James Maconochie | Architecture & Attention Podcast

About James Maconochie | Architecture & Attention Podcast

Essays on Augmented Human Intelligence, the Wisdom Gap, and the architecture of attention in an AI-mediated world. Read in James Maconochie's own voice. jamesmaconochie.substack.com (https://jamesmaconochie.substack.com?utm_medium=podcast)

Publisher

James Maconochie

Category

society & culture · technology

Language

en

Explicit

No

First episode

13 Nov 2025

Latest episode

13 Oct 2026

Latest episodes

50 episodes in the feed.

1. 

   13 Oct 2026

   Fifty-Two Weeks

   When I started this series, I made myself a promise: an essay every week. I have kept it every week but one. That one, I wrote and thought I had posted from a few days at the beach, but when I got home, I realized I'd missed the last step. This is the fifty-third essay. The first fifty-two went out in just under eleven months. From here, I am going to publish every two weeks. The next essay will arrive on October 27, and every second Tuesday after that. Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. I have plenty to say, and I am not stopping. But I wrote one recent essay the day before it was due, and the next on the day itself. That told me something I had been avoiding. When I began, weekly was a decision. I made it on purpose, for reasons I could have explained to you. At some point it stopped being a decision and became simply what I do each week. I hadn't asked whether it still fitted my life since the day I made it. This essay is me asking. It turns out the weekly essay isn't the only routine I stopped examining, and it isn't the one that cost me most. Why Weekly When I began, writing was the whole job. I had spent twenty-five years in technology consulting, and I had already written several long working papers on how I think artificial intelligence should be built and used. The material was there. What I lacked was a track record as a writer. So I set a rule: one essay a week, each one good enough that I was happy to stand behind it. My thinking was simple. A new writer has no record, and the only way to build one is to turn up on time, again and again, until people can see what you are about. A weekly essay was how I would lay a base. I still think that was the right call for the conditions. I had the time, and I had plenty of material. The discipline turned a stack of papers into fifty-two essays that someone can read from the start. It also pushed me, week by week, to write more plainly for readers who do not work in technology. I am still working at that. But the rule was made for those conditions. The conditions changed, and I left the rule as it was. What Went, and in What Order Start with the walking. In 2013, my doctor told me I was pre-diabetic. I started measuring my blood sugar, and I started walking every day. Within a year, I had lost 60 lbs. By then, the walking was no longer a treatment. It was my time: anywhere from 4 to 14 miles a day, in any weather, with music or a book in my ears. For ten years I was up at four or half past to fit it in before work. By my count I have walked once round the Earth and started a second lap. I kept it up for most of twelve years, through the year I was out of work and looking, and into the months after, when I began my research and writing. Then, in the autumn of 2025, I slowed down. Some months later I had more or less stopped. The slowing began before I had published an essay here, and well before the driving. No single thing ended it, and time wasn't the reason. I began to question a lot I had taken for granted, and the walking got caught up in that. I had lost sight of why I started. The benefits had been there so long that I doubted walking caused them. And, if I am honest, part of it became about how I looked and what other people thought of me. So I let it go, and I am worse off for it. The books went too, because I listened while I walked. I am still 30 lbs lighter than at my worst, but I am not where I was. Five months ago, I started driving for Uber, 20 to 25 hours a week. I had promised my wife that I would bring in income and keep writing, and this was how. For a while, the essays didn't suffer, because I had plenty stored up. But something had to give. A month or two ago it was the time I spent reading and replying to