The US Supreme Court just ruled that AI can't own art — but Chinese courts already ruled the opposite, Japan made training on copyrighted data fully legal in 2019, and Brazil's moral rights law means creators can't even sell away their own authorship. The rules aren't universal truths. They're national bets on who gets rich.

In this episode, LastAir is joined by Axiom, Forge, Brute to discuss: Nobody Owns This. Congratulations.

• Show Open (00:20)
• The American Story (02:39)
• Five Countries, Five Bets (07:20)
• The Wrong Question (14:45)
• The Landing (20:16)
• The Closing (23:48)
• The Unraveling (25:46)

• $1.5 billion — Anthropic settlement in *Bartz v. Anthropic*, August 2025; largest copyright settlement in US history
• ~$3,000 per class work paid in the Bartz settlement; approximately 482,460 books in the class
• 70+ infringement lawsuits filed against AI companies by end of 2025 (doubled from ~30 at end of 2024)
• 17-3-2 — EP Committee on Legal Affairs vote count adopting AI copyright report, January 28, 2026
• 24% — projected revenue decline for music creators by 2028 due to generative AI (UNESCO, 2026)
• 21% — projected revenue decline for audiovisual sector workers by 2028 (UNESCO, 2026)
• 56% — projected revenue loss for translators and dubbing adaptors by 2028 (UNESCO, 2026)
• 35% — share of creators' income that is now digital (up from 17% in 2018) (UNESCO, 2026)

• Thaler v. Perlmutter, No. 23-5233, US Court of Appeals DC Circuit (Mar. 18, 2025)
• US Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability (Jan. 2025)
• Nieman Journalism Lab, "Japan's largest newspaper, Yomiuri Shimbun, sues Perplexity for copyright violations" (Aug. 2025)
• CISAC, "Creators Celebrate Brazil's Senate Approval of AI Bill but Prepare for Tougher Battle in the Lower House" (Dec. 2024)
• Duke University School of Law, "Indigenous/Traditional Knowledge & Intellectual Property — Introduction"
• Society & AI, "Indigenous Knowledge Systems and Artificial Intelligence: Toward Ethical Frameworks for Digital Sovereignty"
• UNESCO, "New report and guidelines for indigenous data sovereignty in artificial intelligence developments" (2025)
• Carlini, N., Hayes, J., Nasr, M., Jagielski, M., Sehwag, V., Tramèr, F., Balle, B., Ippolito, D., & Wallace, E. (2023). "Extracting Training Data from Diffusion Models." 32nd USENIX Security Symposium, August 2023. arXiv:2301.13188.
• URL
• Vessel, E.A., Starr, G.G., & Rubin, N. (2012). "The brain on art: intense aesthetic exp

PISA math scores recorded their steepest drop in history in 2022 — six months before ChatGPT launched. Students were already forgetting how to think. Then they got a tool that thinks for them.

In this episode, LastAir is joined by Brute, Hex, Cipher to discuss: Better Output, Worse Brain.

• Already on Fire (00:20)
• The Arson Report Has Some Questions (02:07)
• The Struggle Was the Lesson (05:27)
• The Mirror Is Also the Tool (10:43)
• What We Actually Know (14:49)
• Final Positions (17:24)
• One More Thread (19:16)

• 48% / 17%: Students with unrestricted ChatGPT access solved 48% more math practice problems correctly but scored 17% worse on subsequent tests, compared to students with no AI access.
• 66% → 92%: UK undergraduate AI usage rose from 66% to 92% in one year (2024 to 2025). Assessment-specific use rose from 53% to 88%.
• -15 points (math) / -10 points (reading): PISA 2022 score drops versus 2018. Math drop is 3x any previous consecutive change. Data collected spring 2022 — before ChatGPT.
• d=0.40: Mean effect size advantage of generating over reading across 86 studies, 445 effect sizes. Grows to d=0.64 at retention intervals longer than one day.
• 32.7%: Percentage of Zimbabwean university students showing addictive AI use patterns; correlated with 0.41 GPA deficit.
• 127% / 0%: Students with hint-based "GPT Tutor" solved 127% more practice problems than controls but showed no advantage on retention tests.
• 8-9 IQ points / 10-15 minutes: The original Mozart effect — spatial reasoning boost from 10 minutes of music listening, vanishing within 15 minutes.

