#48 Adam Robinson: Winning at the Great Game (Part 2)

Episode Summary

Executive Summary: The episode centers on Adam Robinson’s warning that AI, quantum computing, and data aggregation will radically reshape work, privacy, and power. He argues schools teach compliance rather than transferable learning, that experts must simplify to a few governing variables, and that real improvement comes from rehearsal and feedback. Chess, Fisher’s genius, and grading bias are used as illustrations of how mastery, subjectivity, and human judgment actually work.

Main Topics: AI, automation, and the future of work (Priority: 5/5): Robinson argues that algorithms and robots will increasingly outperform humans in nearly every procedural task, threatening jobs, identity, and social stability. He sees universal basic income as a likely response but stresses the psychological devastation of mass unemployment. Quantum computing and the collapse of privacy (Priority: 5/5): He describes quantum computing as a major accelerant for AI and a threat to encryption, arguing it could make secrets, security, and data protection far harder to sustain. Data ownership, surveillance, and platform power (Priority: 5/5): The discussion explains how Google, Facebook, Amazon, and similar platforms gather behavioral data to build anticompetitive advantages. Robinson cites examples like Tinder and Target to show how companies infer sensitive truths from seemingly mundane behavior. Learning, rehearsal, and how education should work (Priority: 5/5): Robinson argues schools emphasize rote following and note-taking instead of rehearsing the actual skills tested in real life. He says effective learning requires breaking tasks into sub-skills and practicing them under realistic conditions. Chess as a model for mastery and self-improvement (Priority: 4/5): His chess background and Bobby Fischer stories illustrate focused practice, pattern learning, and the value of playing to win. He uses chess to show how experts simplify complexity and learn through repeated review of actual performance. Decision-making under uncertainty and information overload (Priority: 4/5): Robinson cites Paul Slovic’s horse-handicapper study to argue that more information does not necessarily improve accuracy; it often increases confidence without improving judgment. He favors reducing complex domains to a few key variables. Feedback, self-awareness, and teaching on multiple levels (Priority: 4/5): He emphasizes that people are often blind to how they come across and need honest, structured feedback. He also argues that every interaction teaches multiple lessons, intentionally or not.

Key Arguments: AI is not yet full intelligence, but machine learning is progressing fast enough that systems will soon outperform humans in speed, reliability, and cost across many tasks. Mass unemployment caused by automation would be psychologically damaging because people derive identity from economic contribution, not just from free time. Quantum computing could make encryption obsolete, creating a world where secrecy and data protection become extremely difficult. Big tech’s data accumulation creates a compounding, anti-competitive advantage that smaller competitors cannot realistically match. Good learning means rehearsing the exact cognitive and behavioral skills required in real situations, not rereading notes or passively reviewing content. Experts succeed by narrowing attention to a few essential variables rather than trying to process everything. Confidence can rise even when accuracy does not; more information can reinforce bias rather than improve judgment. Honest feedback is essential for growth, but social dynamics often suppress it, so people need intentional systems to obtain it. Fischer’s success came from simplicity, focus, repetition, and a willingness to play to win rather than merely avoid loss.

Data Points: Chess rating, Magnus Carlsen: 2820 - Robinson cites Carlsen as the best human chess player in the world. Chess rating, top AI trained on human games: about 3,300 - He says software trained on human games already far surpasses the best human players. Chess rating, AlphaZero after self-play: about 3,600–3,700 - DeepMind’s self-play system reportedly leapt beyond the trained version in about four hours. AlphaZero vs. prior software match result: 72–28 - Robinson says the self-play engine beat the reigning program in a 100-game match with no losses. Stephen Hawking estimate of AI survival odds: 1 in 20 - Used to emphasize the seriousness of existential AI risk. Probability of not surviving AI (implied): 19 in 20 - Derived from Hawking’s cited estimate. Horse handicappers, round 1 accuracy: 17% - Paul Slovic study with five pieces of information. Horse handicappers, round 1 confidence: 19% - Their confidence roughly matched their accuracy initially. Horse handicappers, round 4 confidence: 31% - With 40 pieces of information, confidence rose while accuracy stayed flat. Horse handicappers, round 4 accuracy: 17% - Accuracy did not improve despite much more information. Teacher bias effect from female name: +0.25 grade - Same essay was graded more favorably when a female name appeared at the top. Teacher bias effect from male teacher grading female student: +0.5 grade - Male teachers graded female students more leniently. Indentation effect: ~0.73 grade difference - A male student graded by a male teacher lost nearly three-quarters of a grade if paragraphs were not indented. Fischer’s rise in world ranking: Age 12 good, age 13 elite, age 14 U.S. champion, age 15 top 8 in world - Illustrates rapid self-driven improvement. Fischer prep with the narrator: 2 weeks - Robinson spent two weeks with Fischer at Grossinger’s while Fischer prepared for the Spassky match. Fischer’s sleep schedule: 11–12 a.m. wake-up; 3–4 a.m. sleep - Used to illustrate finding optimal performance times. Fischer opening repertoire: Very limited; repeated the same openings for 16–17 years - Robinson uses this to argue for depth over breadth. Robinson’s chess study archive: ~700 Fischer games - He hand-copied and studied hundreds of Fischer games as a teenager.

Pivotal Quotes: "Sometime in the next few decades, an algorithm or a robot powered by algorithms is going to do whatever you do better, faster, more reliably in every respect, cheaper." — Adam Robinson: Explaining why automation and AI threaten many kinds of work. "The key to learning any skill, really, if there's anything I said today, like that was super important. The key to learning any skill, rehearse it." — Adam Robinson: Summarizing his view that learning requires practice of the exact required sub-skills. "The court jester was the one person who was empowered by the king always to tell the truth." — Adam Robinson: Discussing the need for honest feedback systems and why leaders need truth-tellers.

Implications: Listeners should expect AI, data power, and automation to reshape careers and institutions; survival will depend on reducing complexity, practicing real-world skills, and building honest feedback loops before systems become too opaque to control.