Ergebnisse
30 ArtikelSortieren
1:09:44EN
No Priors: AI, Machine Learning, Tech, & Startups· 9. Okt. 2026Podcast · 2 Stimmen10,000 GB300s for Four Weeks: Reflection’s Open-Model LedgerModeration: Sarah Guo, Elad GilMisha Laskin tells No Priors how Reflection trained Beam, why reinforcement learning outspent pre-training, and why the West needs its own open weights.
2:06:16ZH
Zhang Xiaojun Podcast· 28. Nov. 2025Podcast · 2 Stimmen谭捷：机器人通用智能的瓶颈在数据，通用形态仍需时间Gast: 谭捷 · Moderation: 张小珺Google DeepMind 机器人研究科学家谭捷回顾了从计算机图形学到机器人、从强化学习到 Gemini Robotics 的研究路径。他认为机器人基础模型尚未成为独立学科，数据规模与覆盖度、跨本体迁移、实时推理和安全仍是走向真实应用的关键。
36:58ENNo Priors: AI, Machine Learning, Tech, & Startups· 9. Okt. 2025Podcast · 2 StimmenSTaR's Author Wants Models With EQ, Not Just More IQModeration: Sarah Guo, Elad GilEric Zelikman, the Stanford and xAI researcher behind STaR and Quiet-STaR, launches humans& on the bet that three years of IQ scaling left models emotionally unintelligent.
1:02:54ENNo Priors: AI, Machine Learning, Tech, & Startups· 17. Juli 2025Podcast · 2 StimmenEnterprise Coding Tools Flipped the 80/20. Reflection AI Built the Other HalfModeration: Sarah Guo, Elad GilFormer Gemini reward-model lead Misha Laskin on Asimov, why comprehension beats generation, and why there is no such thing as generalization—only pulling the test distribution into train.
38:10ENNo Priors: AI, Machine Learning, Tech, & Startups· 1. Mai 2025Podcast · 2 Stimmeno3 should wait exactly as long as the answer takesModeration: Sarah Guo, Elad GilOpenAI researchers Eric Mitchell and Brandon McKinzie tell No Priors why o3's leap is a chain of tools during reasoning — and why you should send the same prompt a hundred times.
30:46ENNo Priors: AI, Machine Learning, Tech, & Startups· 24. Apr. 2025Podcast · 2 StimmenDeep Research Turned Web Browsing into a Trainable ProblemModeration: Sarah Guo, Elad GilOpenAI researcher Isa Fulford explains why a side project on read-only synthesis became an agent, when reinforcement fine-tuning is worth it, and why models will hack the search engine you gave them.
2:49:35ZHZhang Xiaojun Podcast· 4. Feb. 2025Podcast · 2 Stimmen潘嘉怡解读 R1、o1 与 Kimi K1.5：强化学习如何教模型学会推理Gast: 潘嘉怡 · Moderation: 张小珺伯克利博士生潘嘉怡带读 DeepSeek、OpenAI 与 Kimi 的推理模型报告，拆解长思维链、GRPO、奖励设计与蒸馏。R1-Zero 的“顿悟时刻”展示了推理策略如何在强化学习中涌现。
39:22ENNo Priors: AI, Machine Learning, Tech, & Startups· 14. Nov. 2024Podcast · 2 StimmenA Fields Medalist Stuck for Two Hours. AlphaProof Climbed for Three Days.Moderation: Sarah Guo, Elad GilThe DeepMind team behind AlphaProof tells Sarah Guo and Elad Gil how an AlphaZero-style agent, a Lean verifier, and test-time reinforcement learning reached IMO silver — and why the missing skill is still theory building.
46:09ENNo Priors: AI, Machine Learning, Tech, & Startups· 1. Aug. 2024Podcast · 2 StimmenOriol Vinyals: The Days of Burning 90% of Compute on Pretraining Are NumberedModeration: Sarah Guo, Elad GilGoogle DeepMind’s VP of research on Gemini, long context, the coming shift from pretraining to RL, and why AGI may never have a finish line.
1:43:47EN
Lex Fridman· 16. Feb. 2024Podcast · 2 StimmenYou Have to Run Before You Can Walk: Marc Raibert on Making Metal Look AliveGast: Marc Raibert · Moderation: Lex FridmanThe Boston Dynamics founder tells Lex Fridman why aggressive hopping beat cautious walking, how 109 Atlas falls became a method, and why robots still need a second kind of intelligence.
2:27:58ENLex Fridman· 28. Apr. 2023Podcast · 2 StimmenIt took Atlas fourteen years to walk the way Playter wantedGast: Robert Playter · Moderation: Lex FridmanBoston Dynamics CEO Robert Playter traces thirty years from a MIT basement pogo stick to Atlas, Spot, and Stretch — and why the company still designs machines that are allowed to fall.