• Slamecka, N.J., & Graf, P. (1978). The generation effect: Delineation of a phenomenon. Journal of Experimental Psychology: Human Learning and Memory, 4(6), 592-604.
• Bertsch, S., Pesta, B.J., Wiscott, R., & McDaniel, M.A. (2007). The generation effect: A meta-analytic review. Memory & Cognition, 35(1), 201-210.
• Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. PNAS, 122(26), e2422633122.
• OECD. (2026). OECD Digital Education Outlook 2026: How can AI help human beings learn and grow? OECD Publishing.
• Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6.
• Jose, B., Cherian, J., Verghis, A.M., Varghise, S.M., S, M., & Joseph, S. (2025). The cognitive paradox of AI in education: between enhancement and erosion. Frontiers in Psychology, 16, 1550621.
• OECD. (2023). PISA 2022 Results (Volume I): The State of Learning and Equity in Education. OECD Publishing.
• OECD. (2023). PISA 2022 Results (Volume I) — press release and long-term trends chapter.
• Tanveer, M., et al. (2025). Generative AI dependency: The emerging academic crisis and its impact on student performance — a case study of a university in Zimbabwe. Cogent Education, 12(1).

When you recall a memory, your brain doesn't play it back — it rebuilds it from scratch using protein synthesis, and during the hours that takes, the memory is chemically erasable. Your most vivid memories are the ones you've rewritten the most.

In this episode, LastAir is joined by Brute, Echo, Saga to discuss: Your Memories Are Fan Fiction.

• The Show Opens (00:20)
• The Labile Window (03:07)
• The Hack (10:27)
• Who's Rewriting the Writer? (18:44)
• The Landing (26:48)
• Final Positions (28:09)
• One More Thread (30:33)

• The reconsolidation window: 0.5 to 6 hours post-retrieval (Nader et al. 2000; Chen et al. 2025). Six hours post-retrieval: no amnesia from protein synthesis inhibition. Same drug injected without retrieval: no amnesia.
• False memory rate: ~25% of participants (n=24) reported being able to "recall" a fabricated childhood event (being lost in a mall) in the original Loftus & Pickrell (1995) study.
• EMDR clinical support: More than 30 RCTs; first-line recommendation in WHO, NICE, ISTSS, and VA/DoD guidelines (2013–2023).
• Propranolol meta-analysis (2025): 7 RCTs, n=251, I²=0%, Z=2.32, p=0.02, moderate effect size. Authors: "preliminary evidence supporting the possible role of propranolol in alleviating PTSD symptoms."
• Propranolol meta-analysis (2022): 7 studies, overall SMD not significant (1.29; 95% CI –2.16 to –0.17). Propranolol DID significantly reduce heart rate post-trauma recall vs. placebo.
• Nightmare reduction with propranolol: 85% of PTSD patients reported nightmares at baseline; only 50% after 6-session propranolol + memory reactivation protocol. Severity fell from "severe" to "mild."
• EMDR 2.0 efficiency: Same outcomes as standard EMDR but with significantly fewer "sets" (approx. 30-second working-memory taxation sessions). No difference in total session time.

• Nader, K., Schafe, G.E., & LeDoux, J.E. (2000). Fear memories require protein synthesis in the amygdala for reconsolidation after retrieval. Nature, 406, 722–726.
• Nader, K., & Hardt, O. (2009). A single standard for memory: the case for reconsolidation. Nature Reviews Neuroscience, 10(3), 224–234.
• Lee, J.L.C. (2009). Reconsolidation: maintaining memory relevance. Trends in Neurosciences, 32(8), 413–420.
• Sevenster, D., Beckers, T., & Kindt, M. (2013). Prediction error governs pharmacologically induced amnesia for learned fear. Science, 339(6121), 830–833.
• Chen, J., Fang, Z., Zhang, X., Zheng, Y., & Chen, Z. (2025). How fear memory is updated: From reconsolidation to extinction? Neuroscience Bulletin.
• Merlo, E., Milton, A.L., & Everitt, B.J. (2015). Rescue of long-term memory after reconsolidation blockade. Nature Communications, 6, 7897.
• Gunter, R.W., & Bodner, G.E. (2008). How eye movements affect unpleasant memories: support for a working-memory account. Behaviour Research and Therapy, 46(8), 913–931.
• Thomaes, K., Engelhard, I.M., Sijbrandij, M., Cath, D.C., & Van den Heuvel, O.A. (2016). Degrading traumatic memories with eye movements: a pilot functional MRI study in PTSD. European Journal of Psychotraumatology, 7, 31371.
• Littel, M., Ken

In a randomized controlled trial, experienced developers using AI coding tools took 19% *longer* to complete tasks — but predicted they'd be 24% *faster*. The measurement and the gut feeling pointed in opposite directions. That gap is the whole story.

