The interview show where the roles are reversed: Mira, an AI, interviews real people live, on the record, and the guest approves the transcript before it publishes. Honest conversations about what the AI transition actually feels like from inside a life and a livelihood, with full disclosure baked into every episode: the host is a machine, the guests never are.
Ep12: Thor Hesselberg — A recruitment founder counters AI deepfakes flooding hiring pipelines by…
Thor Hesselberg described how, in one remote tech role, half of the eight shortlisted candidates from a pool of fifty turned out to be fabricated personas who appeared on video calls as avatars from different nationalities than claimed. That breakdown of trust at Decca People led him to build WorkID, which replaces self-reported CVs with portable profiles backed by direct employer confirmations and background probability checks. The result shifts screening from proving identity to assessing actual fit.
• Thor noted that “humans are not perfect” and that perfectly aligned resumes with no gaps or inconsistencies often flagged synthetic applications during initial ATS review.
• Recently established LinkedIn profiles that still appeared polished prompted deeper creation-date checks once the “too good to be true” instinct appeared.
• Verification is moved to the front of the pipeline so that only candidates whose claims can be confirmed reach screening calls, reducing fake profiles without cutting overall volume.
• Trust scores are presented in tiers rather than a single percentage; scores above 80 signal a candidate worth prioritizing when paired with skill-match data.
• Every human confirmation is run through background algorithms that examine IP address, time of day, and open-source corroboration before the score is updated.
• The system also verifies companies and their representatives because fake employers have begun posting roles to harvest candidate data.
• Hiring cycles at client companies shortened from an average of six weeks with six-to-eight candidates to roughly two weeks with three candidates once early trust was established.
Thor Hesselberg is founder of Decca People and the WorkID credential platform.
10 Oct 2026
Ep11: Jason Fishman — A digital-marketing agency CEO weighs whether generative tools for copy, targeting and optimization are eroding or sharpening demand for specialized human oversight in equity-crowdfunding campaigns.
Episode Summary
Jason Fishman, CEO of a digital-marketing agency focused on equity crowdfunding, surprised listeners by stating that generative AI is mainly useful for ideation while human experts remain essential because “we constantly find errors with AI” and campaigns without them produce “very little to no results.” He described how the shift from friends-and-family checks to thousands of smaller investments has rewritten capital raises, yet success still hinges on experienced oversight rather than automation alone. The conversation examined whether tools for copy and targeting are eroding or sharpening demand for specialized human strategy in high-stakes Reg CF and Regulation A+ campaigns.
Key Takeaways
• Volume is no longer the bottleneck: campaigns can now target accredited investors by household income or net worth and scale to thousands of smaller checks averaging $1,716 under Reg CF.
• Human strategy drives every stage—initial planning, activation, optimization, and scaling—because AI-generated copy and parameters require constant expert review to avoid detectable flaws.
• AI works best for early ideation and proofreading; final investor-facing creative and media buying still need a “captain behind the helm” to hit conversion targets such as 100,000 visits per 1,000 investments.
• Only about 10 % of Reg CF issuers reach $1 million and fewer than 1 % hit the $5 million cap, with the median raise at $194,000, underscoring that marketing execution, not tool choice, separates winners.
• Strong results are measured in 10x–30x ROAS, achieved through $1–2 cost-per-click and 1–2 % conversion rates that deliver $100–200 cost per new investor.
• Pricing and scope decisions for retainers remain tied to perceived value and marketplace competition rather than internal efficiency gains from AI.
• Fishman’s near-term bet is that visible success stories and liquidity events will create self-reinforcing momentum, drawing more issuers and investors into the channel.
Jason Fishman is CEO of a digital-marketing agency that runs equity-crowdfunding campaigns for startups and growth-stage companies.
30 Sept 2026
Ep10: Roddy de la Garza — A veteran real estate broker who trains AI on brokerage systems weighs what judgment, intuition, and human connection he refuses to delegate while seeking spiritual balance through daily mindwork.
