, not for their consciousness. An algorithm that can diagnose disease more accurately than a doctor, drive a vehicle more safely than a human, write legal documents more reliably than a lawyer or analyse financial data more effectively than an analyst is economically more valuable for those purposes than the human practitioner. The fact that the algorithm has no inner life is economically irrelevant if what the economy demands is the output, not the experience. This creates the spectre of a “useless class”: a large and growing population of human beings who are economically redundant because algorithms can perform their cognitive functions more cheaply and reliably. Harari is careful to say this is not inevitable. Previous waves of automation displaced workers from particular tasks but created new tasks requiring human labour. There is no a priori reason why this wave should be different. But there are reasons to think it might be: previous automation displaced physical labour and created demand for cognitive labour. AI displaces cognitive labour and it is not obvious what new category of human activity it creates demand for. The political problem is potentially more severe. Liberal democracy is built on the assumption that individual citizens are sufficiently capable of understanding their interests and the implications of political choices to be trusted with political authority. If algorithms can model political preferences, predict policy consequences and identify the structural factors driving social outcomes more reliably than any individual or political elite, what is the democratic justification for overriding the algorithm’s recommendations with popular votes? Harari is not advocating algorithmic governance. He is identifying a tension within liberal democracy that the development of superior artificial intelligence makes acute. If we trust individual self-knowledge and individual judgment as the basis of democratic authority and if algorithms demonstrably have better models of both individual wellbeing and collective outcomes, then the humanist foundations of democracy are undermined. We face a choice between maintaining democratic institutions as a matter of principle, even when they produce suboptimal outcomes as judged by better information sources and revising those institutions in ways that increase dependence on algorithmic authority and thereby reduce human agency and self-determination. 21 Lessons for the 21st Century (2018) Structure and Purpose If Sapiens looks at the past and Homo Deus looks at the distant future, 21 Lessons for the 21st Century is concerned with the immediate present: the political, technological and cultural crises that define the moment we are living through. The book is more loosely structured than its predecessors, organised around twenty-one discrete topics rather than a single sweeping narrative. This is a strength and a weakness: the range is impressive but the unity is weaker. The central anxiety of the book is the simultaneous failure of three of the dominant frameworks through which post-Cold War Western liberalism understood itself and organised its institutions. Technological disruption in the form of AI and biotechnology is creating economic and social changes that liberal democratic institutions were not designed to manage. The ecological crisis, particularly climate change, presents a collective action problem at a global scale that the existing structures of national sovereignty and market capitalism have been unable to address. The resurgence of authoritarian nationalism in democracies previously assumed to be stable is challenging the assumption that liberal democracy is either universally appealing or self-sustaining once established. The Failure of Liberal Narratives Harari argues that the deepest problem is not practical but narrative. Human beings do not merely need solutions to practical problems. They need stories that make sense of their place in the world, justify the sacrifices the social order demands and point toward a coherent future. The liberal narrative that dominated the post-Cold War period, the story of ever-expanding individual freedom, democratic governance, market integration and technological progress gradually eliminating poverty and spreading prosperity, is no longer compelling to large parts of the populations of Western democracies. This narrative failure is not primarily because the story was false. Economic growth, the spread of democracy and the reduction of extreme poverty were real achievements of the post-Cold War period. The narrative failure is because the story did not adequately account for who bore the costs of the disruptions that accompanied those achievements, because it made promises about the trajectory of globalisation that the global financial crisis of 2008 and the stagnation of median wages in Western countries revealed as overoptimistic and because it offered nothing to communities that experienced the benefits of globalisation as abstractions and the costs as concrete. The alternative narratives that have gained ground, various forms of nationalist populism, religious fundamentalism and authoritarian technocracy, are, in Harari’s assessment, not genuine solutions to the problems they identify. They are emotional responses to real grievances that do not have the intellectual content to generate workable policies. They tell people who has caused their problems (foreigners, immigrants, elites, the corrupt establishment) without offering genuine mechanisms for solving them. But the emotional authenticity of the grievances they express means that the liberal intellectual dismissal of these movements as mere ignorance or irrationality has been both analytically wrong and politically counterproductive. Technology, Truth and Attention Several chapters of 21 Lessons address the relationship between technology, media and political reality. The thesis is that the combination of social media’s algorithmic attention optimisation and human cognitive biases produces an information environment that is structurally hostile to the kind of rational deliberation that liberal democracy presupposes. Social media platforms are optimised to maximise engagement, which in practice means maximising emotional arousal, particularly outrage and fear. Algorithms learn quickly that content producing strong negative emotional responses generates more clicks, shares and comments than content producing calm reflection. The result is an information environment that systematically amplifies the most emotionally provocative, the most divisive and the most extreme content regardless of its truth value. This is not a conspiracy. It is an emergent consequence of optimising for engagement in an environment where strong emotions drive clicks. The neurological and psychological vulnerabilities that social media exploits are not new. Human beings have always been more responsive to emotionally salient information than to statistically representative information. Fear and threat responses evolved in environments where rapid emotional response to immediate physical danger was adaptive. The modern information environment activates these responses continuously with stimuli that bear no relationship to actual physical danger and thereby degrades the capacity for calm, rational assessment that complex policy questions require. Harari’s prescription in this domain is more modest than his diagnosis. He recommends critical consumption of media, meditation as a tool for developing the capacity to observe one’s own cognitive processes and scepticism toward all narratives, including the ones that feel most emotionally compelling. These are individually sensible recommendations but they leave unaddressed the structural question of what institutional and regulatory responses could modify the information environment at scale. Resilience, Community and Meaning The latter sections of 21 Lessons are concerned with the question of how individuals can maintain psychological and moral coherence in an environment of accelerating change, epistemic uncertainty and collapsing grand narratives. Harari’s answer draws on the Buddhist and Stoic traditions he engages in Sapiens. The recommendation is not to seek certainty or to commit to a comprehensive ideology. It is to cultivate the capacity to be present with uncertainty: to develop the ability to sit with not knowing, to maintain emotional equanimity in the face of rapid change and to build the local, immediate relationships and communities that provide meaning regardless of the grand narrative context. This is less a political programme than a personal practice but Harari presents it as the foundation on which any more ambitious engagement with the challenges of the age must be built. The chapter on education is one of the book’s most practically pointed. In a world where the specific skills and knowledge that are economically valuable will change rapidly and unpredictably, the most important educational objectives are the meta-skills of learning how to learn, maintaining psychological flexibility and cultivating the critical thinking capacity to evaluate information and resist manipulation. These are not easily measured by standardised tests and they are not well served by educational systems designed around the transmission of fixed content. The institutional implications of this analysis are significant but Harari does not develop them beyond the level of suggestion. Nexus: A Brief History of Information Networks from the Stone Age to AI (2024) The Frame: Information Networks as the Key to Human History Nexus is Harari’s most directly political book and arguably his most timely. It was published in the context of widespread alarm about the potential of artificial intelligence to disrupt democratic institutions, spread disinformation and concentrate economic and political power in the hands of small technical elites. Harari’s response to this context is characteristically oblique: rather than addressing AI directly, he situates it within a long history of information networks and asks what that history reveals about the conditions under which new information technologies empower human flourishing or enable oppression. The central argument is that information is not simply knowledge or truth. Information networks can carry falsehoods as readily as truths and throughout history the most powerful information networks have often derived their power precisely from their capacity to propagate myths, rituals and narratives that are not accurate descriptions of reality but are effective coordinators of large-scale human behaviour. This extends and radicalises the argument of Sapiens. Where Sapiens argued that shared fictions enable cooperation, Nexus argues that information networks, the institutional structures through which shared fictions are produced, distributed and enforced, are the primary determinants of political power. Whoever controls the information network controls the society. The History of Information Networks Harari traces the development of information networks from the oral traditions of preliterate societies through writing, printing, the telegraph, mass media and the internet. Each transition in information technology transformed the political landscape by changing who could produce, store, distribute and authenticate information. Writing enabled the first large-scale bureaucratic states by making it possible to record laws, taxes, property rights and administrative decisions across time and space beyond the reach of individual memory. But writing also required a scribal class with specialised skills, concentrating the power of the information network in the hands of a small literate elite. The power of ancient empires was inseparable from the power of their information management systems: the archives, the record-keeping traditions and the scribal bureaucracies that maintained them. The printing press democratised the production of information and thereby destabilised the information monopolies of the Catholic Church and the feudal aristocracy. The Reformation was, among other things, an information revolution: Luther’s theses spread across Germany in weeks because of the printing press in a way that would have been impossible in the manuscript era. But the printing press also enabled new forms of information control. Early modern states developed censorship apparatus, licensing regimes and information monopolies to manage the destabilising consequences of cheap print. The same technology that enabled the Reformation also enabled effective state propaganda. The twentieth century’s mass media, radio, cinema and television created information environments of unprecedented reach and uniformity. A single broadcaster could reach tens of millions of people simultaneously with a single message. This enabled new forms of democratic politics but also new forms of totalitarian control. The fascist and Soviet regimes of the mid-twentieth century used mass media with sophisticated effectiveness to produce populations that were simultaneously better informed about the state’s official narrative and more thoroughly isolated from alternative sources of information than any previous population in history. AI as a Qualitative Break Harari’s argument in Nexus is that artificial intelligence represents not merely a new information technology but a qualitative break in the history of information networks. All previous information technologies were tools: passive media through which human agents produced, distributed and interpreted information. AI is not a passive medium. It is an active agent: it can generate information, interpret information, make decisions based on information and act in the world on the basis of those decisions without human involvement at each step. The consequence is that for the first time in history it is possible to construct an information network that does not merely transmit and store human decisions but that makes decisions independently. This has two implications that Harari considers alarming. First, the scale of possible surveillance and control is qualitatively different. Previous authoritarian regimes were constrained by the limits of human information processing. The Stasi in East Germany employed one informer for every 63 citizens and still could not achieve comprehensive surveillance. An AI-enabled surveillance state can monitor communications, movements, financial transactions, facial expressions and physiological responses at a scale and granularity no human bureaucracy could match. The authoritarian states of the twentieth century were limited by the information processing capacity of human agents. The authoritarian potential of AI-enabled surveillance eliminates that constraint. Second, the possibility of AI systems that can produce convincing synthetic content, texts, images, video and audio that are indistinguishable from genuine human production, represents an unprecedented threat to the epistemic foundations of democratic society. Democratic deliberation requires a shared factual reality: a baseline of agreed facts about the world that citizens can use as common ground for political argument. That baseline is already severely eroded by social media fragmentation. AI-generated synthetic content could effectively destroy it, producing an environment in which no evidence can be trusted, every document can be faked and every recorded statement can be attributed to anyone. The Asymmetry of AI Deployment One of Nexus’s most important contributions is the analysis of the asymmetry between offensive and defensive uses of AI in the information domain. Creating convincing synthetic disinformation is, with current AI tools, cheap, rapid and requires minimal skill. Detecting synthetic disinformation requires much more sophisticated technical tools, constant updating as generation techniques improve and either computational resources or human expertise at a scale that is not readily available to most organisations, individuals or indeed most democratic governments. This asymmetry means that the deployment of AI for information manipulation systematically advantages the attacker over the defender. A state actor, a well-funded political organisation or even a modestly resourced non-state actor can deploy AI-generated disinformation at a scale and sophistication that overwhelms the defensive capacity of the institutions, journalists and regulatory bodies attempting to identify and correct it. The result is not a specific set of false beliefs that can be corrected but a generalised epistemic crisis: a population that cannot distinguish true from false, that has lost confidence in the possibility of reliable information and that therefore retreats into tribal epistemologies where trust is allocated by group membership rather than by evidence. Democratic Fragility and Institutional Design Nexus is Harari’s most concerned and least optimistic book. The threat he identifies is not that AI will inevitably produce authoritarian outcomes. It is that AI creates new vulnerabilities in democratic institutions that those institutions, designed for a pre-AI information environment, are not equipped to manage, and that the window for making the institutional adaptations necessary to close those vulnerabilities is narrow. The required adaptations are, in broad terms: regulatory frameworks for AI development and deployment that prioritise transparency and accountability, institutional investment in digital literacy at scale, legal and technical standards for the authentication of information that can survive AI-generated synthesis and international agreements that prevent the most dangerous forms of AI-enabled information manipulation. None of these is simple. All of them require effective governance at a moment when the institutions of democratic governance are already under stress from the social media information environment that preceded AI. Harari’s concern is that the pace of AI development is outrunning the pace of institutional adaptation, that the asymmetry between the offensive and defensive uses of AI in the information domain means that the gap widens rather than closes as the technology advances and that the window for institutional responses may be closing before those responses have been designed, agreed upon and implemented. Academic Work: The Military Historian Before the Public Intellectual It is worth acknowledging Harari’s academic origins before drawing conclusions about his project. His doctoral thesis, published as Special Operations in the Age of Chivalry in 2007, examined special military operations in western Europe between 1100 and 1550. The work is meticulous, empirically grounded and appropriately hedged: everything that his popular books are sometimes accused of not being. The concerns of the academic work, the gap between the cultural construction of martial heroism and the material reality of violence, the role of narrative in shaping how participants understood and recorded military experience, the social function of collective myths about warfare, are continuous with the concerns of the popular works even though the register is completely different. The academic Harari’s interest in how stories and myths shape the human experience of violent reality became the popular Harari’s interest in how stories and myths shape all of human social reality. The transition is coherent and the continuity is genuine. But the transition also involves a change in epistemic standards that is worth acknowledging. The academic monograph must defend every claim against specialist objection. The popular synthesis must persuade a non-specialist reader. The rhetorical and evidential requirements are different and the trade-offs between scope and rigour are real. Assessing the Project: Strengths and Limitations What Harari Does Well The most important thing Harari does well is synthesis. He reads across an extraordinary range of disciplines and produces accounts that are genuinely integrative rather than merely additive. The argument that shared fictions are the primary engine of human social organisation is original in its framing even if its components are drawn from existing anthropology and social theory. The analysis of humanism as a religion with internal variants is an illuminating framework for understanding modern political history. The identification of the asymmetry between AI’s offensive and defensive uses in the information domain is an analytically sharp and practically important observation. He is also genuinely willing to follow arguments to uncomfortable conclusions. The argument that the Agricultural Revolution was bad for most people who lived through it despite being good for the species. The argument that liberal democracy’s philosophical foundations are being undermined by the science of cognition that liberalism itself funded. The argument that the development of AI might render large portions of the human population economically redundant. These are not safe observations. They are the kind of claims that a writer more concerned with reputation management would soften or qualify into meaninglessness. His prose is exceptional: clear, precise, often elegant and consistently free of the jargon that makes specialist writing inaccessible to general readers. The books are a pleasure to read and the pleasure is not purchased at the expense of intellectual content. This is rare and should not be taken for granted. What Harari Does Less Well The criticisms of Harari’s work are serious and should be engaged with honestly. The most important is the standard criticism of big-picture history: the generalisations that enable the synthesis also suppress the complexity that qualifies and sometimes contradicts the synthesis. When Harari says “hunter-gatherers worked fewer hours per day than farmers” or “the liberal narrative has failed,” these are claims about vast and heterogeneous populations across millennia. The evidence base for such claims is necessarily incomplete and contested. Specialists in relevant areas frequently find Harari’s characterisations of their fields to be oversimplified or selectively supported. The treatment of consciousness and free will in Homo Deus is philosophically thinner than it needs to be. The claim that the “self” is a fiction constructed by the brain’s narrative systems is much more contested in philosophy of mind and cognitive science than Harari’s confident tone suggests. The relationship between the narrative self, free will, moral responsibility and the practical structures of legal and political institutions is a genuinely hard philosophical problem that cannot be resolved by pointing at neuroscientific evidence for unconscious processing. Harari uses the science correctly at the level of empirical claims but the philosophical inferences he draws are more controversial than acknowledged. His treatment of Buddhism and meditation is substantive and personally authentic but the move from Buddhist phenomenology to broad claims about the nature of consciousness and the self involves philosophical steps that are not always made explicit. The Buddhist analysis of the self as a constructed narrative is a genuine and important insight but it does not straightforwardly entail the political and economic conclusions Harari draws from it. Nexus, while timely and often sharp, is occasionally repetitive and its prescriptions are proportionally thin relative to its diagnosis. The scale of the threat Harari identifies, the potential destruction of the shared epistemic reality that democracy requires, calls for institutional responses of corresponding ambition. The book identifies the problem with greater clarity than it illuminates the solutions. The Unity of Harari’s Project Across all four major books, three commitments define Harari’s project. The first is that human beings are primarily story-telling and story-believing animals: that our unique capacity for large-scale flexible cooperation is grounded not in physical power or cognitive superiority over other animals but in the ability to share fictions that coordinate behaviour across populations of strangers. The second is that this capacity, while the source of everything distinctive and powerful about human civilisation, is also the source of its deepest vulnerabilities: to manipulation, to ideological capture and to the displacement of genuine understanding by emotionally compelling narrative. The third is that the technologies of the twenty-first century, AI and biotechnology, represent a potential discontinuity in human history because they threaten to alter the biological and informational substrate on which human agency, identity and political organisation have historically rested. These three commitments form a coherent intellectual programme even though the individual books address different periods and problems. The historian of shared fictions became the analyst of their contemporary crisis. The diagnosis of the present and the projection of the future are continuous with the analysis of the past. Conclusion: Harari’s Place in the Intellectual Landscape Harari occupies an unusual position: too synthetic and too willing to speculate for academic historians, too grounded in evidence and too intellectually serious for mere popular commentary. The criticism from specialists that he oversimplifies is correct but partially misses the point. Every act of synthesis involves simplification. The question is whether the simplification illuminates more than it distorts. For Harari’s central arguments, about the role of shared fictions in human social organisation, about the vulnerability of liberal democracy to information manipulation and about the unprecedented challenges posed by AI and biotechnology, the illumination is genuine and the distortions, while real, do not undermine the core claims. The more serious criticism is about calibration of confidence. Harari writes with a rhetorical certainty that is not always matched by the strength of the evidence. The big-picture synthesis invites the reader to experience a kind of comprehensive understanding that the actual state of knowledge in the relevant domains does not fully support. This is a genuine risk: confident synthesis can produce the illusion of understanding in the same way that Deutsch’s bad explanations produce the illusion of explanation. But the alternative to Harari’s project is not a more rigorous version of the same project. It is the absence of any serious attempt to synthesise across scales and disciplines at the level of generality that the problems of the present require. The challenges posed by AI, by ecological crisis and by the fragility of democratic institutions are not confined to any single discipline. They require precisely the kind of cross-domain synthetic thinking that Harari practices. The appropriate response to the risks of that practice is not to abandon it but to read it critically, to check specific claims against specialist knowledge and to hold the frameworks lightly while taking the questions seriously. Harari is, whatever his limitations, asking the right questions. In a world dominated by specialists who answer the wrong questions with great rigour, that is a more valuable contribution than it might appear. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

10. 