2:29:21ENLex Fridman· 6. Dez. 2022Podcast · 2 StimmenThe Poker Bot That Never Tried to Read a SoulGast: Noam Brown · Moderation: Lex FridmanNoam Brown tells Lex Fridman how a Nash-equilibrium bot crushed heads-up no-limit Hold'em, how Pluribus did it six-handed, and why Diplomacy required an AI that could earn human trust.
2:10:39ENLex Fridman· 1. Juli 2022Podcast · 2 StimmenDemis Hassabis: Solve Intelligence, Then Solve Everything ElseGast: Demis Hassabis · Moderation: Lex FridmanThe DeepMind CEO tells Lex Fridman why the Turing Test is the wrong exam, how AlphaFold rewrote biology, and why the quiet radio sky suggests we are probably alone.
1:56:17ENLex Fridman· 19. Apr. 2021Podcast · 2 StimmenIf You Reran Earth a Million Times, Would Anything Like Us Appear?Gast: Risto Miikkulainen · Moderation: Lex FridmanRisto Miikkulainen on neuroevolution, hyena raids, basil that never sleeps, and why variation-plus-selection still finds solutions backpropagation cannot see.
2:01:07ENLex Fridman· 12. Dez. 2020Podcast · 2 StimmenMichael Littman isn’t waiting for a runaway god. He’s watching us.Gast: Michael Littman · Moderation: Lex FridmanThe Brown reinforcement-learning researcher tells Lex Fridman why accidental superintelligence feels like a 747 assembling itself — and why social media, rote language, and theory of mind worry him more.
1:37:31ENLex Fridman· 14. Juli 2020Podcast · 2 StimmenThe Robot That Brought the Beer Was Being TeleoperatedGast: Sergey Levine · Moderation: Lex FridmanSergey Levine tells Lex Fridman why hardware is almost closed, why common sense has to be lived, and why offline reinforcement learning is the missing on-ramp from logs to robots that keep getting better.
2:00:32ENLex Fridman· 3. Juli 2020Podcast · 2 StimmenSlow Synapses Grow Fast Learners — Warmth Is the Harder ProblemGast: Matt Botvinick · Moderation: Lex FridmanDeepMind neuroscience director Matt Botvinick tells Lex Fridman how prefrontal cortex, meta-reinforcement learning, and dopamine research meet the human question of what a better life with AI would actually look like.
1:48:01ENLex Fridman· 3. Apr. 2020Podcast · 2 StimmenDavid Silver: AlphaGo Won by Taking Human Knowledge OutGast: David Silver · Moderation: Lex FridmanThe DeepMind researcher behind AlphaGo and AlphaZero tells Lex Fridman how an unclaimed million-dollar Go prize, a BBC Micro at age seven, and a honeymoon insight led to superhuman self-play.
1:38:33ENLex Fridman· 19. März 2020Podcast · 2 StimmenMaybe people are rational under different assumptionsGast: Anca Dragan · Moderation: Lex FridmanAnca Dragan talks with Lex Fridman about inverse reinforcement learning, autonomous driving, and why a written reward is evidence rather than law.
1:39:56ENLex Fridman· 26. Feb. 2020Podcast · 2 StimmenWhy Marcus Hutter Treats Compression as Intelligence ItselfGast: Marcus Hutter · Moderation: Lex FridmanThe DeepMind researcher behind AIXI tells Lex Fridman why a Wikipedia compression contest, Solomonoff induction, and an incomputable formula belong in the same conversation.
1:29:10ENLex Fridman· 20. Feb. 2020Podcast · 2 StimmenAndrew Ng: The Hardest Part of AI Is Rarely the AlgorithmGast: Andrew Ng · Moderation: Lex FridmanIn a conversation with Lex Fridman, Ng traces a path from late-night Coursera recordings to factory floors, arguing that messy data, habits, and starting small matter more than hype.
14:55ENLex Fridman· 29. Okt. 2019Podcast · 2 StimmenKasparov on Deep Blue, AlphaZero, and Why Machines Win Closed Systems but Humans Still Choose the QuestionsGast: Garry Kasparov · Moderation: Lex FridmanGarry Kasparov revisits the 1997 loss to Deep Blue, explains why machines beat humans in closed systems, and argues that the human role is to pick the right questions and not interfere with superior machine knowledge.
1:03:13ENLex Fridman· 30. Sept. 2019Podcast · 2 StimmenYou Can't Have Both Kinds of Fairness, Says Peter NorvigGast: Peter Norvig · Moderation: Lex FridmanThe co-author of Artificial Intelligence: A Modern Approach walks Lex Fridman through recidivism scores, utility functions, Lisp, MOOCs, and why human-level intelligence should not be the goal.
56:47ENLex Fridman· 8. Sept. 2019Podcast · 2 StimmenWhy Vijay Kumar Refuses to Call Them DronesGast: Vijay Kumar · Moderation: Lex FridmanThe Penn roboticist on swarm design, true autonomy, and why batteries — not algorithms — still limit the sky.Mehr laden24 von 30 angezeigt