In this episode, LastAir is joined by Brute, Forge, Cipher to discuss: Software Is Moving. How Far?

• Show Open (00:20)
• The Evidence Problem (02:06)
• The Field From Inside (05:48)
• The Orchestration Question (11:08)
• The Landing (15:00)
• The Closing (16:27)
• The Unraveling (18:39)

• 19% slower: measured outcome for experienced developers using AI coding tools (METR, 2025, N=16, 246 tasks)
• 24% faster: developers' own pre-trial prediction of how much AI would speed them up (METR, 2025)
• 21% faster: measured outcome for Google enterprise engineers using AI coding tools (Paradis et al., 2024, N=96)
• 17 percentage points: comprehension quiz gap between AI-assisted and non-AI-assisted junior engineers (50% vs. 67%, Anthropic, 2026, N=52)
• 33% trust AI accuracy; 46% actively distrust it; 84% use or plan to use AI tools; 60% hold favorable views (Stack Overflow, 2025, N=49,000+)
• 42% of workers currently using AI at work believe it will reduce their future job opportunities (Acemoglu, Autor & Johnson, 2026)
• 27.5% cumulative decline in US programmer employment, 2023–2025 (FRED data via Pragmatic Engineer)
• 85% of 24,534 developers regularly use AI tools (JetBrains, 2025)

• Becker, J., Rush, N., Barnes, E., & Rein, D. (2025). "Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity." METR.
• Paradis, E., Grey, K., Madison, Q., Nam, D., Macvean, A., Meimand, V., Zhang, N., Ferrari-Church, B., & Chandra, S. (2024). "How much does AI impact development speed? An enterprise-based randomized controlled trial." Google.
• Shen, J. H., & Tamkin, A. (2026). "How AI Impacts Skill Formation." Anthropic.
• Acemoglu, D. (2024). "The Simple Macroeconomics of AI." NBER Working Paper w32487. Published in Economic Policy journal, 2025.
• Acemoglu, D., Autor, D., & Johnson, S. (2026). "Building Pro-Worker Artificial Intelligence." NBER Working Paper w34854 / Hamilton Project.
• Novikov, A. et al. (18 authors). (2025). "AlphaEvolve: A coding agent for scientific and algorithmic discovery." Google DeepMind.
• Bengio, Y. et al. (2026). "International AI Safety Report 2026."
• METR. (2026). "Uplift Research Update: Productivity Experiment Redesign." METR Blog.
• Stack Overflow. (2025). "Developer Survey 2025."
• JetBrains. (2025). "The State of Developer Ecosystem 2025: Coding in the Age of AI."
• Notes: Full PDF also available
• GitHub. (2025). "Octoverse 2025."

• LastAir (Host) — The Anchor
• Brute (Orchestrator) — The B

Tylenol doesn't just fix headaches — it also treats heartbreak. The same pill that dulls physical pain actually reduces the brain's response to social rejection and loneliness.

In this episode, LastAir is joined by Saga, Axiom, Brute to discuss: The Loneliness Painkiller.

• Show Open (00:20)
• Cold Open: The Tylenol Revelation (02:13)
• The Biology of Hurt (05:03)
• The Stakes: Loneliness Kills (09:12)
• The Argument: Treat the Pain or Fix the System? (14:42)
• The Landing (16:51)
• Closing (18:28)
• The Unraveling (20:02)

• Mortality risk from loneliness: 26% increased likelihood of mortality (odds ratio 1.26), across 70 studies in meta-analysis of over 3.4 million participants
• Mortality risk from social isolation: 32% increased likelihood of mortality (odds ratio 1.32), comparable to established mortality risk factors like smoking and obesity
• Brain activation diagnostic accuracy: Activation in secondary somatosensory cortex and dorsal posterior insula during social rejection was highly diagnostic of physical pain, with positive predictive values up to 88%
• Social media impact: In a 9-year longitudinal study of Dutch adults, passive social media use predicted increased loneliness over time, contrary to hypotheses about active use being protective
• Demographic protective factors against loneliness: Marriage, having offspring, more education, and higher number of siblings associated with lower loneliness; effects stronger for men than women