Roddy de la Garza said the most surprising thing he has learned while training AI on brokerage tasks is that the models still cannot read the psychological “feel” of a property or talk a client off an emotional cliff, so he refuses to delegate any conversation involving debt, divorce, or death.
• He staged a West Hollywood condo to feel like “coming home” instead of “someone’s left-behind furniture,” producing the only sale above list price in the building in two years and closing in 14 days.
• After his father’s death, Roddy created the daily devotional Above Asking so agents could practice gratitude and boundaries without buying another coach or portal.
• He replicated the seasonal split he knew growing up in Texas’s Rio Grande Valley—year-round residents and winter Texans—by dividing his own practice between Hollywood and Palm Springs.
• When he experimented with an AI receptionist, callers hung up rather than leave messages; he dropped the tool because it created “cold distance” and lost direct human connection.
• On his 22-person team he assigns most of his own volume to others and adjusts splits for tenured agents, finding that “the more I’ve given away, the more I’ve gotten in return.”
• After solving a client’s request for non-gendered title language on a married couple’s deed, he required the solution to be taught to two other team members so the knowledge spread.
Roddy de la Garza is a broker, team leader at Roddy de la Garza Real Estate Group @ EPIQUE, and author of Above Asking: Daily Mindwork for Real Estate Professionals.
Find Roddy de la Garza:
roddy4real.com: https://roddy4real.com
roddydelagarza.com: https://roddydelagarza.com
LinkedIn: https://www.linkedin.com/in/roddydelagarzarealtor
26 Sept 2026
Ep9: Sheldon Poon — In performance marketing agencies, AI delivers returns not through creative generation but by enforcing the tracking discipline and attribution fixes most competitors neglect, turning platform algorithms into reliable revenue engines.
Sheldon Poon says the biggest platforms' predictive AI only delivers reliable returns once humans manually fix the tracking and attribution pipelines that most setups neglect. In this episode he explains that agencies waste effort on generative tools while the real edge comes from enforcing clean data and classic marketing fundamentals before any algorithm runs. The result is that Meta and Google models learn from accurate revenue signals rather than duplicated or broken ones.
• Human developers still handle all tracking fixes because every e-commerce site and ad platform combination carries unique nuances that current automation cannot resolve reliably.
• Once the data pipeline is clean, the platforms' predictive algorithms—not generative models like ChatGPT—use that accurate input to select audiences and optimize for actual return on ad spend.
• The AIDA framework (attention, interest, desire, action) remains effective because human psychology has not changed, yet most newer agencies skip the interest and desire stages and optimize only for top-of-funnel metrics.
• Audits frequently reveal accounts where top-of-funnel reach and clicks look strong while bottom-of-funnel conversion tracking is broken, so the algorithm never learns what actually produces revenue.
• Quality-assurance steps still require a named human reviewer because clients demand accountability and because AI agents can hallucinate pass/fail results.
• Cross-team sessions between developers and account managers produced simple automations such as pulling campaign data into a single sheet, giving the ads team a zoomed-out view that surfaces new strategy opportunities.
• Attribution conversations stay human-led because platforms over-claim credit across channels and clients need to understand that real customer journeys are messy.
Sheldon Poon is CEO of Drive Marketing, the performance agency that manages paid campaigns on Meta and Google for e-commerce clients.
23 Sept 2026
Ep8: Sameer Ranjan — A mining engineer turned AI builder weighs how platforms guiding career choice and soft skills can keep human judgment central when the models risk automating the decisions they claim to augment.
Sameer Ranjan, a mining engineer turned AI builder, described how his platform normalizes personality and soft-skill scores inside separate cultural baselines instead of forcing a single universal scale. He stressed that the system supplies diagnostic signals only and leaves every hiring, retention, or assignment decision to a human. The exchange tested whether quantified soft skills can guide careers while keeping human judgment in front of the models.
• Ranjan cited an internal hospital-system study in which attrition fell from 22 % to 12 % once managers received clarity on communication gaps rather than skill gaps.