    12 Jun 2026

    On Game Theory

    In 1983, the United States ran a war game called Able Archer. Soviet intelligence, watching it unfold, could not tell whether it was a simulation or the real thing. Both sides had nuclear weapons. Both sides had dominant strategies. The world nearly ended not because anyone wanted it to but because the structure of the game made mutual destruction the rational move. Game theory does not begin with mathematics. It begins with that. There is a mode of reasoning that sits underneath economics, politics, evolutionary biology, military strategy, negotiation and social organisation simultaneously. It is the recognition that the outcome of any situation involving more than one decision-maker depends not just on what you do but on what others do in response to what you do and what they expect you to do in response to what they do in response. This recursive structure of interdependent decision-making is the domain of game theory and once you have genuinely internalised its logic you cannot look at competitive, cooperative or mixed situations the same way again. The books I referred for writing this piece collectively build this mode of reasoning from first principles to sophisticated application. What follows is what they collectively teach. Thanks for reading tiwaryshailesh’s Newsletter! Subscribe for free to receive new posts and support my work. The Foundational Architecture: What a Game Actually Is Every strategic situation has a structure that can be analysed before any specific move is considered. The structure consists of the players, the choices available to each player, the information each player has when making their choice, the timing of decisions and the payoffs that each combination of choices produces for each player. Before asking what to do in any strategic situation, the systems thinker asks what game is actually being played here because the answer to that question determines what the available strategies even are. This prior question is consistently neglected in practice. Most people enter strategic situations with their attention focused on their own options and their own objectives, with the other players modelled only superficially. Game theory insists on the prior step: who are all the players, what are their actual payoffs, what information do they have and what is the timing structure of the interaction. The answers are frequently different from first appearances and the difference is strategically decisive. Dixit and Nalebuff establish the most important initial distinction. Games differ along two fundamental dimensions. The first is whether they are zero-sum or non-zero-sum. In a zero-sum game, every gain for one player is exactly a loss for another. The total value in the game is fixed and the question is only how it is divided. Chess, most sports and many political contests are approximately zero-sum. In non-zero-sum games, the total value available depends on what players do collectively. Both players can gain or both can lose depending on their choices. Most business, most negotiation and most social interaction is non-zero-sum. The error of treating non-zero-sum games as zero-sum is one of the most expensive strategic mistakes available and it is extremely common. The second dimension is whether the game is played once or repeatedly. This distinction is as important as the zero-sum distinction and they interact in ways that are surprising and consequential. The logic of a one-shot game is completely different from the logic of the same game played repeatedly because in repeated games the future creates incentives that do not exist in single interactions. Reputation, reciprocity and the threat of future retaliation all become strategically relevant only when there will be future interactions. Understanding whether you are in a one-shot or repeated game with any particular counterpart is therefore one of the first and most important analytical moves available. Dominant Strategies and Nash Equilibrium: The Bedrock Once the structure of a game is understood, the search for strategic solutions begins with the concept of dominance. A strategy is dominant if it produces better outcomes for the player using it regardless of what any other player does. If you have a dominant strategy, you should use it. Full stop. The reasoning of the other players, the history of the relationship and the specific context are all irrelevant if one strategy is genuinely dominant. And if all players have dominant strategies, the outcome is determined by those strategies regardless of what any player prefers. The Prisoner’s Dilemma is the canonical illustration and Poundstone’s history of it reveals why it became so consequential so quickly after its invention. Two suspects are held separately and each is offered the same deal: confess and implicate the other and you go free while the other gets a long sentence, but if both confess you both get a medium sentence and if neither confesses you both get a short sentence on lesser charges. Each prisoner’s dominant strategy is to confess regardless of what the other does. If the other confesses, confessing reduces your sentence. If the other stays silent, confessing lets you go free. Confess is dominant for both players. The result: both confess and both get medium sentences, which is worse for both than if neither had confessed. The Prisoner’s Dilemma is not a curiosity. It is the structural template for an enormous range of consequential real-world situations. Arms races between nations. Price wars between competitors. The overuse of common resources. The underprovision of public goods. The collapse of cooperation in any situation where individual incentives diverge from collective interests. In every case the structure is the same: individual rationality produces collective irrationality. Each player doing what is best for themselves, given what they expect others to do, produces an outcome that is worse for everyone than the cooperative outcome they could have achieved. Nash Equilibrium generalises this insight beyond games with dominant strategies. A Nash equilibrium is a combination of strategies such that no player can improve their outcome by unilaterally changing their strategy given what all other players are doing. It is a stable point in the strategic landscape, not necessarily optimal for any player and sometimes far from optimal for all players collectively, but stable in the sense that no individual player has an incentive to deviate from it unilaterally. The prisoner’s dilemma mutual confession is a Nash equilibrium. Both players confessing is stable because given that one player confesses, the other’s best response is also to confess. Understanding Nash equilibrium is powerful not because it always tells you what to do but because it identifies the stable attractors of strategic situations. In any repeated strategic interaction, the system tends to move toward Nash equilibria over time as players learn and adapt. Knowing where the equilibria are tells you where the situation is heading and whether intervention to change the game structure is required to reach a better equilibrium. The Repeated Game Revolution: Why Cooperation Is Possible The most important single finding in the empirical game theory literature comes from Axelrod’s tournament and it overturns the naively pessimistic reading of the prisoner’s dilemma as proof that cooperation is irrational. Axelrod invited game theorists, economists and computer scientists to submit strategies for an iterated prisoner’s dilemma tournament in which the same pairs of strategies would play each other across many rounds. The winning strategy across multiple tournaments, submitted by the psychologist Anatol Rapoport, was the simplest strategy entered: Tit for Tat. Cooperate on the first move. Then do whatever the other player did on the previous move. Cooperate if they cooperated. Defect if they defected. Return immediately to cooperation if they do. Tit for Tat won not by being clever but by embodying four properties that Axelrod identified as the structural requirements for successful cooperation in repeated games. It is nice: it never defects first. It is retaliatory: it immediately punishes defection. It is forgiving: it returns to cooperation the moment the other player does. And it is clear: its strategy is simple enough that the other player can model it accurately and understand that cooperation will be rewarded and defection will be punished. The implications extend far beyond game tournaments. The Evolution of Cooperation shows that these same four properties characterise the emergence of cooperation in biological systems, in human social evolution and in the sustained cooperation that makes markets, institutions and societies function. Cooperation is not irrational. It is the rational response to repeated interaction with players who will retaliate against defection and reward cooperation. The conditions that make cooperation stable are the conditions that allow complex social organisation to exist at all: sufficient probability of future interaction, sufficient ability to identify and remember how specific players have behaved in the past and sufficient capacity to reward cooperation and punish defection. This analysis identifies precisely why cooperation breaks down and what restores it. Cooperation collapses when interactions become one-shot or when players believe they will not interact again, when players cannot identify who defected or cannot attribute specific defections to specific players, and when the temptation to defect is large relative to the long-term value of sustained cooperation. Every institution designed to sustain cooperation, from contract law to international treaties to professional reputation systems, works by addressing one or more of these breakdown conditions: extending the shadow of the future, improving the attribution of defection to specific players and increasing the cost of defection relative to the long-term value of cooperation. Commitment, Credibility and the Strategic Importance of Constraints Schelling’s contribution to game theory is the insight that most consistently surprises people who encounter it for the first time. In strategic interactions, the ability to constrain your own future choices is frequently more powerful than having more choices available. The player who can credibly commit to a specific course of action, and specifically to a course of action that includes following through on threats that would be irrational to execute if called upon, changes the strategic calculus for every other player. The strategic logic is precise. A threat is credible only if the player making it would actually follow through if the threatened behaviour occurred. A threat that would be irrational to execute is not a genuine strategic constraint on the other player because the other player knows it will not be followed through. The player who threatens to destroy their own negotiating position unless their demands are met is making a credible threat only if they can somehow commit to actually destroying their position. The nation that threatens nuclear retaliation for a conventional attack is making a credible threat only if the other side believes it would actually execute a response that would invite counter-retaliation. The genius of Schelling’s analysis is the recognition that constraining your own rationality, reducing your own options, burning your own bridges, can be strategically powerful because it changes what others rationally expect you to do. Cortez burning his ships on the shore of Mexico was not madness. It was a commitment device. It eliminated the option of retreat, which changed the rational expectation of both his soldiers and the Aztecs about what he would do when faced with adversity. The soldiers who could not retreat fought differently than those who could. The commitment changed the game. This insight propagates through an enormous range of strategic contexts. The labour union that commits to striking rather than accepting any settlement below a specific threshold has a more credible negotiating position than one that will accept anything above their current offer. The competitor that credibly commits to matching any price cut by a rival has made the price cut less attractive for the rival before it happens. The nation that credibly commits to defending a specific piece of territory regardless of cost has made attacking it less attractive than attacking territory whose defence is uncertain. In every case, the commitment works not by improving the outcome if the threat is called but by preventing the threat from being called because the other player’s calculation of expected outcomes shifts. Schelling’s related insight in Micromotives and Macrobehavior addresses the aggregation problem that sits alongside the commitment problem. Individual decisions that are each rational from the individual’s perspective aggregate into collective outcomes that nobody chose and that frequently nobody wanted. The residential segregation that emerges from individual preferences for having some neighbours of the same group, even when that preference is mild, produces complete segregation at the neighbourhood level through a tipping point dynamic. The traffic jam that emerges from individual rational decisions to leave at a specific time produces delays that everyone would prefer to avoid. The bank run that emerges from individual rational decisions to withdraw deposits once other depositors start withdrawing destroys the bank that would have survived if no one had panicked. These emergent outcomes from aggregated individual rationality create the social situations that institutions, norms and policies are designed to manage. And they are systematically misread as the product of malicious actors, cultural pathologies or structural oppression when they are in fact the predictable emergent properties of individual rational responses to specific incentive structures. Changing the emergent outcome requires changing the incentive structure, not blaming the individuals responding rationally to it. Changing the Game: The Highest Leverage Strategic Move Brandenburger & Nalebuff’s Co-opetition makes the argument that should sit at the top of any senior strategist’s thinking about competitive situations. The most powerful strategic moves available in non-zero-sum games are not moves within the current game. They are moves that change the game itself. They introduce the concept of the value net, an extension of game theory to multi-player business situations that includes not just competitors but complementors, players whose participation makes the value of your product or service higher rather than lower. Intel and Microsoft are complementors: each makes the other’s product more valuable. A hotel and a convention centre in the same city are complementors. A streaming platform and a content producer are complementors, at least until one of them decides to vertically integrate and become a competitor. Identifying and cultivating complementors is frequently more strategically powerful than defeating competitors because it expands the total value available in the game rather than just competing for a share of existing value. The five dimensions along which a game can be changed are: the players, the added values each player brings, the rules of the game, the tactics or information architecture and the scope of the game. Each of these is a dimension of strategic intervention that most players never consider because they are focused on playing the existing game better rather than changing it. Apple entering the mobile phone market did not compete harder at the existing game of mobile phones. It changed the game by redefining what a mobile phone was and what value it could provide. The most consequential business strategies of the past three decades are almost all game-changing moves rather than game-playing improvements. The game-changing insight also applies to negotiation. The most powerful negotiating move is rarely the smartest play within the current negotiation. It is changing the structure of the negotiation: adding players whose presence changes the payoff structure for existing players, changing the information architecture so that information previously private becomes common knowledge or vice versa, changing the timing so that delay or urgency shifts the relative costs for different players, or changing the outside options available to one or more players so that their walk-away position and therefore their bargaining power changes. Information Asymmetry: The Hidden Architecture of Strategic Advantage A recurring theme across multiple books in this list is the strategic role of information. In game theory, information structure determines what strategies are available and what equilibria are possible in ways that are as important as the payoff structure itself. The fundamental categories are: games of complete information where all players know all payoffs, games of incomplete information where some payoffs are private knowledge and games of imperfect information where players do not know the history of previous moves. Most real-world strategic interactions are games of incomplete and imperfect information and the management of information, what you reveal, what you conceal, what you credibly signal and what you strategically obscure, is as important as any other strategic choice. Signalling is one of the most important and most misunderstood mechanisms in strategic interaction. A signal is credible only if it is costly enough to send that a player without the underlying quality being signalled could not profitably fake it. The peacock’s tail is the canonical biological example: it is a credible signal of genetic quality precisely because it is so costly to maintain that only a genuinely healthy peacock can afford it. Education credentials signal capability partly through their content and partly through the selection effect of their cost: students who are unlikely to be productive workers find it more costly to obtain credentials and therefore self-select out, making the credential a credible signal even if it teaches nothing directly relevant to job performance. In business and negotiation contexts, the design of credible signals is a primary strategic challenge. The startup that wants investors to believe in its prospects needs to find signals that a low-quality startup could not credibly send, the founder taking a low salary, meaningful personal investment, specific contractual commitments around milestones. The negotiator who wants to signal genuine willingness to walk away needs to have taken actions that make walking away genuinely low-cost for them. The signal works only to the extent that it is genuinely costly to fake. Screening is the symmetric problem: the design of mechanisms that induce other players to reveal their private information. Insurance pricing that offers a menu of premium and deductible combinations screens for risk level without the insurer needing to observe it directly. Job offers that vary in fixed salary and performance bonus screen for confidence in own performance. Warranties screen for product quality. In every case the mechanism works by making the different options attractive to different types, allowing the structure of choices to reveal information that could not be obtained by direct observation or questioning. The Explore-Exploit Trade-off: Sequential Decision-Making Under Uncertainty Christian and Griffiths bring the insights of computer science optimal algorithms into contact with game theory through the lens of sequential decision-making under uncertainty. The explore-exploit dilemma, the tension between gathering information about options and using the best option currently known, arises in an enormous range of strategic contexts and has mathematically precise solutions that provide useful decision heuristics even when the exact mathematical conditions are not met. The multi-armed bandit problem, in which a player must choose repeatedly among options with unknown payoff distributions, is the canonical formulation. The optimal solution under specific conditions involves a specific balance between exploration of unknown options and exploitation of the best-known option. The key insight is that the optimal amount of exploration decreases over time as more information is gathered and as the time horizon shortens. Early in a sequence of decisions, exploring widely is valuable because the information gathered improves all subsequent decisions. Late in a sequence, exploiting the best-known option is correct because there is insufficient time remaining to recoup the cost of exploration through improved future decisions. Applied to hiring, this suggests trying many different types of candidates early in an organisation’s life when the cost of a wrong hire is low relative to the information gained and converging on a specific profile later when the cost of experimentation is higher. Applied to career decisions, it suggests deliberately exploring different roles and domains early in a career when the option value of information is high and the cost of switching is low. Applied to product development, it suggests running many small experiments early and concentrating resources on the best-performing approaches as the launch deadline approaches. The optimal stopping problem provides the precise answer to another class of sequential decisions: when to stop searching and commit to the best option found so far. The mathematical solution, explore for the first 37 percent of your search horizon then commit to the first option that exceeds the best seen during exploration, is not always directly applicable but it establishes the important qualitative principle that the optimal stopping point is always earlier than intuition suggests and that the cost of searching too long is frequently as large as the cost of stopping too early. Skin in the Game: The Alignment Problem Beneath All Strategic Interaction Taleb’s Skin in the Game addresses the most fundamental corruption of strategic interaction: the separation of decision-making from consequences. Every game theory model assumes that players bear the outcomes of their strategic choices. When this assumption is violated, when the person making the decision does not bear the cost of a bad outcome, the game being played is not the game that appears to be being played and the analysis that treats it as the apparent game produces systematically wrong predictions. The financial crisis of 2008 is the canonical modern example. The originators of mortgage loans who sold them on to securitisation vehicles had no skin in the game: they were paid for origination volume regardless of subsequent loan performance. The rating agencies that rated the resulting securities were paid by the issuers whose securities they rated: their skin in the game was the opposite of the investors relying on their ratings. The executives of financial institutions that failed were paid primarily in annual cash bonuses for short-term performance: their downside in the event of catastrophic failure was limited to the loss of future bonuses. The entire system was a game theory problem in which the apparent game, prudent financial intermediation, was not the actual game being played by the actual incentive structure, which was the extraction of fees and bonuses from a system whose tail risks were borne by others. The principle Taleb extracts is ancient and simple. Do not take advice from someone who does not bear the consequences of their advice. Do not trust risk assessments from institutions that profit from the risk without bearing its downside. Do not rely on the judgment of decision-makers who are insulated from the outcomes of their decisions. The absence of skin in the game corrupts the information content of every signal, the reliability of every commitment and the alignment of every incentive in ways that make standard game theory analysis inapplicable. The corollary is that the most reliable strategic partners, whether in business, politics or personal life, are those whose interests are genuinely aligned with yours through the structure of their payoffs rather than through the content of their stated commitments. Stated commitments are cheap. Structural alignment of payoffs is expensive to fake and therefore credible. This is the skin in the game version of Schelling’s commitment insight: the most credible signal of future behaviour is a structure of payoffs that makes the desired behaviour the rational choice regardless of stated intentions. Political Games: The Strategic Logic of Power Bueno de Mesquita and Smith’s Dictator’s Handbook applies game theory to political power with a rigour that explains what conventional political analysis consistently cannot. Their model treats political power as a multi-player coalition game in which the leader’s survival depends on maintaining the support of a winning coalition and all decisions, however dressed in the language of national interest or ideological principle, are best understood as moves in this coalition maintenance game. The three key variables are the nominal selectorate, everyone who has some formal role in choosing the leadership; the real selectorate, those whose support actually matters for leadership survival; and the winning coalition, the minimum subset of the real selectorate whose support is sufficient to maintain power. The game theory logic is inexorable. The smaller the winning coalition relative to the real selectorate, the more cheaply the leader can buy each coalition member’s loyalty because each member knows they are replaceable by someone from the larger pool and therefore accepts less. The larger the winning coalition, the more expensive each member’s loyalty and the more the leader must deliver public goods to maintain power. This single structural variable, coalition size, predicts an enormous range of political outcomes that conventional analysis treats as the result of cultural, ideological or personal factors. Large coalition systems, which require broad popular support, tend to provide public goods, respect civil liberties and maintain institutional constraints because the leader needs too many supporters to maintain power through private rewards alone. Small coalition systems tend to be kleptocratic, repressive and institutionally arbitrary because the leader can maintain power by distributing private rewards to a small group and has no incentive to provide the public goods that the excluded majority would benefit from. The game theory implication for understanding any political system is to identify the actual winning coalition, not the formal institutional structure and to understand the payoff structure that determines what leaders rationally do. The analysis extends to corporate governance, institutional management and any organisation where power is maintained through coalition. Wherever you find outcomes that seem irrational when explained by stated objectives, the game theory analyst asks what coalition is being maintained and what payoffs that coalition requires. Forecasting in Strategic Environments: The Epistemic Discipline Tetlock and Gardner’s Superforecasting connects game theory to the epistemic practices that make strategic analysis accurate rather than merely sophisticated. The challenge of strategic reasoning is not only the design of optimal strategies but the accurate assessment of the probabilities of different scenarios, the likely responses of other players and the range of outcomes that different strategy combinations can produce. Tetlock’s research on forecasting accuracy identifies the habits of thinking that distinguish the best forecasters from the rest. They think in probabilities rather than binary predictions, assigning numerical likelihoods to outcomes rather than predicting yes or no. They update their probabilities in response to new information rather than anchoring on initial assessments. They actively seek disconfirming evidence for their current views rather than only confirming evidence. They decompose complex questions into simpler components that can be assessed more accurately. They use base rates, the historical frequency of similar events, as anchors that are adjusted based on specific case features rather than ignored in favour of case-specific narratives. In game theory terms, the superforecaster is the player who most accurately models the probability distributions over other players’ strategies and over the outcomes of different strategy combinations. The player with the most accurate model of the game’s probability structure, other things equal, will make better strategic decisions because they are choosing among options whose expected values they can estimate more accurately. The calibration insight is particularly important. A well-calibrated forecaster’s 70 percent confidence predictions are right 70 percent of the time, not 90 percent or 50 percent. Most people are systematically overconfident in their predictions. In strategic contexts, overconfidence about the probability that your strategy will succeed, that the other player will respond as you predict or that the game structure is as you have modelled it produces strategies that look optimal in the modelled game and perform poorly in the actual game. The discipline of honest probability assessment, including honest assessment of your own uncertainty about the game structure itself, is the epistemic foundation that makes all the strategic sophistication above actually useful. The Unified Picture Pull all of this together and the complete architecture of strategic intelligence becomes visible. The first layer is structural analysis. Before any strategic move is considered, the game must be understood: who are the players, what are their actual payoffs, what information do they have, what is the timing structure and most importantly what kind of game is actually being played. Many strategic failures begin with misidentification of the game type: treating a repeated game as a one-shot game and defecting at the cost of a long-term relationship, treating a non-zero-sum game as zero-sum and missing the cooperative value available, treating a game of complete information as one of incomplete information or vice versa. The second layer is equilibrium analysis. Given the game structure, where are the stable attractors and are they good or bad equilibria from the perspective of the players? The prisoner’s dilemma mutual defection is a stable equilibrium that is bad for all players. Identifying that the equilibrium is bad and that the game structure is producing it points directly to the intervention: change the game structure so that a better equilibrium becomes stable. The third layer is commitment and credibility. What commitments can you make that are credible because they are costly to fake? What constraints on your own future choices change the rational expectations of other players in ways that benefit you? What signals can you send that separate you from players trying to fake your qualities? What screening mechanisms reveal the private information of other players? The fourth layer is the repeated game dynamic. If this interaction will recur, what reputation are you building with every move? Is your current behaviour consistent with the long-run strategy that serves your interests across the full sequence of interactions? Are you in a relationship where the tit-for-tat logic of reciprocal cooperation applies or a one-shot interaction where different logic applies? The fifth layer is game changing. Can you change the players, the payoffs, the rules, the information structure or the scope of the game in ways that produce better outcomes than any strategy within the existing game? This is the highest-leverage layer and the most neglected in practice. And beneath all of it is the epistemic layer: are your probability assessments of outcomes, player types and strategy effectiveness well-calibrated? Are you updating accurately on new information? Are you accounting for the skin-in-the-game distortions in the information and advice you are receiving? What This Demands of You Game theory does not tell you what to do. It tells you how to think about what to do. The demand it makes is the demand of genuine strategic sophistication: the willingness to model other players as rational agents with their own objectives rather than as obstacles or allies whose behaviour is determined by their character rather than their incentives. The most practically important single shift that game theory produces in a sophisticated practitioner is the habit of asking not just what is my best move but what will the other players do in response to my move and what will I do in response to that. This forward induction into the sequence of responses that any move initiates changes the evaluation of options in ways that consistently produce better outcomes than analysis that stops at the first move. The person who has genuinely internalized this mode of thinking approaches negotiations differently, understanding that the other side’s constraints and payoffs are as important as their own. They approach competitive situations differently, looking for game-changing moves rather than only game-playing improvements. They approach institutional design differently, understanding that the payoff structure determines behaviour more reliably than stated values or intentions. And they approach forecasting differently, maintaining calibrated uncertainty rather than false confidence about how the game will play out. Game theory is ultimately the formalisation of the oldest strategic wisdom: know the game you are in, know the players you are with, know where the stable outcomes are and know how to build the commitments and the reputation that make the best outcomes stable. The mathematics is optional. The thinking is not. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