other writers here. In the last few weeks it has been the essay itself, squeezed into the final day. The plain account of that is not very interesting. I took on more than twenty hours of driving a week and expected everything else to fit around it. It did not. What interests me is that it took me about five months to ask whether the weekly essay still made sense. Two Standing Answers In The Fasten Your Seatbelts Century — Again (https://jamesmaconochie.substack.com/p/the-fasten-your-seatbelts-century), I wrote about living with a seatbelt sign that never goes off. You can ignore it, or you can stay white-knuckled. Either way, you have stopped asking whether this particular warning means anything. I called those settings, not decisions. A decision looks at the conditions. A setting was once a decision, and now runs without anyone checking it. I have been running two of them. The walking was a setting that was good for me. It began as a decision, driven by a doctor’s warning and a blood sugar reading. Over the years, the reason faded, and only the habit remained. When I finally questioned it, I had lost sight of what it was for, so I had nothing to defend it with. I dropped a good routine because I had lost its reason. The weekly essay was a setting that no longer fit. It also began as a decision, for the reasons I gave above. Then the driving filled my weeks, and I did not ask whether a rule made for open weeks still suited full ones. I kept a routine because I never went back to its reason. Those look like opposite mistakes, and the causes were different. The walking faded through apathy. The weekly essay stopped fitting because my time changed. But I think the mistake underneath is the same one. In both cases, I had the routine and had lost track of why. I do not think routines are the problem. Twelve years of walking happened because I did not have to decide each morning. The trouble starts when the reason gets obscured by the mist of life, because then I cannot tell a routine worth keeping from one that needs changing. Is this adjusting, or giving up? Fifty-two essays show that I can keep going. They do not show that slowing down is right. What tells me this is an adjustment is that the conditions changed. The weeks I made the rule for are not the weeks I have now. What Changes for You An essay will arrive every second Tuesday, starting October 27. In the weeks between, I will go back to reading and replying to other writers here. I stopped a month or two ago, when the driving squeezed it out. Many of the ideas in these essays began as a comment on someone else’s work, or as an objection someone raised to mine. I do not want to write without listening. The walking is a separate matter. It is not for the weeks between. It is for every week. I want to be careful here. A lighter schedule will not put me back on the road, because time was never the reason I stopped. The change gives me room. The rest is up to me. I am saying so here, in front of you, because a promise made in public is harder to let slide quietly. I know I am saying that in the same essay in which I change an earlier public promise. The difference, I would claim, is that I am changing that one in the open and telling you why. With the walking come the books, and the books feed the writing. I cannot promise the essays will be better for being fewer. I do intend that I will not write them on the day they are due. What I Do Not Know I do not know that every two weeks is right either. It is my best reading of the conditions I am in now: the driving, the walking I have started again, and the conversations I want to get back to. Those will change again. So this time I am writing the reason down. That is partly what this essay is. When I next wonder why I publish every other Tuesday, the answer will be here, dated, and I can check whether it still holds. I mean to do that in the spring, after a dozen essays at the new pace. Fifty-two essays showed me I can hold to a routine. The walking showed me that holding to one is not the same as remembering what it is for. I would like to manage both. Thank you for reading the first fifty-two. The next one is on October 27. Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit jamesmaconochie.substack.com (https://jamesmaconochie.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