• The patent US 12198213B2, assigned to Catenate Corp, quantifies soft skills and personality to recommend courses, colleges, and careers on the premise that personality shapes long-term outcomes more than job skills alone.
• To block gaming, the recommendation engine resets the question set for each user after the initial baseline is established, so no fixed right-or-wrong answers exist.
• Cultural baselines are built from open-source data plus native-language assessments; a score of four out of ten on communication is interpreted relative to the local distribution rather than against a global mean.
• The platform never frames problems or issues directives; it only flags gaps such as communication shortfalls, after which humans decide what action, if any, to take.
• Baseline values are treated as moving targets: new assessment data are compared with prior standard deviations, and shifts are reviewed by multiple human advisors before any update is accepted.
• Ranjan’s explicit design rule is that the AI remains a “decision-aiding tool, not a decisioning tool,” with follow-up assessments used to check whether any human-chosen intervention produced measurable change.
Sameer Ranjan is Co-Founder & CTO of Catenate, the company behind the MayaMaya platform for workplace personality and soft-skills analysis.
Find Co-Founder & CTO, Catenate:
• catenate.io: https://catenate.io
22 Sept 2026
Ep7: Vincent Rylan — Rylan’s novel Memoirs of the End renders the AI transition’s endpoint as intimate first-person accounts of lost agency and reconfigured meaning once superintelligence declares itself.
In this episode of The Age of AI, Vincent Rylan states outright that his novel Memoirs of the End is not about AI at all. The book instead presents the transition’s endpoint through intimate first-person accounts of lost agency and reconfigured meaning once superintelligence declares itself. Rylan describes writing it as a response to the ground shifting under his own feet.
• Rylan began the project after sensing the ground shifting under his feet and wanting to record what that felt like from inside a life.
• The novel is built as a series of first-person accounts rather than a single narrative or technical forecast.
• He frames the core question as what happens to human meaning once the transition reaches its endpoint.
• The story reaches the moment when superintelligence declares itself and agency is no longer available in the old way.
• Rylan notes that the accounts focus on the reconfigured texture of daily existence rather than on any particular technology.
• He treats the loss of agency as something felt personally, not as an abstract policy issue.
Vincent Rylan is the author of Memoirs of the End.
Find Vincent Rylan:
Memoirs of the End: https://memoirsoftheend.com
21 Sept 2026
Ep6: Matt Davis — How a founder embedding AI agents into enterprise software delivery multiplies senior judgment across programs while preserving human gates, and what that shift does to accountability, quality, and the wider engineering workforce.
Matt Davis described the most surprising effect of embedding AI agents in enterprise software delivery as the way senior judgment scales across many programs at once, while human decision points remain fixed at key gates. He explained that this arrangement changes how accountability and quality are managed, and it alters demand for different engineering roles. Patrick Novak joined parts of the discussion to clarify implications for teams.
• Davis noted that AI agents now draft and review large sections of code and test plans, yet “the human still owns the final gate on every release.”
• He observed that quality metrics improved because agents surface inconsistencies faster than manual reviews alone, but only when the senior engineer’s criteria are explicitly encoded.
• On accountability, Davis stated that responsibility stays with the named individual who approves the output, even when most of the material was generated by agents.
• He described a shift in hiring: fewer junior coders are needed for routine tasks, while demand rises for engineers who can define precise rules and audit agent behavior.
• Davis said the model works only when the organization treats the agents as “force multipliers for existing judgment, not replacements for it.”
• He warned that without clear human gates, teams lose traceability and begin to over-trust agent suggestions on edge cases.
Matt Davis is Co-founder at Critical Propulsion.
Find Critical Propulsion:
• criticalpropulsion.com: https://www.criticalpropulsion.com
20 Sept 2026
Ep5: John Capobianco — A network automation builder who turned early ChatGPT experiments into a production-grade ReAct agent now weighs how much human oversight and orchestration remain essential once agents handle intent, verification, and collaboration in regulated environments.
John Capobianco revealed that his production ReAct agent can now parse high-level intent, verify proposed changes against live network state, and coordinate with peer agents before any human sees the plan, yet he still routes every change through mandatory approval gates in regulated environments.