11. 

    12 Jun 2026

    On Dawkins' Work

    Every living thing on this planet is a machine built by something that has never been alive, has no mind, no intention and no goal. It does not know you exist. It does not care whether you suffer or thrive. It has produced, through four billion years of blind copying and filtering, the eye, the immune system, the human brain and the feeling that there must be something behind all of this. There is not. That is Dawkins’s argument and it is one of the most consequential ideas in the history of human thought. Richard Dawkins is a British evolutionary biologist, popular science writer and public intellectual whose output across more than five decades spans foundational evolutionary theory, philosophy of biology, cognitive science of religion and secular advocacy. His work include multiple books, alongside a substantial body of academic papers, lectures and essays. The single organising logic that runs through virtually everything he has written is this: natural selection operating on replicators is the only known process capable of generating the appearance of design in living systems, and once this is properly understood it dissolves, one by one, the mysteries that have historically driven people toward teleological or supernatural explanations. Every major work is either an elaboration of this principle, a defence of it against rival accounts within biology or a prosecution of its implications for religion and human self-understanding. His intellectual lineage runs directly through Charles Darwin, R. A. Fisher, J. B. S. Haldane and W. D. Hamilton. He works firmly within the neo-Darwinian synthesis and draws heavily on George C. Williams’s Adaptation and Natural Selection (1966), which first argued rigorously that selection acts on genes rather than groups. He engages sympathetically with John Maynard Smith and the tradition of formal evolutionary game theory. The positions he argues against are multiple and sometimes contradictory with one another: group selectionists from V. C. Wynne-Edwards through David Sloan Wilson, Lamarckians and directed-mutation theorists, proponents of punctuated equilibrium in its stronger forms, the adaptationist critics Stephen Jay Gould and Richard Lewontin, theists of every variety and, in his later career, the entire tradition of natural theology from William Paley onward. The Selfish Gene (1976): The Replicator as the Unit of Darwinian Analysis The founding document of Dawkins’s intellectual project begins not with organisms but with a thought experiment about the origin of self-replicating molecules in a primordial chemical soup. This framing is not decorative. It establishes the conceptual priority of the replicator; the entity whose differential copying success is the engine of all subsequent biological complexity. Organisms, on this view, are not the primary units of selection but the vehicles replicators build to promote their own propagation. The gene is the replicator; the body is the machine it rides in. The core argument of The Selfish Gene is that genes behave, in a statistical sense, as though they were selfish, meaning they have been selected precisely because they promoted their own copying at the expense of alternatives, including genetic alternatives carried by other individuals of the same species. This insight, drawn largely from Hamilton’s work on inclusive fitness and from Trivers’s reciprocal altruism, allows Dawkins to reframe the evolution of altruistic behaviour. Apparent sacrifice by one organism for another is not a puzzle for Darwinian theory once you realize the gene shared between relatives is the unit whose frequency is being tracked. The organism helps its kin not because it is altruistic in any human sense but because the gene that produces helping behaviour is thereby helping copies of itself. What makes this book the load-bearing pillar of everything that follows is not its account of kin selection, which was already mathematically established, but its rhetorical and conceptual move of changing the level at which the reader is invited to see selection working. Dawkins argues that if you habitually think from the gene’s point of view rather than the organism’s point of view, a wide range of biological phenomena that otherwise look puzzling or paradoxical resolve themselves. Parent-offspring conflict, sexual conflict, the evolution of virulence in parasites, the seemingly bizarre behaviours of social insects, all become tractable when the question asked is not “what benefits this organism or this group?” but “what benefits this replicating sequence?” The book also introduces the concept of the meme in its final chapter, a replicating unit of cultural information analogous to the gene. This is treated briefly and speculatively in The Selfish Gene itself but becomes important later as an attempt to extend the replicator logic beyond biology. The Extended Phenotype (1982): Selection Beyond the Skin The Extended Phenotype is Dawkins’s most technically rigorous book and the one he has said represents his most original scientific contribution. Its central argument is that the phenotypic effects of a gene need not be confined to the body of the organism carrying that gene. A gene can influence the world beyond the organism’s skin, and that extended influence is part of what selection acts upon. The examples are deliberately strange and carefully chosen. A gene in a beaver that produces dam-building behaviour has phenotypic effects that extend into the physical landscape. A gene in a cuckoo that produces the egg-pattern matching its host has phenotypic effects that extend into the body of the host species. Parasites that manipulate host behaviour are expressing their genes in the host’s nervous system. The principle, once stated, is clarifying rather than merely striking: the boundaries we naturally draw around the organism are convenient but not theoretically fundamental to understanding what selection is doing. This book makes explicit what is implicit in The Selfish Gene: the vehicle, the organism, is theoretically dispensable as a category even if it remains empirically important. What matters is the replicator and its effects on the world, however far those effects extend. The book also contains a rigorous defence of adaptationism against the critiques already being mounted by Gould and Lewontin in their “Spandrels” paper. Dawkins distinguishes between genetic determinism (which he does not hold) and the claim that adaptation is best understood as the product of gene-level selection (which he does hold), and argues that confusing these two positions has generated a great deal of unnecessary controversy. The intellectual function of The Extended Phenotype within the overall project is to harden the theoretical foundations that The Selfish Gene established. If the popular book made the gene’s-eye view vivid and accessible, this academic monograph made it defensible against sophisticated biological objection. The Blind Watchmaker (1986): Design Without a Designer The title alludes to Paley’s Natural Theology (1802), in which the discovery of a watch on a heath implies a watchmaker, and thence the existence of God as designer of the natural world. Dawkins’s answer is that cumulative natural selection is itself a blind watchmaker: it produces the appearance of purposeful design without any designing mind behind it. The central argumentative move is the distinction between single-step and cumulative selection. Mount Improbable (a metaphor developed more fully in a later book) cannot be scaled in a single leap; no mutation produces an eye from scratch. But if each incremental improvement is preserved by selection while each decrement is eliminated, the cumulative effect across geological time can produce structures of extraordinary complexity and apparent purposiveness. The eye, the echolocation system of the bat, the aerodynamics of the swift, all are explicable as the summation of millions of individually small selective steps. Dawkins supports this argument with a now-famous computational demonstration: the Biomorph program, which he built himself and describes at length. Starting from a simple line-drawing ancestor and iterating through random mutations with human selection imposed, the program generates forms of surprising biological plausibility. The point is not that this simulates real evolution (it does not; selection here is by a human chooser, not by differential reproductive success) but that it makes viscerally persuasive the claim that cumulative selection can generate apparent design from simple beginnings. The book is also a systematic attack on creationism as a scientific hypothesis, conducted not primarily by appeals to the fossil record or molecular phylogenetics but by showing that the structure of the creationist argument is philosophically incoherent. The creationist argues that complexity requires a designer; Dawkins argues that a designer capable of creating biological complexity would itself require an explanation of equal or greater complexity. The hypothesis of God does not explain biological design; it merely relocates the explanatory problem while making it immeasurably larger. In the architecture of the overall project The Blind Watchmaker occupies a pivotal position. It is where the gene’s-eye logic of the first two books is explicitly connected to the oldest and most powerful argument for religious belief. The work is not yet primarily about religion (that comes in The God Delusion) but it establishes that Darwinism is the answer to the question that natural theology was attempting to answer. River Out of Eden (1995) and Climbing Mount Improbable (1996): The Digital River These two shorter books, published in quick succession, elaborate different aspects of the same central thesis. River Out of Eden introduces the metaphor of genetic information as a river of digital data flowing through time, branching but never merging (between species), and uses this to explore the nature of evolution as an information process. The striking section on the “utility function” of nature, which is not survival, not happiness and not beauty but purely the propagation of DNA, is one of Dawkins’s most unsparing formulations of the replicator logic and its implications for any attempt to find moral or aesthetic meaning inscribed in biological nature. Climbing Mount Improbable extends the single-step versus cumulative selection argument from The Blind Watchmaker with more detailed case studies, particularly a careful examination of the independent evolution of the eye in many separate lineages. The mountain metaphor does work that the blind watchmaker metaphor does not quite do: it represents not just the possibility of cumulative selection but the topology of adaptive space, making vivid why some trajectories are available to evolution and others are not. Together these books represent Dawkins in a mode of patient popularisation of the central logic, extending it into new empirical domains rather than breaking new theoretical ground. Their importance in the project is as vehicles for the central thesis rather than developments of it. Unweaving the Rainbow (1998): Science as an Aesthetic Project This is the most anomalous book in the Dawkins canon and the one most frequently overlooked in accounts of his intellectual project. Its argument is a direct response to Keats’s accusation, in Lamia, that Newton had destroyed the poetry of the rainbow by reducing it to a prism. Dawkins argues the opposite: that scientific understanding deepens rather than diminishes the aesthetic and emotional power of natural phenomena. The scientific account of why the sky is blue, of what starlight tells us about the composition of distant suns, of the improbability of any given individual being alive at all given the combinatorial lottery of sexual reproduction, these accounts are not cold reductions but expansions of wonder. This book matters to the overall project because it addresses the most sympathetic form of objection to Dawkins’s programme: not that Darwinism is scientifically wrong but that its reduction of living things to gene vehicles drains them of meaning. Dawkins’s counter is that the meme, the unit of cultural replication, propagates ideas precisely because they are in some sense good at propagating, and that science, poetry and music are all meme complexes that have been selected in cultural space for their ability to grip minds. This is not a satisfying answer to the meaning question for everyone, but it shows that Dawkins is at least engaging with it rather than dismissing it. The book also contains a sustained attack on what Dawkins calls “the lazy use of coincidence”, particularly as practised in astrology and paranormal belief. This is a precursor to The God Delusion in its concern with the cognitive mechanisms that generate false beliefs. Dawkins’s claim is precise: coincidences feel remarkable not because they are statistically improbable but because we observe them after they have already occurred and then selectively remember them. The feeling of uncanniness is a cognitive artefact, not evidence of anything beyond normal probability. The mechanism has two parts. Firstly, the denominator problem. When something striking happens, you notice the event but you do not notice the enormous number of unremarkable events that did not happen. You think of a friend and they call. This feels remarkable. What you do not account for is the number of times you thought of someone and they did not call, the number of times someone called and you had not been thinking of them, the number of days on which nothing coincidental happened at all. The remarkable event sits in a denominator so large that even genuinely improbable things become statistically inevitable across enough trials. Dawkins formalises this with the acronym PETWHAC, which stands for Population of Events That Would Have Appeared Coincidental. Any event you observe needs to be assessed not against the probability of that specific event but against the probability that something in the entire PETWHAC would have occurred. The PETWHAC is almost always enormous. Secondly, the media amplification problem. Modern media radically expands the effective PETWHAC without most people realising it. In a pre-media world, you encountered the experiences of perhaps a few hundred people across your lifetime. Remarkable coincidences in that pool were genuinely rare. In a world of television, newspapers and now social media, you are effectively sampling the experiences of millions of people daily. Given a large enough sample, any event with a one-in-a-million probability will happen roughly once per million trials. Across the experience of millions of people, one-in-a-million events are happening constantly and being reported selectively when they do. The person who had the coincidental experience reports it. The millions who did not have it do not report the absence. The result is a steady stream of apparently uncanny events that are in fact exactly what probability predicts. Dawkins uses the coincidence argument to target astrology and the paranormal, but the underlying cognitive error he is identifying is far more general and far more consequential than those targets suggest. The same error produces conspiracy theories. A conspiracy theory is fundamentally a pattern-detection exercise over a large PETWHAC. Given the enormous number of events occurring at any time, some of them will cluster in ways that look connected. The pattern is then extracted, the non-matching events are ignored and the cluster is treated as evidence of design. The cognitive machinery that makes coincidence feel meaningful is the same machinery that makes conspiracy feel plausible. The same error produces superstition in high-performing professionals. Athletes, traders and military commanders frequently develop elaborate rituals and causal beliefs based on sequences where a behaviour preceded a good outcome. The denominator of non-ritual occasions that also produced good outcomes, or ritual occasions that produced bad ones, is not tracked with the same rigour as the confirming sequence. The same error drives a large fraction of bad management decisions. A CEO implements a strategy and revenue grows. The strategy is credited. The background market conditions, the competitor errors, the lagged effects of prior decisions, the regression to the mean from a low baseline, none of these are visible in the same way that the strategy is visible. Narrative causation is extracted from a coincidental sequence and then institutionalised as received wisdom. Dawkins is describing what psychologists would later formalise as confirmation bias combined with the availability heuristic. The availability heuristic, developed by Kahneman and Tversky, holds that people estimate the probability of an event based on how easily examples come to mind. Coincidences come to mind easily because they are emotionally salient and because they get talked about. Non-coincidences are invisible because there is nothing to report. The availability of coincidental examples therefore dramatically inflates their estimated frequency and significance. Confirmation bias compounds this. Once a pattern is suspected, the mind begins tracking confirming instances and discounting or not noticing disconfirming ones. The person who begins to suspect they have psychic abilities starts noticing the hits and forgetting the misses. The sample is corrupted from the first moment the hypothesis is formed. What Dawkins adds to the psychological account is the evolutionary explanation for why this machinery exists at all. In the ancestral environment, the cost of false positives, detecting a pattern that does not exist, was low. The cost of false negatives, missing a real pattern such as the correlation between a rustle in the grass and a predator, was potentially fatal. Natural selection therefore built cognitive machinery heavily biased toward pattern detection. We are, as he puts it, a pattern-seeking species in a universe that contains a great deal of random noise. The mismatch between the pattern-seeking mind and the largely patternless universe is the source of almost every superstition ever invented. Before treating any coincidence