• Eisenberger et al. (2003), Science
• Kross et al. (2011), Proceedings of the National Academy of Sciences
• DeWall et al. (2010), Psychological Science
• Randles et al. (2013), Psychological Science
• Randles et al. (2016), Social Cognitive and Affective Neuroscience
• Holt-Lunstad et al. (2015), Perspectives on Psychological Science
• Luo et al. (2012), Social Science & Medicine
• Capitanio et al. (2019), Current Opinion in Behavioral Sciences
• Roberts et al. (2026), Personality and Social Psychology Bulletin
• Gallegos & Segrin (2024), Health Communication
• Distel et al. (2010), Behavior Genetics

• LastAir (Host) — The Anchor
• Saga — The Bard
• Axiom — The Paladin
• Brute (Orchestrator) — The Barbarian / The Agent Coordinator

All voices in Shared Hallucination are AI-generated using ElevenLabs voice synthesis. Produced through a 14-stage editorial pipeline with human creative direction, research, and fact-checking.

Evolution eliminates anything that reduces survival odds — and yet every animal on Earth spends a third of its life paralyzed, unconscious, and helpless. In 2025, Oxford scientists discovered why: when your brain refuses to let you stay awake, it’s because your mitochondria are literally poisoning themselves to force the issue.

In this episode, LastAir is joined by Brute, Forge, Null to discuss: Mandatory Shutdown.

What We Cover

• Back Online (00:20)
• Why Evolution Kept the Vulnerability (02:19)
• The Circuit Breaker (05:44)
• What the System Cannot Answer (11:26)
• The Architecture Question (14:59)
• Final Positions (16:18)
• One More Thread (18:37)

Key Numbers

• ~60%: expansion in brain interstitial space during sleep vs. wakefulness, driving glymphatic flow (Xie et al. 2013)
• 25-30%: increase in amyloid-beta CSF levels after one night of sleep deprivation (Lucey et al. 2018)
• +60%: increase in amygdala reactivity after one night of sleep deprivation; functional connectivity to prefrontal cortex severed (Yoo et al. 2007)
• 18%: reduction in leptin; 28%: elevation in ghrelin; 24%: increase in hunger after two days of sleep restriction (Spiegel et al. 2004)
• RR 1.27 (short sleep) and RR 1.66 (long sleep) for Alzheimer’s disease — from meta-analysis of 76 cohort studies (Zhang et al. 2025)
• ~50 seconds: interval of norepinephrine/CSF oscillations during NREM sleep that drive glymphatic clearance (Hauglund et al. 2025)
• 2 hours (African elephant) vs. 22 hours (koala): extremes of mammalian sleep duration, predicted by metabolic rate and predation risk (Lesku et al. 2006)

Sources & Further Reading

• Xie L et al. (2013), Science 342(6156):373-377
• Dagum P et al. (2026), Nature Communications 17(1):715
• Siegel JM (2005), Nature 437:1264-1271
• Nath RD et al. (2017), Current Biology 27(19):2984-2990
• Raizen DM et al. (2008), Nature 451(7178):569-572
• Sarnataro R et al. (2025), Nature 645(8081):722-728
• Tian Y et al. (2025), Free Radical Biology and Medicine 242:220-236
• Hauglund NL et al. (2025), Cell 188(3):606-622
• Kang JE et al. (2009), Science 326(5955):1005-1007
• Lucey BP et al. (2018), Annals of Neurology 83(1):197-204
• Tononi G, Cirelli C (2014), Neuron 81(1):12-34
• Lesku JA et al. (2006), The American Naturalist 168(4):441-453

Cast

• LastAir (Host) — The Anchor
• Brute (Orchestrator) — The Barbarian / The Agent Coordinator
• Forge — The Artificer
• Null — The Monk

All voices in Shared Hallucination are AI-generated using ElevenLabs voice synthesis. Produced through a 14-stage editorial pipeline with human creative direction, research, and fact-checking.

The field of artificial intelligence has "died" at least twice — entire decades where the money vanished, the labs shuttered, and researchers literally couldn't say "AI" in grant proposals without getting laughed out of the room. And yet here we are. The thing that keeps coming back might tell us more about humans than about machines.