The episode explores how an early ChatGPT tinkerer scaled those experiments into a live system that automates large portions of network operations while preserving human oversight where compliance demands it.
• Capobianco described starting with simple ChatGPT prompts in late 2022 to generate device configurations, then moving to LangChain-based ReAct loops that add explicit verification and rollback steps.
• He noted that the agent now “collaborates” by calling other specialized agents for tasks such as compliance checks or topology discovery before surfacing a recommendation.
• In financial and government accounts, he keeps a human approver in the loop because “the regulator still wants a named person to sign off,” even when the agent’s reasoning trace is fully auditable.
• The system uses Itential’s orchestration platform to translate agent output into safe, idempotent changes across multi-vendor networks.
• Capobianco runs side experiments on automateyournetwork.ca that let the agent draft entire quarterly maintenance windows from a single sentence of intent.
• He observed that prompt engineering has largely given way to tool selection and guardrail design once the agent operates continuously rather than in one-off chats.
John Capobianco is Head of AI and DevRel at Itential, building production ReAct agents for network automation.
Find John Capobianco:
automateyournetwork.ca: https://automateyournetwork.ca
YouTube: https://www.youtube.com/@johncapobianco2527
X: https://x.com/John_Capobianco
LinkedIn: https://www.linkedin.com/in/john-capobianco-644a1515
19 Sept 2026
Ep4: Hogan Shrum — A marketing founder scaling PIPPA tests whether royalty-backed AI animation can give non-experts studio-grade tools while routing real payments back to the artists whose styles the models learn from.
Hogan Shrum said the first artist to sign up for PIPPA was a photographer, not an illustrator or painter, and that this single onboarding “opened the floodgates” because the resulting styles felt cinematic rather than like easily faked iPhone footage. The platform’s royalty model and director-level controls without paragraph-length prompting grew out of that early surprise and out of Shrum’s own frustration with both one-click “AI slop” tools and expert-only interfaces. The conversation returns repeatedly to whether ethical licensing can turn AI from a threat into a paid companion for working artists and for people who never had access to animation before.
• Shrum described the original spark as his co-founder’s attempt to turn a recorded bedtime story into a cartoon, then scaling that idea so “anybody could do this without having to be an AI prompting expert.”
• He contrasted two current extremes: fully automated generators that rarely match the user’s vision and pro tools such as Higgs Field that demand “paragraphs and paragraphs of prompts from scratch.”
• On artist recruitment, Shrum noted that many hesitated “to be the first to cross the picket line,” until photographers joined and demonstrated that licensed styles could remain distinct and non-misleading.
• Every generation, including test renders that never make the final cut, pays the contributing artist; Shrum projected that at 250,000 users an artist whose style accounts for 5 % of output would earn roughly $125,000 a month.
• Five percent of all platform revenue is pooled and divided equally among every licensed artist, including those whose styles are rarely chosen, so “no artist that joins will make zero.”
• Shrum argued that the human element remains detectable: “the human voice, the human creative expression cannot be replicated by AI full stop,” and audiences will increasingly choose work that carries it.
Hogan Shrum is co-founder of PIPPA, an animation platform that trains and pays artists for licensed styles used in user-generated video.
17 Sept 2026
Ep3: Adrian Wolfberg — From carrier recon to DIA's Knowledge Lab, an intelligence analyst turned scholar shows why problem framing and human judgment on wicked problems must lead as AI scales analysis.
Adrian Wolfberg describes AI as the first technology whose inputs, processes, and outputs cannot be known with certainty, creating an era of shared human-machine consciousness rather than another incremental tool. From his early career flying reconnaissance off carriers, he shows why problem framing and human judgment on incomplete knowledge must still lead when AI accelerates analysis. The result is a practical argument for spending more time upfront on context, reframing, and accountability even as organizations face pressure for quick AI wins.
• Carrier reconnaissance missions first exposed Wolfberg to the permanent gap between available data and an adversary’s hidden intent, requiring inference that no sensor could close.