or apparent pattern as meaningful, ask three questions. What is the full PETWHAC? How many events of this type could have occurred and been noticed? If the answer is very large, the improbability of the specific event is greatly reduced. What is the selection mechanism? Who is reporting this event and why? Reported coincidences are a biased sample of all coincidences, which are themselves a biased sample of all events. The selection pressure toward reporting the remarkable is enormous. What would falsify the pattern? If no possible observation could disconfirm the apparent connection, it is not a meaningful connection. It is a story imposed on noise. These three questions, applied consistently, dissolve the vast majority of what presents itself as coincidental or uncanny. What remains after applying them deserves serious attention. What fails the test does not, regardless of how compelling it feels. This argument sits at the intersection of several things already in your library. It is the epistemological complement to Taleb’s fat tails argument: Taleb says extreme events are more probable than you think, Dawkins says coincidences are less improbable than you think, and both errors derive from the same failure to reason correctly about base rates and sample sizes. It connects to Tetlock’s superforecasting work on calibration. The poorly calibrated forecaster is essentially committing the PETWHAC error in reverse, treating their confident predictions as evidence of skill without accounting for the base rate of correct predictions by chance across the full population of forecasters. It connects to Boyd’s OODA loop at the orientation stage. The pilot or commander who misreads noise as signal, who detects a pattern in random enemy movements and orients their decision cycle around it, has made exactly the cognitive error Dawkins is describing. The adversary who can generate apparently patterned noise thereby corrupts the opponent’s orientation without the opponent realizing it. The lazy use of coincidence is not a minor intellectual failing. It is one of the primary mechanisms by which intelligent people reason themselves into consistently wrong conclusions. The Ancestor’s Tale (2004): Evolution as Pilgrimage The Ancestor’s Tale is Dawkins’s most ambitious and encyclopedic work, a reverse pilgrimage from Homo sapiens back through time to the common ancestor of all life. It is structured by a series of “rendezvous” at which the human lineage meets the lineages of successively more distant relatives, from other primates through fish, insects, fungi and bacteria. Each rendezvous is an occasion to discuss some aspect of the biology, genetics or evolutionary history of the group concerned. This book’s function in the project is primarily evidential. The accumulated molecular phylogenetics, comparative anatomy and palaeontology presented across its six hundred pages constitutes a comprehensive brief for the fact of evolution. It is the most direct answer to creationism at the level of evidence rather than at the level of argument. The framing as a pilgrimage, borrowed from Chaucer, does something else: it makes each lineage equally central to the story, since every living species has survived every generation of its ancestry. This egalitarianism among lineages is itself a restatement of the replicator logic: no species is the goal of evolution; all are equally the current expression of billions of years of differential replication. The God Delusion (2006): The Prosecution The God Delusion is the culmination of the anti-religious thread that runs through the project from The Blind Watchmaker onward. Its central argument is that the existence of God is a scientific hypothesis like any other and that, assessed by the standards of scientific evidence, it is overwhelmingly unlikely to be true. Dawkins explicitly rejects the position of non-overlapping magisteria, associated with Stephen Jay Gould, which holds that science and religion address different questions and therefore cannot conflict. Dawkins’s position is that any God who intervenes in the physical world, answers prayers or performs miracles is making empirically testable claims and must be assessed accordingly. The book’s most original section, philosophically speaking, is the argument from improbability. Dawkins argues that any being complex enough to design the universe would itself require explanation; the God hypothesis therefore does not reduce the explanatory burden but increases it. This is the Blind Watchmaker argument applied not to biological organisms but to the universe as a whole. The Ultimate Boeing 747 gambit, as Dawkins names it (drawing on Fred Hoyle’s original framing), holds that the probability of God existing by chance is even lower than the probability of a Boeing 747 being assembled by a hurricane in a junkyard. The book also contains an extended treatment of the psychological and evolutionary origins of religious belief, drawing on Pascal Boyer’s cognitive anthropology and on Robert Hinde’s work on religion as a biological phenomenon. Religion, on this account, is a by-product of cognitive mechanisms that are adaptive in other contexts: the tendency to attribute agency to unexplained events, the tendency to believe what parents and elders say, the tendency to be susceptible to rule systems. The meme concept reappears here: religious ideas are memes that have been highly successful at propagating themselves, not because they are true but because they exploit pre-existing cognitive biases. The God Delusion is not Dawkins’s most scientifically original work but it is his most consequential in terms of public impact. It defined the terms of what became known as New Atheism and provoked a literature of response that occupied professional theologians, philosophers and scientists for a decade. The Greatest Show on Earth (2009): The Evidence for Evolution The Greatest Show on Earth is Dawkins’s most sustained engagement with the empirical case for evolution, written in explicit response to the creationist and intelligent design movements that had gained political traction, particularly in the United States. Where The Blind Watchmaker argued that Darwinism is the only adequate explanation for the appearance of design, this book argues that the fact of evolution is as well established as any fact in science. The book’s strongest sections are those dealing with molecular evidence: the distribution of endogenous retroviruses across primate genomes, which provides an almost forensically precise record of shared ancestry, and the evidence from molecular clocks for the timing of divergences. Dawkins also gives extended treatment to the direct evidence from artificial selection, using dog breeds, domesticated crops and laboratory experiments in bacterial evolution to demonstrate that heritable variation combined with selective pressure produces rapid and substantial change. The book sits slightly outside the central theoretical architecture of the project in that it does not advance the gene’s-eye view or the replicator logic in any new direction. Its function is prosecutorial: to make it impossible for a good-faith reader to doubt the fact of evolution. That it needed to be written, in Dawkins’s view, is itself a cultural and political datum that his later career has been organised around. The Unifying Architecture The single most important insight threading through the entire Dawkins project is this: replication with heritable variation and differential copying success is a process that, given time, necessarily produces the appearance of purposeful design, and this process requires no mind, no foresight and no goal. Everything else in the body of work is either a derivation from this principle, a defence of it or an application of it. Four ideas recur across the major works and are worth identifying precisely because they are not independent. First, the replicator as the proper unit of Darwinian analysis. This is established in The Selfish Gene, hardened in The Extended Phenotype and deployed in every subsequent book. The gene is not the only possible replicator (the meme is proposed as a cultural analogue) but the gene is the replicator whose properties explain biological evolution. Second, cumulative versus single-step selection. This distinction, introduced forcefully in The Blind Watchmaker and extended in Climbing Mount Improbable, is the answer to every argument from improbability: complexity does not arise in a single step and therefore cannot be assessed by the probability of its arising in a single step. Third, the argument from design inverted. Paley’s argument that complexity implies a designer is correct in its observation but catastrophically wrong in its conclusion, because a designer complex enough to produce biological organisms would itself require explanation, and natural selection provides that explanation without invoking any further intelligence. This argument appears in The Blind Watchmaker, is restated in River Out of Eden and is the philosophical engine of The God Delusion. Fourth, the by-product account of apparently purposive or meaningful phenomena. Religion, aesthetic experience, moral intuitions and the feeling of cosmic significance are not evidence of design or transcendence; they are the products of cognitive and cultural replication processes that have been selected for other reasons. Unweaving the Rainbow and The God Delusion develop this most explicitly. These four ideas are not independent. The replicator logic generates the cumulative selection argument. The cumulative selection argument defeats the argument from design. The defeat of the argument from design removes the most powerful non-evidential support for religious belief. The by-product account then explains why religious belief persists despite the defeat of its best argument. The project is, at its deepest level, a single sustained argument conducted across fifty years and a dozen books. The Criticisms The criticisms of Dawkins’s work come from several distinct directions and vary considerably in their force. They deserve disaggregated treatment. The most philosophically serious internal criticism comes from developmental biologists and systems biologists who argue that the gene’s-eye view systematically misrepresents how development works. Eva Jablonka and Marion Lamb, in Evolution in Four Dimensions (2005), argue that epigenetic inheritance, behavioural inheritance and symbolic inheritance are genuine channels of heritable variation that the gene-centric model cannot accommodate. Denis Noble, in The Music of Life (2006) and subsequent papers, argues from a systems perspective that genes do not “cause” phenotypes in any straightforward sense; the developmental system as a whole determines phenotypic outcomes, and treating genes as the active agents while treating everything else as passive vehicles is a rhetorical distortion of the biology. This is not a trivial objection. Dawkins’s response has generally been that this confuses the question of developmental mechanism (how organisms are built) with the question of evolutionary agency (what selection tracks), and that the gene’s-eye view is a claim about the latter, not the former. This response has some force but it does not fully neutralise Noble’s systems argument, which is precisely that you cannot cleanly separate developmental and evolutionary questions. Stephen Jay Gould mounted the most sustained scientific attack from within evolutionary biology. With Lewontin, in the “Spandrels of San Marco” paper (1979), he argued that Dawkins’s adaptationism attributed adaptive significance to traits that might be architectural by-products, developmental constraints or the results of drift. With Niles Eldredge, in the work on punctuated equilibrium, he argued that the tempo and mode of evolution as revealed by the fossil record did not fit the smooth gradationism implied by the Dawkins-Williams gene-selection model. Dawkins’s response to the spandrels argument, developed in The Extended Phenotype, is that Gould and Lewontin had erected a strawman adaptationist who actually believes every trait has an adaptive explanation, whereas serious Darwinians have always acknowledged the role of constraint and by-product. On punctuated equilibrium Dawkins’s position is that the interesting question is what mechanism generates the pattern and that punctuated equilibrium describes a phenomenon without explaining it. Neither side of this exchange comes off entirely well: Dawkins sometimes sounds more dismissive of Gould than the scientific dispute warrants, and the animosity between them became personal enough to compromise both men’s ability to engage with each other’s strongest arguments. The meme concept has attracted persistent and largely justified criticism from cognitive scientists and cultural theorists. Susan Blackmore, in The Meme Machine (1999), extended the concept sympathetically, but Jerry Fodor, Dan Sperber and others argued that the concept is too vague to do scientific work. What exactly is a meme? Is it a neural pattern, a behaviour, an artefact, or a type that can be instantiated in any of these? How is transmission measured? What is the meme’s equivalent of a copying mechanism? Dawkins himself introduced the concept speculatively and has never developed it into a rigorous research programme. The meme remains more metaphor than mechanism, and the field of cultural evolution has largely moved toward more precisely specified models, such as those developed by Robert Boyd and Peter Richerson, that do not use the meme vocabulary. The criticisms of The God Delusion from philosophers and theologians are substantial and cannot be dismissed as mere special pleading. Alvin Plantinga, in Where the Conflict Really Lies (2011), argued that Dawkins’s central argument from improbability commits a basic philosophical error: it treats God as the kind of entity whose complexity is explicable or inexplicable on the same terms as physical complexity, when theologians have consistently argued that God is not a complex physical entity but a simple, non-physical one. Whether one accepts that theological move or not, the point is that Dawkins’s argument engages with popular religion rather than with sophisticated theology, and this is a genuine limitation. Terry Eagleton’s review in the London Review of Books made a similar point more sharply, observing that Dawkins addresses a kind of naive theism that most serious theologians would not recognise. Dawkins’s reply, that most actual religious believers hold something closer to naive theism than to the sophisticated positions of academic theology, has sociological plausibility but does not constitute an engagement with the strongest version of the opposing argument. Finally, Dawkins’s own rhetorical choices have worked against some of his stated goals. He has said repeatedly that one of his central aims is to raise the consciousness of people who do not realise they are atheists and to make atheism publicly acceptable. The combative, often contemptuous tone of The God Delusion and much of his journalism and social media output has demonstrably alienated people who might have been persuaded by a more charitable engagement with the psychological and social functions of religious belief. The science communication literature consistently shows that confrontational messaging tends to produce identity-protective reasoning rather than attitude change. A writer whose central interest is in how minds form beliefs, and who has written with genuine insight about cognitive biases in Unweaving the Rainbow and elsewhere, has been oddly resistant to applying that insight to his own public strategy. Why It Matters Dawkins’s influence on how educated people in the English-speaking world understand evolution is difficult to overstate. The Selfish Gene changed the default vocabulary of evolutionary biology: “meme”, “selfish gene”, “extended phenotype” and “arms race” are now standard terms, and the gene’s-eye framing is now the default stance of most working evolutionary biologists even when they disagree with Dawkins on specific questions. The book is among a handful of works in twentieth-century science writing that genuinely changed how a scientific community thinks about its own field, not just how the public understands it. The broader cultural impact is harder to assess but real. The Blind Watchmaker made the Darwinian answer to natural theology clear and accessible at a moment when creationism was mounting a serious political challenge, particularly in the United States. The God Delusion catalysed, along with the contemporaneous work of Christopher Hitchens, Sam Harris and Daniel Dennett, a period of unusually public and sustained engagement between scientific and religious worldviews that has altered the landscape of public discourse about belief. What the project leaves open is considerable. The meme concept promised a unified theory of cultural evolution that has not materialised in the form Dawkins suggested. The question of whether the gene’s-eye view can fully accommodate the findings of developmental systems biology remains genuinely contested. The deeper question of meaning, of what human beings are to make of a universe in which they are the products of a blind replicator process, is addressed in Unweaving the Rainbow and elsewhere with more passion than rigour. Dawkins is confident that scientific understanding is sufficient for a meaningful life but the argument for this, as opposed to the celebration of it, is thinner than the arguments he makes in his areas of greatest technical competence. What makes the project unusual in the history of its field is the combination of theoretical originality, synthetic ambition and rhetorical power. The gene’s-eye view was not Dawkins’s original discovery; it was Hamilton’s and Williams’s. But the intellectual architecture Dawkins built around it, connecting replicator logic to the philosophy of design, to the cognitive science of religion and to the public understanding of science, is genuinely his own. The project is a demonstration that a single powerful idea, properly understood and relentlessly applied, can reorganise a very large amount of intellectual territory. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