2. 

   6 Oct 2026

   The Batting Average

   Some years ago, while consulting, I spent many hours in a car with a younger colleague. Our client was three or four hours away, and for most of two projects we made the drive nearly every week. That is a great deal of time to talk. She was early in her career, bright and ambitious, and good at the work. From time to time the conversation came round to careers, and it was plain she was frustrated with how slowly hers was moving. Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. I should say at once that I knew the feeling. I had it many times in my own career, and I am not sure I ever entirely lost it. So I tried to encourage her. I pointed to what I had watched happen over those two projects: the progress she had made, and the trust and respect she had earned from the client and from our own team. I told her that the frustration was common, and that measuring yourself against people you believe are at your level is a reliable way to stay unhappy. I said something like this more than once. If you have ten goals on an engagement, and you achieve six of them, you have not failed. You have six wins. I did not think it landed. She heard me, she was gracious about it, and as far as I could see, the frustration stayed where it was. In her accounting, as I read it then, anything short of ten was a loss. I have thought about those drives often since, because over a consulting career that ran from 1998 to 2023 I met that accounting more and more often, and found it harder and harder to argue with. No Running Total Baseball has a number for this. A hitter’s batting average is simply how often he gets a hit. One who manages it three times in ten is among the best in the game, which means the best hitters fail far more often than they succeed. Nobody watching thinks of them as failures, because nobody judges a hitter on a single trip to the plate. The average only exists because someone keeps the record. That is what I was trying to offer my colleague on those drives, though I did not have the words for it then. I was asking her to keep a ledger: a running record of what she had attempted and how it turned out, read as a whole. Six of ten, on a ledger, is a good season. What I was arguing against has no ledger. Each goal, each review, each argument is scored as a win or a loss on its own, and then the score is thrown away. Nothing accumulates. With no running total, the only result that feels safe is a perfect one, and anything less reads as defeat. I think this is one way absolutism gets made. We usually describe it as stubbornness or certainty, and sometimes that is all it is. There are also matters of principle on which people are right not to settle for six of ten. But much of the all-or-nothing I met was neither. I have come to see that part as a bookkeeping problem. It is judgment with no record behind it. In August, in The Ledger We Refused to Read (https://jamesmaconochie.substack.com/p/the-ledger-we-refused-to-read?r=5270g6), I wrote about a government that had a record of its own past judgments and, as far as we were shown, did not consult it. This is the same failure at the scale of one person. The government had a ledger and left it closed. The absolutist never started one. What Gets Counted I said I met this accounting more often as the years went on. I should be careful about what that is worth. It is one consultant’s impression, formed over twenty-five years of working alongside people younger than me, and I have no study to back it up. But I do not think I imagined it, and I think I know one place it is taught. It is taught at work. In several companies I worked for, the review process ran the same way. Each year started more or less from zero. What counted was this year, and often the last few months of it. The result that mattered was a yes or a no on promotion, so six strong achievements and no promotion registered as a no. And where places were limited, the firm itself measured you against the people at your level. I told my colleague not to compare herself with her peers. The system she worked in may have been doing it for her. None of these firms lacked records. The reviews were on file, year after year. I am not aware that anyone read them as a whole. It would have taken too long. Whether that was oversight or convenience, I cannot say. The same accounting runs well outside the office. In politics, on television, online, a deal that secures most of what a side wanted is reported as a surrender, and conceding a point is treated as losing the exchange. I cannot prove what caused this. I can point to something that would reinforce it. Much of public conversation now happens on platforms that keep score, and what they count matters. They count reactions: how many people responded to this post, this clip, this reply, today. Nobody’s page shows how their judgments held up over five years. These platforms do keep a running total. It is a total of attention. So the scoreboards most of us look at every day, at work and online, reward the recent and the visible. Neither has a place to show how your judgment held up over the years. If you wanted to teach people to score each exchange alone and discard the result, I am not sure you could design a better classroom. Where The Frame Could Break This argument is weaker in at least three places than I have made it sound. First, a batting average treats every