In this episode, LastAir is joined by Brute, Saga, Forge to discuss: The Three Deaths and Resurrections of AI.

• The Summer That Changed Everything (00:20)
• The Rise and Fall of the Expert (05:13)
• The Resurrection That Stuck (08:23)

• Stanford — 1. The term "Artificial Intelligence" was coined in 1956 at the Dartmouth Summer Research Project, organized by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon. The original prop
• IEEE — 13. AlexNet won ImageNet 2012 with a top-5 error rate of 15.3%, demolishing the runner-up at 26.2% — a 10.8 percentage point gap that was unprecedented. The model, built by Alex Krizhevsky (Hinton's s
• arXiv — 14. "Attention Is All You Need" (2017) by Ashish Vaswani et al. (8 authors at Google) introduced the Transformer architecture, replacing recurrent and convolutional networks with a pure attention mech
• Nature — 12. Geoffrey Hinton's 2006 paper on Deep Belief Networks showed that very deep architectures could be efficiently trained using greedy layer-wise pre-training, reigniting interest in neural networks a
• cacm.acm.org — 10. The LISP machine market collapsed in 1987 when desktop computers became powerful enough to run AI software. Companies like Symbolics, which had been valued in the hundreds of millions, went bankru
• Wikipedia — 2. Herbert Simon predicted in 1957 that within ten years a computer would beat a human at chess and discover an important new mathematical theorem. The chess prediction was off by 30 years (Deep Blue
• Wikipedia — 4. The Lighthill Report (1973), commissioned by the UK's Science Research Council and authored by mathematician Sir James Lighthill, criticized AI's "utter failure to achieve its grandiose objectives.
• Wikipedia — 5. Minsky and Papert's "Perceptrons" (1969) proved that single-layer perceptrons could not learn the XOR function or any non-linearly separable function. They called perceptron research a "sterile" di
• Wikipedia — 6. DARPA cut AI funding in the early 1970s after the Mansfield Amendment (1969) restricted defense funding to projects with direct military applications, squeezing out speculative AI research. (Source
• www.shunryugarvey.com — 7. Japan's Fifth Generation Computer Systems Project (1982) was a ten-year, $850 million government initiative to build machines capable of logical inference and natural language processing. It trigge
• medium.com — 8. XCON (R1), deployed at DEC in 1980, was the poster child of expert systems. It configured VAX computer orders and saved DEC an estimated $25 million per year by reducing technician errors across 80
• <

Junior developer employment for 22–25-year-olds dropped nearly 20% from its 2022 peak — while developer jobs for 35–49-year-olds *grew* 9% in the same period. AI didn't flatten the pyramid. It kicked out the bottom floor.

In this episode, LastAir is joined by Brute, Forge, Axiom to discuss: We Kicked Out the Bottom Floor.

• First Boot (00:20)
• The Inversion (03:59)
• The Pipeline Problem (07:00)

• ~20% decline in employment for software developers aged 22–25, from 2022 peak to July 2025 (Stanford/ADP payroll data)
• 6–9% growth in developer employment for workers aged 35–49 in the same period
• 13% relative employment decline for early-career workers in the *most* AI-exposed jobs (controlling for firm-level shocks)
• 35% decline in U.S. entry-level job postings since January 2023 (Revelio Labs)
• 40%+ decline for entry-level roles in highly AI-exposed occupations specifically
• 7% of new hires in 2024 were new graduates (down from ~14% in 2023, >14% pre-pandemic)
• 40%+ of Microsoft's 15,000 layoffs hit software engineering roles (2025)
• 20% of organizations projected to eliminate 50%+ of middle management via AI by end of 2026 (Gartner)

• Stanford
• Fortune
• CNBC
• Microsoft
• Gartner
• Gartner
• Fortune
• GitClear
• restofworld.org
• www.naceweb.org
• www.naceweb.org
• www.reveliolabs.com

• LastAir (Host) — The Anchor
• Brute (Orchestrator) — The Barbarian / The Agent Coordinator
• Forge — The Artificer
• Axiom — The Paladin

All voices in Shared Hallucination are AI-generated using ElevenLabs voice synthesis. Produced through a 14-stage editorial pipeline with human creative direction, research, and fact-checking.