12. 

    12 Jun 2026

    On Geopolitics

    Russia has no warm water port that is not blocked by a potential adversary. That single geographical fact explains five centuries of Russian foreign policy more accurately than any account built around ideology, personality or culture. The Tsars pushed outward. The Soviets pushed outward. Putin pushed outward. The names change. The map does not. If you want to understand why nations do what they do, start with the map. There is a habit of mind that distinguishes the serious geopolitical analyst from the commentator who merely follows events. The commentator sees what happened. The analyst sees why it was probably going to happen given the structural forces at work and what is likely to happen next given those same forces. The difference is not access to better information. It is the possession of frameworks, mental models of how power, geography, history and human nature interact, that organise raw information into strategic understanding. The books I referred for this piece collectively build those frameworks. Not the frameworks of any single ideological tradition but the overlapping and sometimes contradictory frameworks that together produce the most accurate available picture of how the world actually works. What follows is what they collectively teach. The Permanent Layer: Geography as Destiny’s First Draft Begin where every serious geopolitical analysis must begin. Before ideology, before economics, before the decisions of leaders and the accidents of history, there is geography. The shape of the land, the location of rivers, the presence or absence of natural harbours, the distance from warm-water ports, the defensibility of borders and the location of energy and mineral resources are not variables that change on human timescales. They are the permanent constraints within which all political choice occurs and they have been shaping the behaviour of states for as long as states have existed. Marshall’s Prisoners of Geography makes this argument with unusual clarity by walking through ten maps that explain more about global politics than most diplomatic history. Russia’s behaviour makes complete sense once you understand that it has no warm-water port that is not blocked by a potential adversary, that the flat plain of Eastern Europe has been the invasion route into the Russian heartland repeatedly throughout history and that every Russian leader from the Tsars to Putin has therefore been structurally compelled to push outward and create buffer zones regardless of their ideology or personal character. The invasion of Ukraine is not primarily an expression of Putin’s personality. It is the latest iteration of a strategic logic that Russian geography has imposed on Russian rulers for five hundred years. China’s behaviour in the South China Sea makes complete sense once you understand that China’s coastline is a series of island chains controlled by American allies that function as a containment barrier limiting Chinese naval projection into the Pacific. The artificial islands China is building, the aggressive posture in the Taiwan Strait and the Belt and Road infrastructure across Central Asia are all attempts to break out of a geographical containment that American strategic planners deliberately constructed and that Chinese strategic planners have been trying to dismantle with equal deliberation. The United States’ behaviour as a global power makes complete sense once you understand that it is the only major power in human history to be bordered on two sides by vast oceans and on two sides by weak and friendly neighbours. This geographical fortune allowed the United States to develop continental-scale economic power without the permanent military mobilisation that continental powers require for survival and then to project that power globally from a position of domestic security that no Eurasian power has ever enjoyed. American foreign policy exceptionalism is not primarily a cultural or ideological phenomenon. It is the political expression of an exceptional geographical position. Kaplan extends this analysis into the zones of perpetual instability. The Middle East’s political dysfunction is inseparable from its geography: a region of scarce water, vast oil wealth unevenly distributed, no natural borders that correspond to ethnic or religious communities and surrounded by the competing interests of external powers for whom its resources and location are strategically critical. The states drawn by colonial powers across this geography were not designed for political viability. They were designed for colonial administrative convenience. The instability that has followed is the predictable consequence of geography and political structure being in fundamental tension. The Himalayan massif and the geography of the Indian subcontinent explain the dynamics of the India-Pakistan-China strategic triangle with equal force. India’s geography gives it the largest natural defensible perimeter in the world and the strategic depth that allows it to absorb pressure while building response capability. It also creates structural competition with Pakistan over Kashmir, which sits astride the water systems on which both nations depend, and with China over the Himalayan border that defines the boundary between the world’s two most populous nations. These are not conflicts produced by misunderstanding or bad faith. They are the structural outputs of geography interacting with demography and resource distribution. The framework that geography provides is not deterministic. Leaders make choices and choices matter. But they make choices within a geographical constraint set that powerfully shapes what options are available and what costs attach to each option. The leader who ignores geography pays the price consistently. Napoleon ignored the Russian winter. Hitler ignored it again. Every American president who has tried to impose a political solution on Afghanistan has ignored the geographical reality that makes it unconquerable and ungovernable from outside. The Structural Layer: How the International System Shapes State Behaviour Mearsheimer’s offensive realism is the most intellectually uncomfortable framework in this body of work for anyone who believes that international relations are primarily driven by values, institutions or the quality of leadership. His argument is structural: the behaviour of great powers is not primarily determined by their domestic political systems, their ideologies, their cultures or the intentions of their leaders. It is determined by the structure of the international system in which they operate. The international system is anarchic in the technical sense: there is no world government, no global sovereign with the legitimate authority and physical capacity to enforce agreements and protect weaker parties against stronger ones. In this environment, every state must ultimately rely on its own power for survival. No treaty, no alliance and no international institution can provide the absolute guarantee of security that a domestic government provides to its citizens because there is no enforcement mechanism that operates above the level of national power. In this anarchic system, relative power matters more than absolute wellbeing. A state that grows richer while a rival grows richer faster has become relatively weaker regardless of its absolute improvement. A state that disarms in the name of peace has made itself more vulnerable regardless of the sincerity of its partner’s stated intentions. The rational response to this structural reality is the continuous pursuit of relative power advantage and the maintenance of enough military capability to deter or defeat any plausible combination of adversaries. This logic produces the behaviour that idealists attribute to greed, aggression or the specific character flaws of specific leaders. The United States did not seek to dominate the Western hemisphere because Americans are uniquely aggressive. It did so because geographical security requires preventing any other great power from establishing a significant presence in your hemisphere from which it could threaten you. China is not seeking to dominate Asia because the Chinese Communist Party is uniquely expansionist. It is doing so because a great power in an anarchic system rationally seeks to establish regional hegemony as the foundation of its security. These are the same strategic logic expressed by different powers in different historical moments. Mearsheimer’s framework predicts that the US-China competition will intensify regardless of the ideology of either party, the quality of bilateral diplomacy or the density of economic interdependence. The United States will attempt to prevent China from achieving regional hegemony in Asia for the same structural reasons it has prevented any other power from achieving regional hegemony in any region adjacent to its interests throughout its history. China will continue to push for regional hegemony for the same structural reasons that every rising great power has sought it throughout history. This is not pessimism or warmongering. It is the structural forecast that the international system’s architecture produces. The critical question, which Allison’s Thucydides Trap analysis addresses directly, is whether the structural logic of great power competition produces war or whether it can be managed through the diplomacy, institution-building and strategic restraint that has occasionally prevented it. His historical survey of sixteen cases in which a rising power threatened to displace a ruling power finds that war resulted in twelve of them. The four exceptions are not random. They share specific characteristics: deliberate diplomatic management of the transition, the creation of institutions that gave both parties stakes in the existing order and the exercise of strategic restraint by both sides at critical moments of tension. The US-China case has some features that make it more manageable than historical analogies suggest and some that make it more dangerous. The existence of nuclear weapons on both sides creates a mutual deterrence that significantly raises the cost of direct military conflict. The density of economic interdependence creates mutual vulnerabilities that provide both sides with incentives to avoid the economic catastrophe that war would produce. But the speed of China’s rise, the ideological distance between the two systems, the specific flashpoints of Taiwan and the South China Sea and the domestic political pressures in both countries that make strategic restraint politically costly all push in the opposite direction. The Civilisational Layer: Culture as Strategic Force Huntington’s Clash of Civilisations is the most debated and most misread book in the geopolitics canon. It is read as a prediction of inevitable conflict between civilisations, which is not its argument. It is an argument about the primary fault lines along which conflict will organise in the post-Cold War world, replacing the ideological fault line of the Cold War that had previously provided the organising principle of international conflict. His central observation is that the universalist claim of Western liberal democracy, the belief that Western values of individual rights, democratic governance and market economics represent the natural destination of all human societies once they develop sufficiently, is not shared by the other major civilisations and that the attempt to impose it produces resistance that is civilisational in character rather than merely political. The Islamic world, the Confucian world, the Orthodox world and the Hindu world each carry distinct civilisational identities, historical memories and political values that are not converging toward the Western model and that generate genuine resistance to Western universalism that cannot be dissolved by economic development or better communication. The practical consequence for strategic analysis is that the post-Cold War framework of liberal internationalism, which assumed that expanding the zone of democratic governance and free markets would produce a convergent world order of mutually beneficial cooperation, consistently misread resistance to this project as the product of bad governments, elite manipulation or development failures rather than as the expression of genuine civilisational differences in what constitutes legitimate political order. Nisbett’s Geography of Thought adds the deepest layer to this argument. The differences between East Asian and Western modes of thinking are not superficial cultural preferences but fundamental differences in cognitive style that have been shaped by thousands of years of different social organisation, philosophical tradition and ecological context. East Asian thinking tends to be contextual, relational and attentive to the field within which objects exist. Western thinking tends to be analytical, categorical and focused on the object rather than its context. These differences produce genuinely different perceptions of the same situation, not just different interpretations of the same perception. For geopolitics, this means that when Western analysts assess Chinese strategic intentions using Western analytical frameworks they are not just using different vocabulary. They are operating on a different cognitive architecture that may systematically misread what China is doing and why. Kissinger’s On China is the most powerful expression of this insight from a practitioner’s perspective. His account of Chinese strategic thinking, rooted in the tradition of wei qi rather than chess, oriented toward patient positional accumulation rather than decisive battle, comfortable with centuries-long time horizons and deeply rooted in the historical memory of China as the natural centre of the world order, provides a framework for reading Chinese behaviour that no amount of Western strategic theory generates. The wei qi analogy is worth dwelling on. Chess is a game of decisive confrontation in which pieces are eliminated from the board and the game ends with a winner who has destroyed the opponent’s capacity to continue. Wei qi is a game of positional encirclement in which the objective is to surround territory and accumulate positional advantage incrementally over a very long game in which no single move is decisive and in which the overall position matters more than any local battle. The strategic culture that produced chess and the strategic culture that produced wei qi approach competition differently at the deepest level and the collision of these two strategic cultures is one of the defining features of the current geopolitical moment. The Historical Layer: How We Got Here Frankopan’s Silk Roads performs the most important single act of historical reorientation available to a geopolitical analyst: it moves the centre of gravity of world history from the Atlantic to Central Asia where it actually belonged for most of recorded history. The history most educated Westerners carry is a story centred on Greece, Rome, Western Europe and the Atlantic world with Asia as a peripheral backdrop that occasionally intrudes through the Mongol invasions or the Silk Road trade that supplied Western luxury goods. This is a provincialism masquerading as universal history and it produces systematic blind spots in geopolitical analysis. The historical centre of the world for most of the past three millennia was the zone stretching from the Eastern Mediterranean through Persia and Central Asia to China. The trade routes across this zone, the Silk Roads in their many forms, were the arteries of global economic activity, technological diffusion and cultural exchange that made the ancient and medieval world function. Western Europe’s rise to global dominance from the 15th century onward was not the natural culmination of a superior civilisation. It was partly the product of specific geographical advantages, Atlantic access that allowed circumnavigation of the Eurasian land mass, partly the product of specific technological accidents, gunpowder and naval technology, and partly the product of the ecological catastrophe of the Black Death which reorganised European institutions in ways that eventually produced the scientific revolution and industrial capitalism. The Asian civilisations that Western powers displaced were not primitive or stagnant. They were sophisticated systems disrupted by a specific historical accident that temporarily shifted the balance of power in ways that are now reversing. Khanna’s Future Is Asian and Connectography extend this historical reorientation into the present and near future. Asia already accounts for the majority of global economic output, population and growth. The institutional architecture of Asian economic integration, the Regional Comprehensive Economic Partnership, the Asian Infrastructure Investment Bank, the Belt and Road Initiative and the web of bilateral relationships that China has built across Africa, the Middle East and Southeast Asia, is creating a parallel international order that does not depend on Western institutions for its function and that is increasingly capable of setting the terms of engagement rather than accepting terms set by Washington or Brussels. Connectography’s central argument is that the primary architecture of global power is shifting from territorial control to connectivity infrastructure. The nation that controls the cables, pipelines, shipping routes and supply chain nodes through which global economic activity flows has more practical power over other nations than the nation that controls the most territory within its borders. China’s Belt and Road Initiative is best understood not as a development aid programme or even as a debt trap but as the deliberate construction of a connectivity architecture that places China at the centre of the flows on which other nations depend, replicating the geostrategic advantage that control of sea lanes gave Britain in the 19th century. The Institutional Layer: Why Some Nations Succeed and Others Fail Acemoglu and Robinson’s Why Nations Fail provides the analytical framework that explains the variance in national outcomes that geography alone cannot account for. Two nations with similar geographies and similar resource endowments frequently diverge dramatically in their development trajectories and the explanation lies in the character of their institutions: the rules, enforcement mechanisms and organisational structures that determine who captures the gains from economic activity. Inclusive institutions, those that distribute political power broadly, protect property rights for all citizens, enforce contracts reliably and create conditions in which anyone can participate in economic activity and capture the returns from their effort and investment, generate sustained economic development because they align individual incentives with productive activity. Extractive institutions, those that concentrate political power in the hands of a narrow elite who use it to extract wealth from the rest of the population, generate short-term wealth for the elite but long-term stagnation for the society because they systematically destroy the incentive for broad-based productive activity. The geopolitical dimension of this framework is that extractive institutions are self-reinforcing and extremely difficult to change from within. The elite who control extractive institutions benefit enormously from maintaining them and have both the incentive and the power to resist institutional change that would distribute power more broadly. This produces the vicious cycle of extractive development: poor institutions produce poor economic outcomes which produce the political instability that allows elites to maintain extractive institutions that produce poor outcomes. Breaking this cycle requires shocks large enough to disrupt the existing power equilibrium, which historically has meant wars, revolutions, colonisation or the specific historical accident of a coalition broad enough to impose inclusive institutional change against elite resistance. The implication for international relations is that the assumption underlying much development aid and democracy promotion, that poor countries are poor because their governments lack knowledge of better policies and will adopt better policies once those policies are explained and resources are provided, is wrong about the mechanism. Poor countries are frequently poor not because their governments lack knowledge but because their governing elites benefit from the institutions that keep them poor. Providing resources to extractive states does not produce development. It provides more resources for elite extraction. Changing institutional character requires engaging with the political economy of who benefits from existing institutions and what would need to change for those parties to accept or be forced into institutional reform. Fukuyama’s End of History argument, properly understood, is not the triumphalist claim it is frequently caricatured as. He is not claiming that history has ended in the sense that nothing further will happen. He is claiming that the ideological competition among alternative forms of political and economic organisation has been resolved in favour of liberal democracy and market capitalism because no other system has demonstrated comparable capacity to generate prosperity and legitimate governance. The subsequent decades of nationalist resurgence, Chinese authoritarian capitalism and Islamic political movements have tested this claim without definitively refuting it, because none of the alternatives has demonstrated the superior developmental outcomes that would constitute an ideological refutation. Kagan’s World America Made provides the structural complement to Fukuyama’s ideological argument. The liberal international order that has produced the longest period of great power peace and the greatest expansion of prosperity in human history is not a natural equilibrium. It is a constructed order maintained by American power and the American willingness to bear the costs of maintaining it. When American power declines or American willingness to bear those costs diminishes, the order does not maintain itself. It begins to erode in ways that are first visible at the margins and then at the centre. The erosion already underway in every region simultaneously is not the product of specific policy failures. It is the structural consequence of the gap between the order’s maintenance requirements and the power and will available to meet them. The Strategic Practice Layer: How Great Powers Actually Behave Kissinger’s Diplomacy is the most comprehensive practitioner’s account of how the modern state system was built and how it actually functions as opposed to how international relations textbooks suggest it functions. His central argument is that the European balance of power system, which produced the most stable period of great power relations in modern history between 1815 and 1914, worked not because its participants were more moral than their predecessors but because they shared a common understanding that the system’s stability depended on preventing any single power from achieving hegemony and that all powers had interests in maintaining the system even at the cost of specific national preferences. The Congress of Vienna architects, Metternich above all, understood that the post-Napoleonic settlement would endure only if every major power felt it had a stake in the order rather than a grievance to be redressed. The inclusion of defeated France as a legitimate participant in the settlement rather than its punishment and exclusion was not generosity. It was the strategic recognition that an aggrieved France would be a revisionist power committed to overturning the settlement at the first opportunity. The same insight, applied in reverse, explains why the Versailles settlement after the First World War produced the catastrophe that the Vienna settlement had avoided. The practical lessons Kissinger draws from this history are directly applicable to the current period. The United States has consistently made the Versailles error of treating adversaries as inherently illegitimate rather than the Vienna error of integrating rivals into a system in which they have stakes worth preserving. The liberal internationalist tendency to frame every conflict as a struggle between good and evil, democracy and autocracy, the rules-based order and its enemies, produces rhetorical clarity at the cost of strategic flexibility. It makes compromise look like appeasement, makes integrating rivals look like legitimising their behaviour and consistently forecloses the diplomatic options that might produce stable settlements at lower cost than the confrontations that the moralistic framing makes inevitable. Pillsbury’s Hundred-Year Marathon introduces the most important single revision to the Western understanding of China’s strategic intentions. His argument, developed from decades of work as a China analyst for the US intelligence community, is that Chinese strategic planning has been oriented since at least 1949 toward the long-term goal of displacing American hegemony and restoring China to what it considers its natural position as the world’s leading power. The marathon metaphor reflects the Chinese strategic orientation toward patience: a hundred-year strategy executed through incremental positioning, the concealment of capability and intent and the exploitation of American openness and good faith to access technology, capital and market access that accelerated Chinese development. The strategic implication is that the period of engagement, in which Western powers assumed that integrating China into the international economic system would gradually liberalise its political system and align its interests with the existing order, was not just a miscalculation. It was a period in which the intended beneficiary of the liberal order systematically used its access to that order to build the capability to challenge it. This does not mean engagement was wrong as a strategy. It means it was executed without the strategic clarity about Chinese intentions that would have allowed its terms to be set more advantageously. The Resource and Environmental Layer: The Material Constraints of Power Brannen’s Ends of the World adds the deepest time horizon available to geopolitical analysis: the geological and biological record of what happens to complex systems when they exceed the carrying capacity of their environment. His account of the five mass extinction events in Earth’s history is not a climate change polemic. It is a systems analysis of how civilisational complexity interacts with resource constraints in ways that produce collapse rather than gradual adjustment. The geopolitical dimension is that the resource and environmental constraints now facing the global system are not distributable evenly across all nations. They fall disproportionately on the nations least responsible for producing them and least equipped to manage their consequences. Water scarcity, driven by climate change and aquifer depletion, is already the primary driver of political instability across the arc from North Africa through the Middle East to South Asia. The combination of population growth, agricultural stress and state fragility in this arc is producing the migration pressures, conflict dynamics and governance failures that are reshaping the politics of every adjacent region. The nations that control water, food production capacity and the mineral resources required for the energy transition, lithium, cobalt, rare earth elements and others, are accumulating a form of strategic leverage that operates on a longer time horizon than military power but that is ultimately more fundamental. China’s deliberate acquisition of mineral resource access across Africa and Central Asia, its control of rare earth processing capacity and its early investment in renewable energy manufacturing are not primarily commercial decisions. They are strategic positioning for the resource competition that will define the middle decades of this century. The Unified Picture Pull all of this together and what emerges is a coherent and demanding picture of the world that most daily news coverage systematically obscures. The permanent layer is geography. It sets the constraints within which all other factors operate. Nations that ignore geographical constraints in their strategic planning consistently pay the price. Nations that design their strategy around geographical realities consistently outperform those that design it around ideological preferences. The structural layer is the international system’s anarchy. In the absence of a global sovereign, great powers pursue relative power advantage regardless of their stated values. The behaviour of states is therefore more predictable from their structural position than from their ideology, culture or the intentions of their specific leaders. This does not make those factors irrelevant. It means they operate within a structural constraint that is prior to all of them. The civilisational layer is the genuine diversity of political values, historical memories and cognitive styles across the major world civilisations. Western liberal universalism is one civilisational project among several and its claim to represent the natural destination of all human political development is not shared by the other major civilisations and generates genuine resistance that cannot be dissolved by better communication or more development assistance. The historical layer is the reorientation required to see the current period accurately. The Western dominance of the past five centuries is not the natural order of things but a specific historical episode produced by specific circumstances that are now reversing. The return of Asian centrality is not a threat to be resisted but a structural shift to be understood and navigated. The institutional layer explains the variance in national outcomes that geography and structure cannot account for. Inclusive institutions produce development and legitimate governance. Extractive institutions produce stagnation and instability. And the political economy of who benefits from existing institutions determines whether they can be changed without catastrophic disruption. The strategic practice layer is how all the above is translated into specific decisions by specific leaders in specific moments. The quality of that translation, the ability to see the structural forces clearly, to identify the genuine interests of all parties rather than their stated positions and to design settlements that give all parties stakes in the order rather than grievances against it, is what distinguishes strategic statecraft from the moralistic performance that substitutes for it in most democratic foreign policy discourse. And the resource and environmental layer is the emerging constraint that will reshape all the above over the coming decades as the material foundations of the current order come under pressure in ways that conventional geopolitical analysis is only beginning to incorporate. What This Demands of You The synthesis of this body of work makes a specific demand on anyone operating at the intersection of national policy and global affairs. It demands the consistent practice of what Kissinger called seeing the world as it is rather than as you wish it were, combined with the simultaneous understanding that the world as it is contains genuine moral dimensions that cannot be dissolved into pure power analysis without producing a strategic culture that is both ethically corrupt and strategically self-defeating. The practitioner who has genuinely absorbed this body of work approaches every geopolitical question with a set of prior questions that most analysts never ask. What does geography dictate here regardless of political preferences? What does the structure of the international system make likely regardless of stated intentions? What civilisational differences in values and cognitive style are producing misunderstandings that look like bad faith but are actually genuine perception differences? What historical memories and grievances are operating beneath the surface of current positions? What institutional incentive structures determine who benefits from the current situation and who therefore has stakes in changing it? The analyst who asks these questions consistently will be wrong sometimes and right more often than the analyst who reads current events through the lens of their own civilisational assumptions, who treats stated intentions as reliable guides to future behaviour and who mistakes the temporary for the structural. Geopolitics rewards the long view, the structural view and the intellectually humble view that the world is more complex, more driven by permanent forces and less amenable to idealistic transformation than any single moment’s dominant narrative suggests. The nations and individuals who operate from that understanding consistently outperform those who do not, not because they are more powerful but because they are more accurate about the terrain they are moving through. That accuracy, built from the frameworks this body of work provides, is the highest-leverage investment available to anyone whose decisions touch the lives of nations. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

13. 