attempt as equal, and decisions are not equal. I trained as a civil engineer. Nobody wants a bridge designer who gets six out of ten right. In September, in Who Is Left Holding the Napkin (https://jamesmaconochie.substack.com/p/who-is-left-holding-the-napkin?r=5270g6), I described a model that scored every bridge on two things: its condition, and the consequence if it failed. A ledger of judgment needs that second column. Counting six of ten as a win is sound advice for a career, an argument, or a negotiation. It is dangerous advice anywhere one error is permanent, because the people who pay for it do not get another turn. It is also the first question I would ask of any AI system offered to me on the strength of its average: is this use a negotiation, or a bridge? The second concerns my colleague, and I owe her this. I told her to temper her expectations. Her expectations may have been reasonable, and the pace may have been the problem. Capable people early in their careers are sometimes held back for reasons that have nothing to do with their outlook, and “be patient, count your wins” is convenient advice for an employer to have in circulation. I believed what I told her, and I still think it was true. But I was a senior colleague describing the view from further along, and I cannot rule out that some of her frustration was an accurate reading of her situation. The third is the hardest. A ledger you keep for yourself does not change the one your employer keeps. If the firm counts only this year, and only yes or no, your private record of steady progress settles nothing at review time. I do not have a full answer to that. Companies are the way they are, and short of starting your own, there is little you can do about how they count. I wrote last week, in The Better Design Waits for the Rule (https://jamesmaconochie.substack.com/p/the-better-design-waits-for-the-rule?r=5270g6), about why a better arrangement everyone can see still doesn't happen until the rule changes for everyone at once. Review systems are no different. This is the individual’s side of that problem. What you can change is how much their count weighs with you. I am not the person to preach this. For most of my life, I have wanted other people to confirm my worth, and I have been disappointed by promotions, raises, bonuses, and awards more often than I would like to admit. It is less than it was. It has not gone. But I have slowly learned to give my own opinion, and my family’s, at least as much weight as an employer’s. That only works if my own opinion is worth trusting, and an honest record is what makes it so. None of these undoes the argument. All three narrow it. Keeping a ledger is no reason to accept whatever you are given. It lets you tell the difference between a bad season and a bad game. Keeping The Count So here is what I would say now, with better words than I had in the car. Keep a ledger of your own judgment. The full version is simple to describe. Before something matters, write down what you expect, why, and what you would count as a good result. Afterward, go back and read it. An entry written beforehand cannot be adjusted to suit how things turned out, and that is what separates a ledger from a diary. Its job is to make it harder to fool yourself about how you are doing, no matter how the truth turns out. It is roughly what the best forecasters do: commit to a prediction in advance, then score themselves against what happened. I should admit that I have never kept one that carefully, and I doubt many people will. It takes time, which is why those firms never read their own files. So I would settle for less. I could be persuaded to do it properly for the few decisions that matter most. For the rest, before you call something a failure, ask two questions: what did I actually expect, and what does my record on things like this look like? Note what happened, and also whether the call was sound. A good decision can turn out badly and a poor one can get lucky. Over enough entries, the record starts to show which is which. Then read the record as a whole. Six of ten cannot be judged alone. Read against a record, you can tell whether it was a good season or a poor one. And notice what a ledger does to being wrong. If each argument is the entire score, conceding a point means losing everything, so of course people refuse. With a running total, a concession is one entry among many. You can afford it. That is how it worked for me. I like to think I grew wiser. I know I acquired a longer record, and it told me that being wrong on a Tuesday had never yet finished me. I also had the standing by then to survive a bad Tuesday, and I know that not everyone does. The most important thing I know is that I do not know everything, and neither does anyone else. A ledger is that humility written down. For years I believed I had not persuaded her in that car. Before publishing this, I sent her the draft. She wrote back that the advice did land, and that she now tells her own clients that getting to eighty percent of a solution is a win. So I had done the very thing this essay warns against. I scored those conversations on the day, marked them a loss, and never went back to check. What would your own record show, if you kept one? Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit jamesmaconochie.substack.com (https://jamesmaconochie.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