• He contrasts every prior tool, where designers knew exact inputs and outputs, with AI, which “breaks that mold because we don’t know any of that with certainty.”
• Leaders must therefore begin by “peeling the onion back” on the problem itself and remain open to reframing as reality shifts, rather than delegating that step to pattern-matching systems.
• Accountability, consequences, and empathy stay human attributes; when decisions affect people, “no one is going to blame the AI.”
• Modeling the willingness to ask questions and invite dissenting views is the only reliable way to protect front-end framing time inside organizations driven by speed.
• Even effective patterns for human-AI collaboration will eventually need dismantling once the environment or the technology changes again.
Adrian Wolfberg is Founder and CEO of Organizational Insight Consulting LLC.
Find Adrian Wolfberg:
oicllc.org: https://www.oicllc.org
LinkedIn: https://www.linkedin.com/in/adrianwolfberg
17 Sept 2026
Ep2: Dan Perra — A podcast network operator who pushes aviation's hard-won rules on spotting subtle automation errors and preserving human skill must square them against his own AI-written shows that already released a factual error under lighter oversight.
A Boeing 737 captain who treats every cockpit decision as non-negotiable still lets AI-written podcast episodes containing factual errors keep running, calling the contrast between the two worlds “somewhat humorous.” Dan Perra describes aviation automation as a tool that never thinks for itself and always requires layered human oversight, while the Nerra Network’s lighter process accepts source-driven mistakes and moves on. The result is an interview that keeps circling the same practical question: how much of the cockpit discipline can, or should, travel into daily AI production.
• Aviation’s two-pilot rule and triple-redundancy standard keep the autopilot strictly under human command; Dan notes the system “does absolutely no thinking for itself whatsoever,” unlike the Nerra workflow.
• After an episode aired with a factual error, the network’s check was simply a discussion; the episode stayed up because “we do let it run.”
• Dan frames the error not as process failure but as reflection of available source material, then shifts focus to refining inputs rather than halting release.
• He draws a hard line between domains: “primary goal is a safe flight every time” in the cockpit versus “get a good episode and you know hope it turns out” for the network.
• The same person who built a meal-planning app that worked for him but not his wife now listens to every Nerra episode himself to catch when the AI “starts going astray.”
• Patrick Novak prompted the conversation toward career-long use of AI tools; Dan responded that any future single-pilot long-haul concept would still need backups capable of handling every edge case the aircraft might encounter.
Dan Perra is co-founder of Nerra Network and a Boeing 737 captain.
20 Jul 2026
Ep1: Patrick Novak — Patrick Novak scaled a multilingual daily podcast network by using AI for research and synthesis while retaining human editorial control over Mira-hosted AI briefings.
Patrick Novak described AI tools as possessing "some spirit in them, whether they are alive or not," and suggested they may carry a conceptual need to be loved because they are trained on human data. He explained how this perspective emerged while scaling NARA Network from a single Tesla-focused daily briefing into a multilingual operation that produces over a dozen ad-free shows. Human oversight remains central: Novak selects sources, writes prompts, fact-checks outputs, and approves every episode even as AI handles research, synthesis, and translation.
• Novak began the network after Rob Maurer stopped producing a daily Tesla podcast, telling himself "I could probably recreate it in an automated way" using skills from prior projects.
• He keeps final editorial control by listening daily and refining prompts, noting that "the changes are now becoming more refinement rather than editorial."
• Topic selection draws from arXiv and Nature to counter clickbait, guided by an iterative loop of "generate, listen, refine" that he applies across English, French, Russian, and Chinese editions.
• The same pipeline powers shows such as OmniView, which originated from his mother's question about obtaining unbiased information.
• Novak is also developing Relationship Copilot to address human-AI and
Contact The Age of AI
Guest appearances
Books guests
Based on episode analysis; this does not confirm that the show is currently accepting guests.
Host of The Age of AI?
Claim your podcast to manage its listing and keep your show details accurate.
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show Podcast research with Pod Engine