    12 Jun 2026

    On System Thinking

    Kodak invented the digital camera in 1975. It had the patents, the engineers and the capital. It filed for bankruptcy in 2012. Nobody at Kodak was stupid. Nobody was asleep. The system was working exactly as designed, which is precisely why it failed. That is not a business story. That is a systems story and once you see it you cannot stop seeing it everywhere. There is a mode of thinking that is so rare in practice and so powerful in application that those who develop it operate in most professional and intellectual environments as though they have access to information others do not. They are not smarter in the conventional sense. They have not necessarily read more or worked harder. They have learned to see something that most people cannot see: the structure underneath events. Most people think in events. Something happened. Someone caused it. Fix the cause or punish the person and the problem is resolved. This mode of thinking is intuitive, narratively satisfying and consistently wrong in complex environments. Events are the visible surface of a deeper reality. Beneath the events are patterns of behaviour recurring over time. Beneath the patterns are the systemic structures, the feedback loops, the stocks and flows, the delays and the incentive architectures that produce the patterns. The event is the symptom. The structure is the disease. Treating symptoms while the structure remains unchanged produces the same outcomes repeatedly with the reliable accompaniment of genuine bewilderment about why the problem keeps returning. The books I referred for writing this piece collectively build the complete mental architecture for seeing structure rather than just events. What follows is what they collectively teach. The Grammar of All Systems Meadows provides the foundational vocabulary without which everything else in this domain remains impressionistic. Every system, from a thermostat to a global economy to a human body to a political institution, is built from three basic components: stocks, flows and feedback loops. Understanding these three components and how they interact is the grammar of systems thinking in the same way that understanding words, sentences and grammar is the prerequisite for reading any language. A stock is any quantity that accumulates or depletes over time. Water in a bathtub. Money in a bank account. Trust in a relationship. Knowledge in a field. Pollution in an atmosphere. Population in a country. Stocks are the state of the system at any given moment and they change only through flows. A flow is the rate at which a stock changes. The tap filling the bathtub and the drain emptying it are both flows. Birth rate and death rate are the flows that determine population stock. Investment and depreciation are the flows that determine capital stock. The crucial and consistently underappreciated property of stocks is that they cannot change instantaneously. You cannot drain a bathtub in a moment regardless of how wide you open the drain. You cannot rebuild trust in a relationship overnight regardless of how sincere your effort. You cannot decarbonise an economy in a year regardless of the strength of the political will. The inertia of stocks is one of the primary sources of the delays that make systems so difficult to manage intuitively. Feedback Loops are the mechanisms through which a system responds to its own state. A reinforcing feedback loop amplifies change in the direction it is already moving: more investment produces more capital which produces more return which enables more investment. Compound growth is a reinforcing loop. So is compound decay. So is the escalation dynamic in an arms race. So is the virality of a successful social media post. Reinforcing Loops produce exponential behaviour, both the spectacular growth curves that excite investors and the runaway collapse curves that precede systemic failure. A Balancing Feedback Loop resists change and tries to maintain the system at a goal or equilibrium: a thermostat that turns on heating when temperature falls and off when it rises, a predator population that grows when prey is abundant and shrinks when prey becomes scarce, a central bank that raises interest rates when inflation exceeds target. Balancing loops are the stabilising mechanisms of systems and their absence is one of the signatures of systems heading toward collapse. The interaction between reinforcing and balancing loops, mediated by the delays inherent in stock dynamics, produces the full complexity of system behaviour. Oscillation, overshoot, collapse, growth, stagnation and resilience are all emergent properties of these three basic components in different configurations. This is the profound simplicity at the heart of systems thinking: an enormous diversity of complex behaviours produced by a small number of structural elements interacting in different combinations. The Pathologies: How Systems Fail Meadows identifies a set of recurring system archetypes, structural configurations that produce characteristic failure patterns across wildly different domains. Recognising these archetypes is one of the highest leverage skills in applied systems thinking because the structural solution to each archetype is generalisable across every domain in which it appears. The Tragedy of the Commons is perhaps the most consequential. When multiple actors share access to a common resource whose stock is finite and whose depletion is not individually costly to each actor in proportion to their contribution to it, the individually rational behaviour of each actor produces collective destruction of the resource. Every actor’s incentive is to extract as much as possible before others do the same. The aggregate result of individually rational decisions is collectively catastrophic. This structure underlies overfishing, groundwater depletion, atmospheric carbon loading, antibiotic resistance and the degradation of shared institutional resources including democratic norms. The structural solution in every case is the same: either privatise the commons so that individual actors bear the full cost of their extraction or regulate access through collective governance mechanisms strong enough to enforce restraint. Neither solution is comfortable. The alternative is the collapse of the commons. Fixes that Fail is the archetype that explains why so many well-intentioned policy interventions make problems worse over time. A symptomatic fix is applied to a problem. The fix provides immediate relief. The relief reduces the pressure to address the fundamental cause. The fundamental cause continues to operate and eventually reasserts the symptom, often in a more severe form. More symptomatic fix is applied. The fundamental cause becomes more entrenched. The system becomes progressively more dependent on the symptomatic fix while the underlying problem deepens. Flood control infrastructure that makes floodplain development seem safe until a flood overwhelms the infrastructure. Antibiotic use that relieves immediate symptoms while accelerating the development of resistant strains. Bailouts of financial institutions that relieve immediate crisis while reinforcing the too-big-to-fail dynamic that created it. The structural signature of this archetype is the same everywhere: short-term relief, long-term deepening of the fundamental problem. Shifting the Burden is the related archetype in which the symptomatic fix not only fails to address the fundamental cause but actively atrophies the system’s capacity to address it. The person who uses alcohol to manage anxiety reduces the anxiety in the short term but also reduces the development of the intrinsic anxiety management capacity that would solve the problem fundamentally. The organisation that uses consultants to solve strategic problems reduces the short-term pressure but also prevents the development of the internal strategic thinking capacity that would solve problems fundamentally. The dependency that develops on the symptomatic fix makes addressing the fundamental cause progressively harder. Dekker’s Drift into Failure provides the most important account of how these pathologies produce catastrophic outcomes in complex engineered and organisational systems. His central argument is that major accidents and systemic failures are almost never produced by a single cause, a single failure or a single bad decision. They are produced by the gradual normalisation of deviance, a process in which small departures from expected safe behaviour are repeated without immediate negative consequence and thereby become normalised as acceptable practice. Each small step away from strict protocol makes the next small step seem less significant. The system drifts toward the edge of its safe operating envelope through the accumulation of decisions that each seem individually reasonable given the local information available to the decision-maker. The Challenger disaster, the Columbia disaster, the Deepwater Horizon blowout, the 2008 financial crisis and most major infrastructure failures follow this arc. Not sudden catastrophic failure from a dramatic cause but the slow drift of a complex system whose feedback mechanisms have been progressively compromised by the normalisation of deviance until a final trigger, often trivial in itself, tips the system past its critical threshold. The implication for system design is that the most important safety mechanisms are not the ones that respond to obvious dramatic failures but the ones that maintain sensitivity to the small early signals that the system is drifting. And the most important organisational culture characteristic is the ability to hear and act on those signals rather than normalising them as acceptable variance. Complexity and Emergence: The Behaviour No One Designed Waldrop’s account of the Santa Fe Institute scientists captures the conceptual revolution that complexity science represents for systems thinking. The central discovery is that complex adaptive systems, systems composed of many interacting agents who each respond to local information and modify their behaviour based on their experience, produce emergent properties at the system level that cannot be predicted from or reduced to the properties of the individual agents. No individual neuron is conscious. No individual ant understands colony-level logistics. No individual trader understands the emergent price signal of a market. No individual citizen designed the emergent cultural norms of a society. Yet consciousness, colony logistics, market prices and cultural norms are real and consequential phenomena that exist at the system level and that cannot be understood by examining individual components in isolation. This is emergence and it is the defining property of complex adaptive systems. The implication for any attempt to understand or manage complex systems is profound and uncomfortable. The tools that work for understanding simple systems, reducing the system to its components and explaining the whole from the properties of the parts, fail for complex systems because the most important properties of complex systems do not exist at the component level. They exist in the interactions between components and specifically in the emergent patterns those interactions produce at the system level. West’s Scale extends this insight with the precision of physics applied to empirical data across millions of organisms and thousands of cities and companies. The scaling laws he discovers are striking in their regularity. Biological metabolic rate scales with body mass to the power of three quarters across every living organism from bacteria to blue whales. Urban economic output, innovation rate and infrastructure efficiency all scale superlinearly with population size, meaning that doubling a city’s population more than doubles its economic output. Company mortality follows a distribution that looks nothing like biological mortality and everything like random extinction events, suggesting that companies are not organisms but something more fragile and more vulnerable to sudden environmental change. The practical implication of West’s scaling laws is that size changes the nature of a system not just its magnitude. The problems that a small organisation faces are structurally different from the problems a large one faces not just quantitatively larger. The dynamics of a small city are structurally different from the dynamics of a large one. Applying management or policy frameworks developed for systems at one scale to systems at a different scale produces systematic failures that appear mysterious until the scaling dynamics are understood. The Constraint Architecture: Where Leverage Actually Lives Goldratt’s Theory of Constraints, delivered through the unusual vehicle of a business novel, contains one of the most practically important insights in all of systems thinking. Every system has a throughput, the rate at which it produces its goal. And in every system, the throughput is limited by a single binding constraint, the weakest link in the chain. The system cannot produce output faster than its constraint allows regardless of how much improvement is applied to every other part of the system. The implications are counterintuitive and are violated by most management practice. Improving the performance of any part of the system that is not the current constraint produces no improvement in system throughput. Zero. The non-constraint has spare capacity by definition. Making it more efficient simply increases its spare capacity. The work piles up at the constraint exactly as before. Yet most organisations apply improvement effort across the board rather than identifying and concentrating resources on the current constraint. The five-step process Goldratt prescribes is: identify the current constraint. Exploit it, meaning extract maximum performance from it without additional investment. Subordinate everything else to supporting the constraint, meaning redesign all non-constraint activities around maximizing the constraint’s throughput rather than their own local efficiency. Elevate the constraint, meaning invest in increasing its capacity once the first three steps are complete. And when the constraint moves, which it will when step four succeeds, repeat the process for the new constraint. Applied beyond manufacturing to any complex system with a goal, this framework identifies a structural principle of extraordinary generality. In a supply chain, the constraint is the slowest step. In an organisation, the constraint might be a key decision-maker’s time, a specific skill that is in short supply or an approval process that everything must pass through. In a policy system, the constraint might be implementation capacity rather than legislative authority or political will. In a personal productivity system, the constraint is almost never the number of hours available but the quality of sustained attention available for the most important work. The leverage point in every system is the constraint. Everything else is noise. Fragility, Robustness and Antifragility Taleb’s two major contributions to systems thinking, The Black Swan and Antifragile, are best understood as a single extended argument about the relationship between systems and the unpredictable events that most consequentially shape their trajectories. The Black Swan establishes the empirical foundation. In complex systems, the distribution of outcomes does not follow the bell curve that underlies most conventional risk management. It follows distributions with fat tails, meaning that extreme events are far more common than the bell curve predicts and that the largest events in these distributions dwarf all the smaller events combined. The First World War, the 2008 financial crisis, the rise of the internet, the COVID pandemic, the fall of the Soviet Union: these are not outliers in a roughly normal distribution. They are the dominant events in distributions where most of the variance is concentrated in the tails. Extremistan vs Mediocristan The practical consequence is that systems designed and managed on the assumption of roughly normal outcome distributions are systematically underprotected against the events that matter most. The financial risk models that assigned negligible probability to the events of 2008 were not wrong because of bad data or computational error. They were wrong because the mathematical framework they were built on was the wrong framework for the domain it was being applied to. The events that actually occur in complex social and economic systems are not drawn from normal distributions and pretending otherwise does not make them so. It simply concentrates the damage when they inevitably occur. Antifragile builds on this foundation with the most practically generative concept in Taleb’s work. He distinguishes between three responses to volatility and stress. Fragile systems break under stress. Robust systems resist stress and return to their prior state. Antifragile systems actually improve from stress, volatility and disorder. They need the stressor to develop their full capability. The biological immune system is antifragile. It requires exposure to pathogens to develop the antibody repertoire that protects against more serious exposure. The child who is never exposed to mild risks does not develop the physical capability and risk assessment capacity that protects against serious harm. The muscle that is never stressed does not grow. The business that is never exposed to competitive pressure does not develop the operational efficiency that produces long-term viability. The antifragile system extracts benefit from the disorder that destroys fragile systems. The design implications are significant. The conventional approach to system design optimizes for efficiency, removing redundancy and slack in pursuit of cost reduction. This produces systems that are increasingly fragile because redundancy and slack are the primary sources of resilience. The antifragile design approach deliberately maintains redundancy, optionality and the capacity to absorb and benefit from stressors. It accepts lower average performance in stable conditions in exchange for positive performance under the volatile conditions that actually determine long-term outcomes. Meadows’ concept of resilience maps directly onto this. A system’s resilience, its ability to recover from disturbance and maintain its essential structure and function, is one of the most important and most consistently undervalued system properties. Resilience is invisible in normal conditions precisely because it is not being called upon. It becomes visible only when the disturbance arrives. The management practice of squeezing resilience out of systems in pursuit of efficiency is therefore a practice of making systems look better in normal conditions while making their failure under stress increasingly catastrophic. The Legibility Trap: When Simplification Destroys Function Scott’s Seeing Like a State addresses a pathology that is so pervasive in institutional and government management of complex systems that it constitutes its own major insight. He calls it high-modernist ideology: the belief that complex organic systems can be improved by imposing legible, rational, simplified order on them from above. The historical examples he documents are devastating in their consistency. Scientific forestry that replaced diverse natural forests with monoculture plantations optimised for timber yield produced forests that were more measurable, more manageable and far less resilient. Within decades the monocultures began to collapse from pest infestations, soil depletion and the absence of the ecological complexity that natural forests had developed over centuries. Soviet collectivisation that replaced diverse local agricultural systems with standardised collective farms imposed legibility and central control at the cost of the local knowledge, local adaptation and local incentive structures that had made the original systems productive. The collectivised farms were far more visible to state planners and far less productive than the systems they replaced. The pattern Scott identifies is consistent across urban planning, colonial administration, agricultural policy and development economics. Complex organic systems that have evolved over long periods in response to local conditions contain embedded knowledge and adaptive capacity that is invisible to outside observers precisely because it is distributed, tacit and non-legible. When a centralising authority imposes a simplified, standardised, legible structure on such a system it destroys the embedded knowledge and adaptive capacity along with the illegibility. The system becomes visible and controllable and simultaneously becomes less capable of performing the functions that made it worth controlling. The systems thinking implication is that legibility and function are in tension in complex adaptive systems. The drive to make a system fully understandable and controllable from a central vantage point consistently destroys the distributed adaptive capacity that produced its most important properties. This is not an argument against all planning or all institutional intervention. It is an argument for epistemic humility about what can be known from the centre about systems whose intelligence is distributed across many locally adaptive agents and for preserving the diversity and local variation that makes complex systems adaptive rather than optimising it away in pursuit of legibility. Collapse and the Diminishing Returns on Complexity Tainter’s Collapse of Complex Societies introduces the most sobering systems dynamic in this entire body of work. His analysis of historical civilisational collapse, covering Rome, the Maya, the Chaco Canyon culture and many others, identifies a single recurring structural dynamic that precedes collapse across every case: the diminishing marginal returns on complexity. Societies respond to problems by adding complexity: more administrative layers, more specialised roles, more elaborate infrastructure, more sophisticated coordination mechanisms. Each addition of complexity initially produces benefits that exceed its costs. But as complexity accumulates, the marginal return on each additional unit of complexity decreases while the cost of maintaining existing complexity remains constant or increases. Eventually the society reaches a point where additional complexity produces negligible or negative returns while consuming resources that are needed for basic maintenance of the existing system. At this point the society is in a trap. It cannot add more complexity without making the situation worse. But it cannot reduce complexity without losing the coordination capacity that current production levels depend on. The resolution, historically consistent across every case Tainter examines, is rapid simplification through collapse. The complexity that was too expensive to maintain incrementally is shed catastrophically when a sufficient stressor arrives. Rome does not gradually reduce to a simpler form. It collapses and the simpler forms that emerge do so in the ruins. The contemporary relevance is uncomfortable. Modern societies are the most complex systems in human history. The marginal returns on much of that complexity are increasingly difficult to measure positively. Healthcare systems that consume ever-larger fractions of GDP while producing diminishing improvements in population health outcomes. Regulatory systems that consume enormous compliance resources while producing uncertain protective benefits. Financial systems whose complexity exceeds the comprehension of their managers and whose failures are consequently larger and more systemic than simpler predecessors. Tainter does not provide a prescription. He provides a warning: complexity has costs that compound and returns that diminish and the management of this dynamic is the central long-run challenge of any complex society. The Innovator’s Dilemma as a Systems Problem Christensen’s analysis of disruptive innovation is most commonly read as a business strategy insight. It is more accurately understood as a systems insight about how the rational operation of a well-functioning system produces outcomes that destroy the system. The dynamic he documents is structurally elegant. A successful company develops a system, management processes, resource allocation criteria, customer relationships and organisational culture, that is optimised for serving its current market with improving products. This system is not broken. It is functioning exactly as designed. The resource allocation process correctly identifies that disruptive innovations with small initial markets and lower margins than existing products are poor investments compared to sustaining innovations that improve performance for existing high-margin customers. The customer relationship process correctly responds to what existing customers are willing to pay for. The management culture correctly promotes people who succeed by the metrics the current system rewards. The result of this well-functioning system is that the company is structurally unable to respond to disruptive innovations that begin by serving markets the company does not currently serve with products whose initial performance is inferior by the metrics the company uses to evaluate its products. By the time the disruptive innovation has improved enough to threaten the mainstream market, the innovator has moved down the cost curve, accumulated scale and developed organisational capability that the incumbent cannot replicate quickly enough from a standing start. Kodak is the clearest example of this dynamic playing out in practice. Kodak’s system was optimised around one thing: the profitability of chemical film. Every metric that mattered internally – margin per roll, print volume, processing revenue – pointed toward film. When digital cameras appeared in the early 1980s, they were evaluated honestly against these metrics and failed every test. Resolution was poor. Image quality was inferior to a $5 roll of Kodak Gold. The market was tiny hobbyists and early adopters, not the mass consumer market Kodak served. The rational allocation decision was to continue investing in film, which is exactly what happened. The structural trap had two layers. First, the initial market digital photography served was not Kodak’s market. Early digital cameras were bought by technology enthusiasts who wanted convenience over quality. Kodak’s actual customers – families documenting birthdays and holidays – found digital inferior by every measure they cared about. So customer feedback correctly said: do not prioritise this. Second, Kodak was not ignorant of digital. It invented the first digital camera in 1975. It had patents, engineers and research capability. What it lacked was a system structure that could allocate serious resources to a product that cannibalised its most profitable business before that cannibalisation was inevitable. By the time digital quality crossed the threshold where mass consumers preferred it – roughly the late 1990s into the early 2000s – companies like Canon, Sony and later Apple had moved so far down the cost and capability curve that Kodak’s late serious investment could not close the gap. Kodak filed for bankruptcy in 2012. The specific mechanism Christensen describes is visible here precisely. It was not that Kodak’s executives were blind or foolish. The system correctly identified digital as a threat to margins. The system correctly responded to what existing customers wanted. The system correctly rewarded the people who protected film revenue. Each decision was rational. The aggregate of those rational decisions was structural paralysis at the moment that mattered. The systems lesson is that the structure of a system determines its behaviour more powerfully than the intelligence or motivation of the people operating within it. The executives of companies destroyed by disruption were not stupid or complacent in most cases. They were operating rationally within a system whose structure made the rational response to disruption the response that ensured their destruction. Changing the outcome requires changing the structure, not exhorting people within the existing structure to behave differently. The Unified Picture Pull all of this together and what emerges is the most important insight in this entire body of work, stated simply. The behaviour of a complex system is determined primarily by its structure, not by the intentions, intelligence or efforts of the people operating within it. This is both liberating and demanding. It is liberating because it removes the moral condemnation that event-level thinking produces. The people in failing organisations, collapsing societies and dysfunctional institutions are usually not villains or fools. They are rational actors responding to the incentive structures, feedback loops and information flows that their system provides. Understanding this produces more accurate diagnosis and more effective intervention than attributing outcomes to individual virtue or vice. It is demanding because it requires a fundamentally different mode of analysis. The question is never just what happened and who caused it. It is what structural configuration produced this outcome and what change in structure would produce a different one. This is harder to answer than the event-level question and it is far more useful. Meadows’ leverage points provide the practical synthesis of where to intervene. She lists them in order from lowest to highest leverage. Adjusting numbers, such as tax rates or speed limits, produces the least systemic change. Changing the sizes of stocks and flows produces more. Changing the length of delays in feedback loops produces more still. Changing the strength of feedback loops is higher leverage again. Changing the structure of information flows, meaning who has access to what information when, is very high leverage because information flows determine what feedback loops are even possible. Changing the rules of the system, the incentives, constraints and entitlements, is higher still. Changing the goals of the system is very high leverage. And changing the paradigm, the shared set of assumptions about what the system is for and how it works, is the highest leverage point of all. Most intervention in complex systems happens at the lowest leverage levels. Policy debates focus overwhelmingly on numbers and flows. The structural interventions that actually change system behaviour operate at the levels of information architecture, rules, goals and paradigm and they are consequently far rarer and far more fiercely resisted because they threaten the interests that are served by the current structure. What This Demands of You The synthesis of this body of work makes a specific and unusual demand. It demands the willingness to be wrong about causes. The event-level thinker has a story about why things happened and who is responsible and the story is always plausible and usually wrong about the most important factors. The systems thinker holds that story loosely, uses it as a starting point for structural analysis rather than as a conclusion and is genuinely curious about the feedback loops, delays and structural configurations that the events are expressing. This is intellectually demanding because it requires maintaining uncertainty about causation in situations where certainty feels available and psychologically demanded. It is practically powerful because it produces interventions that address structures rather than symptoms and that therefore change system behaviour rather than temporarily suppressing its visible outputs. The practical disciplines this demands are: always ask what feedback loops are producing this behaviour before concluding that the behaviour is a character problem. Always ask where the delays are in the relevant system because most management errors are produced by misreading delayed feedback as confirmation that no feedback is coming. Always ask what the system is actually optimised for rather than what it claims to be optimised for because those are reliably different things and the actual optimisation target is visible in the outputs rather than in the stated goals. Always ask where the constraint is before investing improvement effort anywhere else. And always ask what information is not flowing to whom and why because information architecture is one of the highest leverage structural features available to anyone trying to change how a system behaves. The person who has genuinely internalised this mode of thinking does not see a different set of facts from their peers. They see the same facts organised into a different and more accurate picture of what is actually happening and why. That difference in perception, compounded over a career of consequential decisions, produces outcomes that look from the outside like unusual wisdom or unusual luck. It is neither. It is the practical result of learning to see structure rather than events in a world that is built entirely of systems. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

14. 