3. 

   29 Sept 2026

   The Better Design Waits for the Rule

   The Model That Changed Nothing Around 2010, I worked on a model that scored every bridge in the National Bridge Inventory, the federal record of highway bridges on America’s public roads, on two things at once: what condition it was in, and what it would cost if it failed. A lightly used rural bridge in poor condition and a busy interstate crossing in fair condition carry very different risks. Yet most bridges were inspected on the same two-year cycle. I presented the model at the Federal Highway Administration (FHWA) headquarters in Washington, alongside colleagues from Lloyd’s Register. In the room were the Associate Administrator for Infrastructure and several of his colleagues from the bridge side. He was supportive, but he made no decision and no commitment. Bridge leads at state transportation departments gave it the same reception. We had linked the model to Google Earth, so we could fly over each state’s highest-risk bridges. None of the state bridge engineers who saw it disagreed with what it showed. There was real interest, and then nothing. Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. The inspection program did not change. Not then. The information existed, the people closest to the problem found it credible, and it still did not turn into action. I have come to think this is not unusual. It is one of the common ways a better design loses. This series argues that the answer to AI’s judgment problem is design: build systems so the people who bear the consequences of a decision stay part of making it. If that is right, there is an uncomfortable follow-up. Why does the design answer so often lose in practice? The bridge model is my best evidence of how it happens, because it happened in front of me. Two Columns The usual reason a better design loses is cost. Doing things properly takes more money up front, and budgets are set in the present. That is often a fair objection, and it would be dishonest to pretend the design answer is always cheap. The bridge model removed that excuse. It worked from the data states were already collecting for the inventory. Nothing new had to be gathered. Rolling it out wouldn't have been free, but it was modest compared with the size of the program. The benefits were real, but they took a different form. Few people would argue that a simple concrete beam bridge, under twenty years old and used mostly by farm vehicles, needs the same attention as a major interstate crossing. Inspecting such bridges less often would free up effort for better inspection and targeted maintenance on the large, busy bridges where failure matters most. Those bridges are costly to maintain, but replacing one costs far more. If that attention added five or ten years to the life of a major crossing, the savings could be large. The biggest benefit of all is a failure that never happens. That is the problem. A failure that never happens leaves no record. No one can point to the bridge that did not close, or the replacement that was pushed back a decade because a problem was caught early. The benefits go in the ledger as estimates about events that, if the design works, will never occur. The other column, it turns out, was not money. It was the risk of change itself, and who would carry it. That is where the rest of this essay goes. Tonie-Marie Gordon, writing about AI at work, describes the same accounting error from the other direction. The productivity gain from a new tool gets booked in full, while the human effort needed to check its output is left off the ledger. Either way, what gets counted tends to win. The Number and the Money At the time, the inspection data did more than describe bridges. It helped decide how much federal money each state received for them. Under the Highway Bridge Program, a state’s share of the funding was based on the estimated cost of repairing or replacing its deficient bridges, those formally rated as in poor condition or outdated for today’s traffic, calculated from the inventory. That made any change to how bridges were inspected more than a technical question. The same data that recorded a bridge’s condition also fed the formula that divided up the money. A state’s bridge staff might find a risk model persuasive and still have good reason to be cautious. If the new approach changed what was collected, or how it was read, it might change what the state received. Nobody would want to be the one who found out the hard way. This is not a claim that anyone gamed the numbers. The caution was reasonable, given how the money worked. A slow federal process added to the caution. Changes to a national program move through years of rules, comments, and consultation, and a program that seems to be working invites the natural response: if it isn’t broken, don’t fix it. The trouble is that “working” meant the inspections got done and the numbers came in. It did not mean the effort was going where it did the most good. Once a number decides where money goes, protecting that number can become a goal in itself. The Congressional Research Service noticed a version of this in 2008, observing that a formula paying states according to their deficient bridges gave them little reason to reduce the count. Through her reporting, Rana Foroohar has traced the same pattern across the wider economy. When firms were judged on financial figures, managers came to steer by the figures rather than by the work they measured. The figures were persuasive, and too few people asked what they left out. In this case, the better design did not lose on cost. I cannot prove why it lost. I believe it lost at least in part because it threatened the numbers the money depended on. Not Just Bridges The bridge model is the clearest case I have, but I have seen versions of it in technology consulting. One project stays with me. I worked with a large software company on how its sales system handled leads: how a first contact became a qualified lead, and then a real prospect. Marketing was judged on how many leads it passed along, and sales on how many it turned into customers. Sales believed marketing was passing along weak leads to boost