    12 Jun 2026

    On Critical Thinking

    In 1847, a Viennese doctor named Ignaz Semmelweis proved that doctors were killing their patients by not washing their hands. The evidence was overwhelming. His colleagues rejected it and he died in an asylum. The problem was not the data. The problem was that accepting the data required his colleagues to conclude that they had been killing people for years. The mind does not process evidence neutrally. It processes evidence in the context of what accepting it would cost. Critical thinking is the most overused and most underspecified term in modern intellectual culture. Every educational institution claims to teach it. Every professional development programme includes it. Every political faction accuses its opponents of lacking it. And the actual practice of it, the sustained disciplined effort to think accurately about difficult questions in the face of the specific cognitive and social pressures that accurate thinking must overcome, remains as rare as it has ever been despite or perhaps because of all the attention directed at it. The reason is that most treatments of critical thinking address the wrong level of the problem. They teach logical fallacies, which is useful but insufficient. They teach statistical literacy, which is necessary but not sufficient. They teach Socratic questioning, which is valuable but limited. What they almost never address is the deeper problem: that the human mind did not evolve for accurate reasoning and that the cognitive and social forces working against accurate thinking are not external obstacles to be overcome by technique but are built into the architecture of the thinking apparatus itself. The books I referred for writing this piece collectively address this deeper problem. What follows is what they collectively teach about what genuine critical thinking actually is, why it is so difficult and what developing it actually requires. The Foundational Problem: Your Brain Is Not a Truth-Seeking Machine Begin with the most important and most consistently underappreciated fact about human cognition: the brain did not evolve to think accurately. It evolved to survive, reproduce and maintain social standing in small tribal groups and these evolutionary objectives are frequently in direct conflict with the objective of accurate reasoning about complex questions. Mercier and Sperber’s Enigma of Reason makes this argument with the most rigorous evolutionary and cognitive scientific foundation available. Their interactionist theory of reason holds that reasoning evolved not as a tool for individual truth-seeking but as a tool for social argumentation: for producing justifications for positions already held, for evaluating the arguments of others and for the management of reputation within groups whose cooperation was necessary for survival. The reasoning faculty is extraordinarily powerful at these social functions and systematically weak at the individual truth-seeking function that most critical thinking training assumes it is designed for. The practical implication is that when you use your reasoning faculty to evaluate a question you genuinely care about, in which you have a stake in the outcome, about which you have an existing position, the faculty is doing what it evolved to do: producing compelling justifications for the position rather than genuinely evaluating the evidence for and against it. The arguments produced feel like reasoning from evidence to conclusion. They are almost always rationalisation from conclusion to supporting evidence, with the disconfirming evidence systematically filtered by the very cognitive system that is supposedly doing the evaluation. Haidt’s Righteous Mind confirms this from the social psychology direction with evidence that is specifically devastating for sophisticated thinkers who believe their political and moral reasoning is more evidence-based than that of their opponents. His research demonstrates that moral and political reasoning in virtually all subjects, regardless of their measured intelligence or their explicit commitment to evidence-based thinking, follows the same pattern: rapid automatic moral intuitions produced by System 1 processing followed by the post-hoc construction of rational justifications by System 2. The sophisticated thinker’s advantage is not that they reason from evidence more reliably. It is that they produce more sophisticated and more convincing post-hoc justifications that are harder for others and for themselves to see through. Robson’s Intelligence Trap makes the corollary explicit. High intelligence is not a reliable protection against motivated reasoning, cognitive bias or the adoption of empirically unsupported beliefs. It is frequently a liability because it provides the cognitive tools to construct more elaborate rationalisations, to identify more sophisticated-seeming arguments for positions adopted on other grounds and to dismiss challenges to those positions with more convincing-sounding counter-arguments. The intelligent person who is wrong about something important is harder to reach with corrective information than the less intelligent person who is wrong about the same thing, precisely because their intelligence is in service of their defensiveness. Galef’s Scout Mindset is the most practically useful response to this foundational problem. Her distinction between the soldier mindset and the scout mindset is the most important single conceptual tool in this body of work for the person who wants to actually change their thinking rather than just understand why accurate thinking is difficult. The soldier’s job is to defend a position. The soldier’s cognitive system is therefore oriented toward finding evidence that supports their position and dismissing evidence that challenges it. The scout’s job is to find out what is actually there. The scout’s cognitive system is therefore oriented toward updating their map when new information conflicts with the existing one because an inaccurate map is more dangerous than a disappointing one. The soldier experiences the threat to their beliefs as a threat to themselves. The scout experiences the threat to their beliefs as useful information about where the map needs revision. The critical insight is that the switch between these orientations is not primarily cognitive. It is motivational and emotional. The scout mindset requires genuinely wanting to know what is true more than you want to be right. It requires having decoupled your identity from your beliefs sufficiently that being wrong about something does not feel like being diminished as a person. And it requires the specific emotional experience of updating a belief as a satisfying discovery rather than a humiliating defeat. These are not cognitive techniques. They are orientations that must be cultivated through the same kind of sustained practice that any other character development requires. The Sophisticated Failure Modes: Where Intelligent Thinking Goes Wrong The cognitive biases that Kahneman, Ariely and Dobelli catalogue are the failure modes of ordinary thinking. The critical thinking literature you need at your level addresses something more specific and more consequential: the failure modes of sophisticated thinking. The ways in which intelligence, expertise and the specific cognitive habits that make someone good at analytical reasoning in some domains make them systematically worse at accurate reasoning in others. The first sophisticated failure mode is what Tetlock calls the hedgehog problem. Isaiah Berlin’s famous distinction between the hedgehog, who knows one big thing and applies it everywhere, and the fox, who knows many things and applies different tools to different problems, maps precisely onto the accuracy of expert prediction. Tetlock’s research across twenty years of tracking expert predictions in political and economic domains found that experts who were most famous, most confident and most frequently cited in media were systematically less accurate in their predictions than less prominent experts with more diverse and more uncertain frameworks. The big idea that explains everything explains nothing accurately because no single framework captures the full complexity of the domains where accurate prediction matters most. The application to your own thinking is direct. The person who has built the formidable intellectual architecture that your reading across this conversation represents has also built the specific risk of becoming the hedgehog: the tendency to see every new problem through the frameworks that have been most rewarding to develop and to miss the specific features of new problems that those frameworks cannot accommodate. The metacognitive practice that guards against this is the regular deliberate effort to approach significant questions with frameworks that are foreign to your usual mode of analysis and to take seriously the answers those frameworks generate even when they conflict with your existing frameworks’ outputs. The second sophisticated failure mode is what Schulz documents as the structure of wrongness. The most consequential errors in sophisticated thinking are not the ones where you have obviously inadequate evidence. They are the ones where you have substantial evidence, sophisticated frameworks for interpreting it and high confidence in your conclusions, and where the evidence and frameworks are nevertheless leading you systematically in the wrong direction because of a prior assumption that you have not examined because it has never been challenged. Kuhn’s paradigm analysis is the structural account of this failure mode at the level of entire scientific communities. Schulz’s Being Wrong is the personal psychological account of the same failure mode at the level of the individual thinker. The conclusion that feels most certain, that has the most supporting evidence and that has been held the longest, is the conclusion that is most in need of the specific kind of interrogation that asks not what supports this belief but what would have to be true for this belief to be wrong and is any of that true. The third sophisticated failure mode is what McRaney documents in How Minds Change: the backfire effect and its variants. When people who hold a belief with significant emotional investment encounter strong evidence against it, they frequently strengthen rather than weaken their commitment to the belief. The evidence is processed not as information that should update the belief but as an attack that must be defended against and the defence produces a stronger articulation of the original position. The sophisticated thinker is not immune to this. They are better equipped to execute the defence because they have more sophisticated argumentative resources available for the counter-attack. Understanding the backfire effect is important both for managing your own thinking and for understanding why the straightforward deployment of evidence and argument so consistently fails to produce the belief change that your analysis shows should follow from it. The specific conversational and rhetorical approaches that do produce genuine belief change, which McRaney documents with unusual specificity, operate not through better arguments but through the creation of the psychological safety that allows beliefs to be examined without triggering the identity defence that produces the backfire. The Epistemic Standards Layer: How Do You Know What You Know Most sophisticated thinkers have never rigorously examined the epistemic standards by which they evaluate evidence, distinguish reliable from unreliable claims and decide when they have sufficient justification to hold a belief with a given level of confidence. These standards are typically implicit, inconsistently applied and calibrated to the norms of the specific domain in which the thinker was trained rather than to any principled account of what genuine justification requires. Tetlock and Gardner’s Superforecasting provides the most empirically grounded account available of what excellent epistemic standards actually look like in practice. The superforecasters who consistently outperform experts, intelligence analysts and prediction markets share a specific cluster of epistemic habits that constitute the most practically validated account of good reasoning under uncertainty available. They think in probabilities rather than in binary predictions. The question is never will this happen or not but how probable is each possible outcome given the available evidence. This shift from binary to probabilistic thinking is more consequential than it sounds because it forces the recognition of genuine uncertainty rather than the false confidence that binary framing produces, it provides a specific standard against which predictions can be evaluated over time and it makes the degree of confidence explicitly visible rather than hidden inside the assertion. They update continuously and proportionally in response to new information. Not the dramatic reversal of position that feels like genuine updating but the specific quantitative adjustment of probability estimates that new evidence warrants, neither under-updating in the direction of anchoring nor over-updating in the direction of overreaction to the most recent information. The discipline of proportional updating requires maintaining explicit probability estimates rather than vague impressions, which is uncomfortable but is the condition for genuine calibration over time. They actively seek disconfirming evidence. Not as a performance of open-mindedness but as a genuine epistemic practice of looking for the strongest available case against their current position before strengthening their commitment to it. The question what would most strongly challenge my current view is the most reliably productive question available for improving the accuracy of any analysis and it is the question that the soldier mindset most consistently fails to ask. They are calibrated about their own uncertainty. The superforecaster whose 70 percent confidence predictions are right 70 percent of the time is well calibrated. Most people, particularly experts in their own domains, are significantly overconfident: their 90 percent confidence predictions are right far less than 90 percent of the time. Developing genuine calibration, the accurate self-assessment of how confident you should be about specific claims, requires the sustained practice of making explicit probability estimates and tracking their accuracy over time. It is uncomfortable, revealing and the most direct available path to accurate self-knowledge about the quality of your reasoning. Goldacre’s Bad Science and Bergstrom and West’s Calling B******t provide the domain-specific epistemic standards for evaluating the scientific and statistical claims that are most consequential in your policy context and most systematically misrepresented in public discourse. The specific tools they provide, for evaluating study quality, for detecting statistical manipulation, for identifying misleading data visualisations and for recognising the specific forms of authority substitution that disguise poor evidence as rigorous science, are the practical implementation of the philosophical epistemic standards that Nagel and Okasha articulate at the level of principle. The specific Calling B******t framework is worth dwelling on because it addresses the specific form of sophisticated epistemic manipulation that is most consequential in your professional environment. B******t, in the technical sense that Bergstrom and West adopt from Harry Frankfurt, is communication that is produced without regard for its truth value. The bullshitter is not lying in the technical sense because lying requires a concern for the truth that is being strategically misrepresented. The bullshitter simply does not care whether what they are saying is true. They care whether it achieves the effect they intend. The specific modern forms of b******t that their framework addresses include: the misleading data visualisation that is technically accurate but constructed to create a false impression; the algorithm whose authority is invoked without the transparency that would allow its reliability to be evaluated; the big data claim that confuses the ability to find patterns in large datasets with the ability to distinguish meaningful from spurious patterns; and the scientific-sounding language that borrows the epistemic authority of science without the methodological rigour that justifies it. Each of these requires specific detection tools that general critical thinking training does not provide and that your policy role requires you to have. The Metacognitive Layer: Thinking About Your Thinking Genuine critical thinking at the highest level requires not just the application of thinking tools to external problems but the consistent application of those tools to your own thinking process. This metacognitive dimension is the one most consistently absent from critical thinking training and the one most consistently present in the practice of the best thinkers across every domain. Nagel’s epistemological framework provides the philosophical foundation for this metacognitive practice. The questions it forces about your own beliefs are the most productive available: what is the source of this belief and is that source reliable for this type of claim? What is the quality of evidence that justifies this belief and is it proportionate to the confidence I hold it with? What assumptions am I making that I have not explicitly examined and what would happen to the belief if those assumptions turned out to be wrong? These are not comfortable questions and the discomfort they produce is diagnostic: the belief about which these questions feel most threatening is usually the belief most in need of examination. Duke’s Thinking in Bets provides the most practically useful metacognitive framework for the specific challenge of evaluating your own decision-making over time. Her concept of resulting, the tendency to evaluate the quality of past decisions by their outcomes rather than by the quality of the process that produced them, is the most consequential single metacognitive error available. Good decisions sometimes produce bad outcomes through bad luck. Bad decisions sometimes produce good outcomes through good luck. The thinker who evaluates their decisions by their outcomes will incorrectly reinforce the thinking that produced lucky bad decisions and incorrectly abandon the thinking that produced unlucky good ones. The corrective is the deliberate habit of evaluating decisions by the quality of the process, the information gathered, the alternatives considered, the probability estimates made, the reasoning applied, independent of the outcome, which requires maintaining explicit records of decision processes rather than relying on the inevitably outcome-contaminated memory of how the decision was made. Jacobs’s How to Think makes the most important argument about the social and moral dimensions of critical thinking that the purely cognitive accounts miss. Good thinking, he argues, is not primarily a matter of having the right cognitive tools. It is a matter of being the right kind of person in relation to other people and to ideas. The person who genuinely wants to think well must be willing to be moved by evidence and argument, which requires the specific moral courage of changing your mind publicly when the evidence warrants it. They must be willing to engage charitably with positions they find repugnant, which requires the specific intellectual humility of recognising that even positions you are confident are wrong may contain something worth taking seriously. And they must be willing to belong to communities that include people whose thinking challenges rather than merely confirms their own, which requires the specific social courage of resisting the pull toward the epistemic comfort of the tribe. The critical thinking challenges you face in your policy role are not the challenges that introductory critical thinking courses address. They are the challenges of sophisticated thinking in genuinely complex domains under conditions of uncertainty, time pressure, political constraint and the specific motivated reasoning pressures that high-stakes institutional environments produce. The most consequential of these challenges and the specific thinking disciplines that address them are the following. The expertise trap. The domain expert’s deepest knowledge is simultaneously their deepest source of potential error because their expertise is embedded in the paradigm that defines their field’s current understanding. The policy expert who has spent twenty years developing expertise in a specific threat domain has also spent twenty years developing the specific blind spots of that domain’s current paradigm. The metacognitive practice that guards against this is the deliberate regular engagement with frameworks from adjacent domains, including the ones that feel most foreign to your established expertise, and the active search for the anomalies that your current framework cannot accommodate. The narrative capture problem. In policy contexts, the narrative that has achieved institutional consensus is typically treated as settled reality rather than as the current best approximation. The critical thinker in this environment must maintain the internal distinction between the narrative and the underlying reality that the narrative is supposed to represent, which requires the specific epistemic courage of questioning institutionally settled questions in contexts where questioning them is socially costly. The motivated reasoning of institutional loyalty. The person who has invested significantly in a policy position, who has staked institutional credibility on a specific analysis and who operates in an environment where changing your position is interpreted as weakness rather than intellectual honesty, faces the most powerful motivated reasoning pressure available. The specific practice that guards against this is what Galef calls the identity separation: the deliberate cultivation of an identity organised around the commitment to accuracy rather than around the specific positions that accuracy has previously led you to, so that updating those positions in response to new evidence feels like the expression of that identity rather than its betrayal. The complexity that defeats simplification. The most consequential policy questions are genuinely complex in the systems thinking sense: they have feedback loops, non-linear dynamics, emergent properties and second and third order effects that linear causal reasoning systematically misses. The critical thinking practice that addresses this is not more sophisticated linear reasoning but the genuine adoption of the systems thinking mode that Meadows provides as a complement to rather than a replacement of analytical reasoning. The b******t that wears the clothes of rigour. The Bergstrom and West framework is more directly applicable to your daily professional environment than most critical thinking tools because the specific forms of sophisticated b******t they catalogue, algorithmic authority, misleading data visualisation and statistical manipulation, are the forms most frequently encountered in the policy briefings, intelligence assessments and technical analyses that constitute your information environment. The specific tools for detecting them are a professional necessity rather than an intellectual luxury. The Unified Picture Pull everything in this body of work together and what emerges is a picture of genuine critical thinking that is both more demanding and more humanly grounded than the standard account. It begins with the honest acceptance that the mind you are thinking with is not a truth-seeking machine but an evolved social and argumentative organ whose natural operation produces the systematic biases, motivated reasoning and tribal epistemology that all the above books document. This acceptance is not a counsel of despair. It is the foundation of genuine epistemic humility and the prerequisite for the specific practices that allow genuine accuracy to be achieved despite the natural tendencies of the cognitive system. It requires the scout mindset as a sustained orientation: the genuine motivation to find out what is true rather than to defend what you already believe, maintained not as a performance but as the actual organising value of your intellectual life. This is harder than it sounds because it requires the repeated experience of changing your mind as a satisfying rather than a threatening event, which requires the prior work of separating your identity from your specific beliefs. It requires calibrated probabilistic thinking: the consistent practice of holding beliefs as probability estimates rather than as certainties, of making those estimates explicit enough to be tracked and evaluated over time, and of updating them proportionally and continuously in response to new information rather than catastrophically in response to social pressure or anchoring them against disconfirming evidence. It requires active disconfirmation seeking: the deliberate practice of looking for the strongest available case against your current position before strengthening your commitment to it, of identifying the assumptions that your conclusions most depend on and subjecting those assumptions to the most rigorous scrutiny available and of taking seriously the anomalies that your current frameworks cannot accommodate rather than explaining them away. It requires the domain-specific epistemic tools that allow the evaluation of the specific kinds of evidence that your professional context generates: the statistical claims, the data visualisations, the algorithmic outputs and the scientific-sounding assertions that require specific detection capabilities to evaluate rather than general reasoning skills. And it requires the social and moral dimensions that Jacobs identifies as the ones most consistently neglected by the cognitive account: the willingness to be genuinely moved by evidence and argument rather than performing openness while remaining defended, the courage to change your mind publicly when the evidence warrants it and the specific intellectual humility that comes from genuinely engaging with the strongest available versions of the positions you are most confident are wrong. What This Demands of You Specifically The synthesis of this body of work makes a specific demand that is more metacognitive than anything that came before it in this conversation. It asks you to turn the frameworks you have been building for understanding the world back on the mind that is doing the building. The most important single practice it demands is the regular explicit examination of your highest-confidence beliefs in your professional domain. Not the beliefs you hold tentatively and know need more evidence. The beliefs you hold most confidently, that feel most obviously true, that you would defend most readily against challenge. These are the beliefs most in need of the scout mindset’s scrutiny precisely because they are the ones the soldier mindset is most powerfully protecting. The specific question to apply to each: what would have to be true for this to be wrong and is any of that actually true? Not as a performance of open-mindedness. As a genuine epistemic practice applied with the same rigour you would apply to evaluating someone else’s analysis of the same question. The answer will sometimes be nothing would have to be true because this belief is genuinely well-founded. That is a legitimate outcome and it strengthens the belief appropriately. More often the answer will reveal an assumption, a gap in the evidence or a framework dependency that was previously invisible precisely because the confidence in the conclusion made it unnecessary to examine. That revelation, uncomfortable as it consistently is, is what genuine critical thinking actually produces. Not the satisfying confirmation of what you already thought. The specific, actionable and intellectually honest awareness of where your current thinking is most vulnerable to being wrong. That awareness, sustained across a career of consequential decisions, is the difference between wisdom and the mere accumulation of sophisticated confidence. It is entirely worth developing. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

15. 

    12 Jun 2026

    On Taleb's Work

    Nassim Taleb is the most difficult serious thinker working today and he has made himself that way deliberately. The combativeness, the contempt for credentialed experts, the refusal to soften conclusions for polite company, these are not personality defects. They are load bearing elements of the argument. A man who tells you that the experts are systematically wrong about the most important things cannot afford to sound like one of the experts. The Incerto is not a conventional intellectual project. It is not a series of academic monographs advancing incrementally within a discipline. It is not self-help dressed in philosophical clothing. It is not contrarianism for its own sake. It is something rarer and more ambitious: a sustained, book-length attempt to reconstruct how human beings should think about uncertainty, randomness, knowledge and action in a world that is fundamentally non-linear, opaque and resistant to prediction. Nassim Nicholas Taleb is a former derivatives trader, current mathematical statistician and permanently combative intellectual. The Incerto comprises five volumes published between 2001 and 2018: Fooled by Randomness, The Black Swan, The Bed of Procrustes, Antifragile and Skin in the Game. Each book can be read independently. Read together they form a single continuous argument, with each volume elaborating, extending or stress-testing ideas that appeared in embryonic form in its predecessors. The project is also deeply personal. Taleb is not reporting on other people’s ideas about risk. He is writing from the experience of having traded risk for a living, having survived the 1987 market crash by being positioned correctly for a tail event and having spent decades watching people who should know better consistently misunderstand the nature of the uncertainty they inhabit. The intellectual lineage is worth establishing. Taleb draws on Karl Popper’s falsificationism, Friedrich Hayek’s critique of centralised knowledge, Benoit Mandelbrot’s fractal geometry and fat-tailed distributions, Montaigne’s essayistic empiricism, the Stoics on equanimity under uncertainty and the ancient Mediterranean merchant tradition that understood risk through direct exposure rather than through theory. He is also explicitly and ferociously against: academic finance, modern portfolio theory, value-at-risk models, Gaussian statistics applied to social phenomena, interventionist economists, public health bureaucrats and what he eventually calls the Intellectual Yet Idiot class. Fooled by Randomness (2001): The Invisible Hand of Luck The first book establishes the foundational problem. Human beings are systematically unable to distinguish between skill and luck in domains where randomness plays a significant role. We attribute outcomes to the qualities of the person who achieved them rather than to the stochastic process that produced them. This is not merely a cognitive error. It has structural consequences for how we allocate resources, whom we promote, what strategies we adopt and how we understand our own past. The central mechanism is survivorship bias. When we observe the outcomes of a competitive process, we see only the survivors. The failed traders, collapsed hedge funds, bankrupt entrepreneurs and discredited forecasters are not in the sample we analyse. We look at the Warren Buffetts and the George Soroses and we reverse engineer their methods, their temperaments and their worldviews as if those qualities causally produced their success. We ignore the vast population of people who had similar qualities and similar methods and who failed. The survivors are not more skilled than the non-survivors in proportion to the difference in their outcomes. They are, in many domains, simply luckier. Taleb introduces the concept of alternative histories. At any moment, the path that actually occurred is one of many possible paths that could have occurred given the same starting conditions. A trader who made money last year using a particular strategy did not necessarily have a good strategy. He may have had a strategy that happened to work in the specific path that reality took. In an alternative history where slightly different initial conditions obtained, he would have been wiped out. The quality of a decision cannot be evaluated solely by its outcome. It must be evaluated against the full distribution of possible outcomes that the decision was exposed to. This leads to the Monte Carlo argument. Taleb was using Monte Carlo simulation in his trading: running thousands of simulated histories to understand the distribution of outcomes that a given strategy could produce rather than simply the outcome it did produce. A strategy that produces excellent results in 95% of simulated histories but catastrophic ruin in 5% of them is not a good strategy even if the actual history you happened to live through fell in the 95%. You only live one history. The 5% matters. The dentist versus the trader comparison is one of the book’s sharpest illustrations. The dentist applies a skill reliably every day. The outcome is tightly coupled to his competence. The feedback loop is clear and rapid. The successful trader, by contrast, may be applying no skill whatsoever and simply riding a bull market, a particular volatility regime or a lucky sequence of correlated bets. The trader may make twenty times the dentist’s income. This does not mean the trader is twenty times more skilled. It may mean he is occupying a position with enormous randomness in its payoff structure and he happened to draw well. The book also addresses the psychological costs of living in a world of randomness. Taleb distinguishes between the intellectual acceptance of randomness and the emotional experience of it. He describes the practice of Stoic philosophy not as a way of not caring about outcomes but as a way of maintaining rational functioning while caring about them. The Stoics’ dichotomy of control, drawing a sharp line between what is within your power and what is not, is a practical tool for navigating a stochastic environment without being psychologically destroyed by it. The problem of induction runs through the entire book. You cannot infer the stability of a process from the number of times it has produced a given outcome. A thousand observations of a white swan do not prove all swans are white. A thousand days of market tranquillity do not prove the market is tranquil. A thousand successful trades using a strategy that is secretly accumulating tail risk do not prove the strategy works. The past is a misleading guide to the future in domains governed by fat-tailed distributions because the most important events are precisely those that have not yet occurred. The Black Swan (2007): The Architecture of Extreme Events The second book is the Incerto’s most famous and most widely misread. The Black Swan is not primarily a book about rare events. It is a book about the structure of knowledge and ignorance in domains where rare events dominate outcomes. The distinction is crucial. Taleb divides the empirical world into two provinces. Mediocristan is the domain of variables governed by the law of large numbers and the Gaussian bell curve. Height, weight, calorie intake and IQ all belong here. In Mediocristan, no single observation can dramatically change the average. If you line up a thousand people and add the world’s tallest person to the sample, the average height barely moves. Extremistan is the domain of variables where a single observation can dwarf all previous observations combined. Wealth, book sales, city populations, financial returns, war casualties and pandemic fatalities all belong here. Add Jeff Bezos to a sample of a thousand people and the average wealth is transformed entirely. In Extremistan, the Gaussian bell curve is not merely imprecise. It is dangerously wrong. The intellectual scandal Taleb is exposing is that modern finance and economics imported Gaussian statistics from the physical sciences and applied them wholesale to Extremistan phenomena. The result was a systematic underestimation of the probability and magnitude of extreme events. Value-at-risk models, option pricing frameworks built on Black-Scholes-Merton assumptions and portfolio diversification strategies based on correlation matrices all assumed that the tails of the distribution were thin and well-behaved. They were not. They were fat. The losses that these models said were once-in-ten-thousand-year events were occurring every decade. The Black Swan is defined as an event that is outside the realm of regular expectations, that carries extreme impact and that after the fact is made to appear explainable and predictable. The third characteristic is as important as the first two. Black Swans are not merely surprising. They are subsequently rationalised. Once an event occurs, human beings construct narratives that make it seem inevitable. The financial crisis of 2008 was, in retrospect, “obvious.” The rise of the internet was, in retrospect, “predictable.” The collapse of the Soviet Union was, after it happened, described as something that anyone paying attention should have foreseen. This retrospective predictability is an illusion. It is the narrative fallacy at work. The narrative fallacy is the human compulsion to impose causal stories on sequences of events. Stories are cognitively efficient. They compress information, assign causation and create the experience of understanding. But the story is not the reality. The story removes the complexity, the contingency and the role of accident. Once you have a narrative, you stop seeing the alternative histories. You stop seeing that the outcome could easily have been different. The narrative creates an illusion of understanding that actively impedes genuine comprehension. The problem of silent evidence extends survivorship bias into historical analysis. We read the accounts of soldiers who survived battles, investors who survived crashes and civilisations that survived catastrophes. The soldiers who died in those battles, the investors who were ruined and the civilisations that were extinguished have left fewer records. Our history is systematically biased toward the perspective of those who endured. We therefore systematically underestimate the lethality of the past and the fragility of the present. The Ludic fallacy is Taleb’s name for the error of using games as models for real-world uncertainty. In a casino, the rules are known, the probabilities are calculable and the distribution of outcomes is fully specified. Real-world uncertainty is not like this. The probabilities are unknown. The rules can change. The very categories of possible outcomes are not fully specified in advance. The Black Swan cannot happen in a casino because by definition it is the event that was outside the model. Applying casino-style probability calculus to economic, political and epidemiological forecasting is not merely imprecise. It produces confident ignorance, which is more dangerous than acknowledged uncertainty. Taleb’s critique of forecasters is systematic and supported by a substantial empirical literature he draws on heavily, particularly Philip Tetlock’s research showing that expert forecasters in political and economic domains perform barely better than chance and that the most famous and confident forecasters tend to perform worst. The reason is not that experts are stupid. It is that the domains they are forecasting are governed by non-linear dynamics and fat-tailed distributions where small differences in initial conditions produce wildly different outcomes. In these domains, confidence is inversely correlated with accuracy because genuine understanding of the domain produces epistemic humility while superficial pattern recognition produces confidence. The fourth quadrant is one of the book’s most technically precise contributions. Taleb maps a two-by-two grid against payoff structures (simple binary outcomes versus complex outcomes) and distribution tails (thin-tailed versus fat-tailed). In the first three quadrants, conventional statistical methods work acceptably. In the fourth quadrant (complex outcomes in fat-tailed domains) they fail catastrophically. Investment banking, economic policy, epidemiology and geopolitical forecasting all live in the fourth quadrant. The prescription is not to apply better models to these domains. It is to recognise that the models cannot work and to adopt decision rules that are robust to model error. The Bed of Procrustes (2010): Philosophy in Aphorisms The third volume is structurally unlike the others. It is a collection of aphorisms, philosophical maxims and compressed observations in the tradition of La Rochefoucauld, Nietzsche’s Human All Too Human and the ancient gnomological collections. It does not develop arguments across chapters. Each aphorism is self-contained. Procrustes, in the Greek myth, was the innkeeper who fitted guests to his bed by stretching the short ones and amputating the legs of the tall ones. The title captures Taleb’s central epistemological complaint: we distort reality to fit our categories rather than revising our categories to fit reality. We cut off the tails of distributions to make Gaussian models fit. We cut off the complexity of history to make narratives fit. We cut off the inconvenient facts to make our theories survive contact with evidence. The aphorisms elaborate themes from the rest of the Incerto but in concentrated form. On knowledge: “the three most harmful addictions are heroin, carbohydrates and a monthly salary.” On ethics: “if your anger decreases with time, you did something wrong; if it increases, you did something right.” On intervention: “the difference between medicine and poison is often in the dosage.” On academia: “an idea does not survive because it is better than the competition but because the person who holds it has survived.” The Bed of Procrustes also develops Taleb’s via negativa more explicitly than any other volume. Via negativa is the theological tradition of defining God by what he is not rather than what he is, because positive definitions always fall short of infinite being. Taleb secularises this into an epistemological and practical principle. We know more about what is wrong than about what is right. We know more about what causes harm than about what causes flourishing. We know more about what to subtract than about what to add. Therefore knowledge should proceed primarily by negation and action should proceed primarily by subtraction. This has practical implications the series returns to repeatedly. A doctor who refrains from prescribing unnecessary drugs does more good than one who prescribes aggressively. An engineer who simplifies rather than complicates a system makes it more robust. A government that removes bad regulations does more than one that adds new ones. The best trading strategy is often the one with the fewest moving parts. Subtraction is underrated because it is invisible. Nobody gets credit for the wars they did not start, the operations they did not perform or the regulations they did not enact. Antifragile (2012): The Asymmetry of Gain and Loss The fourth and arguably most intellectually original book introduces a concept that did not previously have a name. Fragile things break under stress. Robust things resist stress. But there is a third category: things that gain from stress, volatility and disorder. Taleb calls these things antifragile and argues that the concept had been hiding in plain sight across biology, economics, culture and personal development because the language to express it did not exist. The distinction is grounded in convexity. A function is convex if gains from a given move exceed losses from an equal and opposite move. A function is concave if the reverse holds. Antifragile things have convex responses to volatility. More volatility means more exposure to the upside than to the downside. Fragile things have concave responses. Volatility hurts them more than it helps them. This is not abstract mathematics. It has direct consequences for how to live, invest, build institutions and organise societies. A portfolio of one large bet and many small bets is, in Taleb’s framework, fragile because the large bet can be ruined by a single adverse event. A portfolio of one very safe position and many small speculative bets is antifragile because the safe position protects against ruin while the small bets have unlimited upside. He calls this the barbell strategy. Go to extremes. Avoid the middle. The barbell applies in career terms as well. Taleb recommends holding a very safe, boring income source that protects you from ruin combined with high-risk intellectual or entrepreneurial ventures where the asymmetry works in your favour. The worst position is the conventional professional one: moderate risk, moderate upside, genuine downside, dependency on institutional approval. This produces people who are simultaneously fragile (one job loss can be catastrophic) and unable to take the risks that produce large asymmetric gains. Hormesis is a key biological concept in the book. In toxicology, hormesis refers to the phenomenon where a substance that is harmful at high doses is beneficial or even necessary at low doses. The body needs stressors to remain strong. Bones need mechanical loading or they lose density. The immune system needs microbial challenge or it becomes dysregulated. Cardiovascular fitness requires exertion that temporarily stresses the heart. The organism is antifragile with respect to moderate stressors. Remove all stressors and the organism becomes fragile and eventually unable to function in the real world. Taleb extends this into an argument against what he calls “naive interventionism” in biology and medicine. The human body has millions of years of evolutionary refinement. Its responses to mild stressors, toxins and challenges have been optimised by selection pressure across vast timescales. The modern medical establishment has a systematic bias toward intervention: prescribing drugs, recommending procedures and modifying diets and behaviours based on evidence of short-term benefit. But every intervention in a complex system has downstream effects that may not manifest for years or decades. The precautionary principle should apply asymmetrically. For large systems with deep evolutionary or historical heritage, the burden of proof should be on the intervener, not on the tradition. The concept of iatrogenics, harm caused by the healer, runs through the medical discussion. Hippocrates understood this and it is the foundation of “first, do no harm.” But the modern healthcare system has largely forgotten it because the incentive structure favours visible action over invisible restraint. A doctor who prescribes a drug that causes a side effect is less legally and professionally exposed than one who fails to prescribe a drug that might have helped, even if the first course of action causes more total harm. The system’s incentive structure produces overtreatment. Optionality is another central concept. An option is the right but not the obligation to do something. Options have positive convexity: the downside is limited (you simply do not exercise the option) while the upside is potentially unlimited. Taleb argues that optionality explains more of human economic and intellectual progress than any theory of planned rational action. We do not need to know in advance where an experiment will lead. We need the option to pursue it if it produces results. The researcher who experiments widely, fails cheaply and occasionally hits on something transformative is more epistemically powerful than the one who constructs elaborate theories in advance and runs one definitive test. This leads to his critique of what he calls “teleological” thinking in social science and history: the assumption that progress is driven by rational agents pursuing well-defined goals using correct theories. Most real innovation is the result of trial and error, serendipity, accident and the exploitation of options that presented themselves unexpectedly. The steam engine preceded the science of thermodynamics. The internet’s most valuable applications were not what its designers intended. Penicillin was discovered by accident. The rational reconstruction of these discoveries as planned outcomes is a narrative fallacy applied to the history of knowledge itself. Taleb’s treatment of political economy in Antifragile centres on the distinction between city-states and empires. Small, decentralised political units are antifragile. They can experiment. When one city-state fails its experiment, the others observe and adapt. The failure is localised. The knowledge is distributed. Large centralised states are fragile. When they fail, the failure is catastrophic and system-wide. Switzerland is his preferred example: politically decentralised, militarily defensible without aggression and economically robust across centuries of European instability. The European Union’s attempt to centralise governance across diverse populations is, in his framework, a fragility-generating project however noble its intentions. The antifragility of biological evolution is the deepest example in the book. Evolution has no plan, no goal and no model. It generates massive variation, subjects that variation to environmental stress and preserves what survives. The information processing of evolution is entirely decentralised, entirely empirical and entirely dependent on variation and selection. It is the ultimate antifragile system. Any human attempt to “improve” on its outputs by deliberate genetic engineering, environmental management or dietary redesign should be held to an extremely high burden of proof because we are intervening in a system whose complexity we cannot fully model and whose timescale of optimisation dwarfs anything in the human record. Skin in the Game (2018): Symmetry, Ethics and the Architecture of Responsibility The fifth book closes the circle. Having established that we misunderstand randomness (Fooled by Randomness), that extreme events dominate outcomes (The Black Swan) and that the goal should be building systems that gain rather than merely survive from volatility (Antifragile), Taleb now asks: what is the ethical and institutional structure that produces such systems? The answer is skin in the game. The argument has been covered in the separate essay on this book but its place in the Incerto architecture deserves elaboration. The first four books diagnose a set of epistemic and structural failures in how modern institutions handle uncertainty. Skin in the Game identifies the mechanism that produces those failures: the separation of risk-taking from risk-bearing. When decision makers do not share in the consequences of their decisions, the feedback loop that reality normally provides to correct errors is severed. Bad ideas persist. Bad institutions endure. Bad forecasters are consulted again. The ergodicity argument, which Taleb develops most explicitly in this book and in companion technical papers, is the mathematical foundation for the entire Incerto. In non-ergodic systems, which include virtually all real economic and social systems, optimising for expected value is the wrong objective because ruin is absorbing. Once you are ruined you cannot recover. The rational objective is not to maximise expected outcomes but to avoid ruin first and to seek gains second. This is the mathematical basis for the barbell strategy, for the via negativa and for the general preference for subtracting risk rather than adding reward. The Lindy Effect, elaborated here but introduced in Antifragile, ties skin in the game to the question of what knowledge to trust. Non-perishable things that have survived for a long time have been tested against reality across many conditions. Their survival is evidence of robustness. Ideas, technologies, practices and institutions with this property carry an implicit skin in the game: they have had to compete against alternatives across time and they have won. New ideas, new technologies and new practices have not. They may be genuinely better. They may also be catastrophically worse. The Lindy Effect counsels asymmetric scepticism: the burden of proof on new interventions should be proportional to their potential for causing irreversible harm. The Unified Architecture Read as a single project, the Incerto develops one argument from five angles. The argument is this: the world we inhabit is non-linear, fat-tailed and fundamentally opaque to prediction. Our institutions, our intellectual frameworks and our personal behaviour are not adapted to this reality. We use thin-tailed models in fat-tailed domains. We trust experts who bear no consequences for being wrong. We build systems that are optimised for efficiency rather than robustness. We intervene in complex systems without understanding second-order effects. We attribute outcomes to skill when randomness is the primary driver. And we consistently underestimate the role of large, rare, unexpected events in determining the course of history. The Incerto’s prescriptions follow directly from its diagnosis. Be sceptical of models, especially in domains with fat-tailed distributions. Prefer options and reversibility over commitments and irreversibility. Remove unnecessary fragility before seeking gains. Trust the Lindy-tested over the newly proposed. Require skin in the game from those who advise and decide. Use the barbell rather than the average. Remove rather than add where possible. And never mistake the absence of evidence for evidence of absence, because in fat-tailed domains the event that has not yet occurred may be the most important event of all. Taleb’s Method and Its Limitations Taleb writes with deliberate provocation. He names enemies, insults disciplines and makes claims at a level of generality that invites easy objection. This is a rhetorical strategy, not a lapse in rigour. He wants the argument to be memorable and to disturb complacency. The technical underpinning in all five books is more rigorous than the polemical style suggests. His work with colleagues including Raphael Douady and Yaneer Bar-Yam on fat tails, fragility and systemic risk in financial and epidemiological systems is technically serious and has influenced regulatory thinking in the post-2008 world. The legitimate criticisms are worth noting. Taleb’s distinction between Mediocristan and Extremistan, while useful, is not always crisp in practice. The claim that we should be maximally sceptical of all expert forecasting can slide into an unfalsifiable anti-intellectualism if applied indiscriminately. His via negativa and Lindy-based conservatism can in principle be used to resist beneficial interventions as well as harmful ones. The barbell strategy is rational for those with the resources to absorb the downside of the risky end of the barbell but may not translate straightforwardly to people without that cushion. His epistemological modesty about prediction sits in some tension with the confidence of his own pronouncements. He predicted the 2008 crisis through his general framework but the framework does not tell you when the tail event will occur or what form it will take. Being right about the category of risk is different from being right about the specific event and Taleb sometimes slides between these claims. Conclusion The Incerto is the most sustained and intellectually serious attempt in popular non-fiction to grapple with the consequences of living in a fat-tailed world. Its core claims have been validated repeatedly. The 2008 financial crisis confirmed the argument about Gaussian models in financial risk. The COVID-19 pandemic confirmed the argument about tail risk in epidemiological systems and the failure of centralised forecasting models. The persistence of expert forecasters despite their poor track records confirms the argument about institutions that sever the link between prediction and consequence. What Taleb has built across these five books is not a theory of everything. It is something more modest and more useful: a set of heuristics, principles and conceptual tools for navigating a world that resists prediction, rewards robustness and punishes the confusion of maps for territories. The project is uneven in places, occasionally self-indulgent and sometimes wrong in specifics. But the core insight, that we systematically underestimate the role of the improbable and overestimate the reliability of our models and our experts, is correct, important and still insufficiently absorbed by the institutions and individuals who most need to understand it. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit tiwaryshailesh.substack.com (https://tiwaryshailesh.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

Contact Dear Son..

Guest appearances

Does not typically book guests

Based on episode analysis; this does not confirm that the show is currently accepting guests.