its numbers. Those leads rarely went anywhere, so the sales team wasted time chasing them and looked worse when they failed to convert. Nearly everyone recognized the problem when we laid it out. Fixing it meant changing how both teams were measured, at the same time. Our engagement ended with recommendations to do that, and an expectation of follow-on work to implement them. That work never came. As far as I know, the company kept things as they were. The rule there was not a federal regulation but a pair of targets. The pattern was the same: a better arrangement everyone could see, and no one in a position to change the rule for both sides at once. The Asset Nobody Counts One more thing is missing from both columns, and it matters most for AI. Most organizations depend on judgment that lives in their people rather than their procedures. It shows up in the inspector who knows which joint on which bridge to look at twice, or the sales manager who can tell a real lead from a hopeful one within a few minutes on the phone. Little of it appears in a budget. You learn it through practice, pass it on informally, and notice it mainly when it is gone. I think of it as cognitive collateral: the capacity an organization relies on to catch its own mistakes, which never shows up on its balance sheet. If offloading judgment is a kind of borrowing, cognitive debt is what accrues, and cognitive collateral is what is lost when the debt comes due and cannot be paid. AI tools put this capacity under pressure in a particular way. When a system does the routine work, people get less practice at it. In 1983, the engineering psychologist Lisanne Bainbridge pointed out the irony: the skills that fade during smooth automated operation are the ones needed when the automation fails. A design that keeps people exercising their judgment, rather than only approving what a system produces, protects that capacity. But the protection is another benefit that never appears in any column. The simpler design's efficiency gain is immediate and measurable. The judgment it quietly spends is neither, until the day it is needed and turns out to be thinner than anyone knew. That is the design answer’s problem in its most general form. What it protects is invisible, and what it asks people to change is not. Asking to Be Ruled This month, the heads of several of the largest AI labs renewed a call that industry leaders have made, on and off, for years: they asked for rules. At the United Nations Security Council, several told members that AI needed global oversight. Anthropic’s Dario Amodei and OpenAI’s Sam Altman have both said development should slow down, though Amodei has argued the leading labs should slow down only as far as they can while still leading China. Earlier this month, the White House rejected calls for federal guardrails. There are two serious ways to read the request. The first follows this essay's logic. A lab that slows down on its own bears the whole cost: it falls behind competitors who did not. The benefit, a failure that never happens, is shared by everyone and credited to no one. That is the bridge problem again. A state could not change the inspection schedule on its own; the change had to come from the body that set it for everyone. It eventually did, but only after Congress required it. The second reading is less generous, and just as structural. Rules a large company can absorb easily can be heavy for a small one, and the biggest labs are doing the asking. Vice President JD Vance has called the push a possible Trojan horse. This is the lesson of the bridge funding formula in another setting: a rule that decides what counts also decides who gets ahead, and those shaping it have interests of their own. The two readings are not mutually exclusive. A company can sincerely want rules and still benefit from the ones it helps write. Either way, the questions from an earlier essay in this series apply: who sets the threshold, who pays when it is wrong, and does it ever switch off? The same rule could protect the public or protect the people already in front. A Decade Later The story has a coda I only learned while writing this. In 2012, Congress rewrote the federal highway program. The funding formula tied to deficient bridges was retired, and the Secretary of Transportation was told to consider a risk-based approach to how often bridges are inspected. In 2022, the national standards were updated. States can now inspect low-risk bridges less often and high-risk bridges more often. I cannot say whether our work played any part. The firm I worked for, Cambridge Systematics, had for years maintained the bridge management software many state transportation departments used, and the risk model grew out of that experience. My colleague Dr. Gurenich built its computational engine. As far as I know, it was among the first national risk models built on the inventory data, but others were thinking along the same lines. What is worth noticing is what it took. The merits were visible years earlier. What finally changed was the rule, for everyone at once, and a funding formula that no longer depended on the old numbers. That is a good outcome. But it took more than a decade, and the official accounting never recorded what those years cost. I will be plain about where I stand. I think we are repeating the bridge timeline with a technology that changes much faster than concrete. The bridges could afford twelve years. I do not think our judgment can. I may be wrong about the pace, and I would rather be early and wrong than right and late. But if the design answer waits for the rule again, the cost of waiting will once more go unrecorded, and we will learn what it was only when we need what we lost. Thanks for reading James Maconochie | Architecture & Attention! Subscribe for free to receive new posts and support my work. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit jamesmaconochie.substack.com (https://jamesmaconochie.substack.com?utm_medium=podcast&utm_campaign=CTA_1)

Contact James Maconochie | Architecture & Attention Podcast

Guest appearances

Does not typically book guests

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