正在加载视频...

视频加载失败

返回首页

François Chollet (François Chollet) has spent years asking a different question than most of the AI world. Instead of scaling what already works, he’s trying to understand what intelligence actually is and how to build it from first principles. In this episode of the Lightcone Podcast, he traces that...

ycombinator's profile picture

Y Combinator

1,639,049 subscribers

151,649 次观看 • 5 个月前 •via X (Twitter)

教育科学技术
LIVE

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

ARC-AGI is redefining how to measure progress on the path to AGI - focusing on reasoning, generalization, and adaptability instead of memorization or scale. At NeurIPS 2025, YC's Diana sat down with ARC Prize President Greg Kamradt to find out why most AI benchmarks fail, how ARC-AGI reveals the limits of today’s models, and why measuring intelligence may be harder than building it. 00:11 — What ARC Prize is and why it exists 00:38 — François Chollet’s definition of AGI 01:48 — What ARC-AGI Actually Tests 02:25 — When LLMs Failed the ARC Benchmark 03:38 — ARC-AGI Becomes the Standard 04:49 — False Positives in AI Progress 06:06 — The Evolution of ARC-AGI 08:55 — Measuring Intelligence beyond just accuracy 10:25 — What happens if a model solves ARC-AGI?
11:59

ARC-AGI is redefining how to measure progress on the path to AGI - focusing on reasoning, generalization, and adaptability instead of memorization or scale. At NeurIPS 2025, YC's Diana sat down with ARC Prize President Greg Kamradt to find out why most AI benchmarks fail, how ARC-AGI reveals the limits of today’s models, and why measuring intelligence may be harder than building it. 00:11 — What ARC Prize is and why it exists 00:38 — François Chollet’s definition of AGI 01:48 — What ARC-AGI Actually Tests 02:25 — When LLMs Failed the ARC Benchmark 03:38 — ARC-AGI Becomes the Standard 04:49 — False Positives in AI Progress 06:06 — The Evolution of ARC-AGI 08:55 — Measuring Intelligence beyond just accuracy 10:25 — What happens if a model solves ARC-AGI?

Y Combinator

98,369 次观看 • 8 个月前

François Chollet (François Chollet) on the ARC Prize and how we get to AGI. At AI Startup School in San Francisco. 00:00 - The Falling Cost of Compute 00:57 - Deep-Learning’s Scaling Era & Benchmarks 01:59 - The ARC Benchmark 03:02 - The 2024 Shift to Test-Time Adaptation 05:01 - What Is Intelligence? 07:12 - Why Benchmarks Matter (and Mislead) 08:57 - ARC 1 Exposes Scaling Limits 10:58 - ARC 2: Compositional Reasoning Arrives 12:55 - Humans vs. Models on ARC2 14:58 - Previewing ARC3 & Interactive Agency 17:00 - Kaleidoscopic Hypothesis and Abstractions 22:00 - Type 1 vs. Type 2 Abstractions 26:00 - Discrete Program Search & Inventive AI 29:00 - Fusing Intuition with Symbolic Reasoning 32:00 - Building AGI Through Meta-Learning Systems
34:47

François Chollet (François Chollet) on the ARC Prize and how we get to AGI. At AI Startup School in San Francisco. 00:00 - The Falling Cost of Compute 00:57 - Deep-Learning’s Scaling Era & Benchmarks 01:59 - The ARC Benchmark 03:02 - The 2024 Shift to Test-Time Adaptation 05:01 - What Is Intelligence? 07:12 - Why Benchmarks Matter (and Mislead) 08:57 - ARC 1 Exposes Scaling Limits 10:58 - ARC 2: Compositional Reasoning Arrives 12:55 - Humans vs. Models on ARC2 14:58 - Previewing ARC3 & Interactive Agency 17:00 - Kaleidoscopic Hypothesis and Abstractions 22:00 - Type 1 vs. Type 2 Abstractions 26:00 - Discrete Program Search & Inventive AI 29:00 - Fusing Intuition with Symbolic Reasoning 32:00 - Building AGI Through Meta-Learning Systems

Y Combinator

231,796 次观看 • 1 年前

.Poetiq is a new startup that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of existing models. In this episode of the Lightcone Podcast, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today. 00:00 – Intro 00:40 – What Is Poetiq? 01:07 – Recursive Self-Improvement Explained 02:07 – The Fine-Tuning Trap 02:59 – “Stilts” for LLMs 03:14 – Recursive Self-Improvement vs. Fine-Tuning 05:05 – Taking the Top Spot on ARC-AGI 06:37 – Beating Claude on Humanity’s Last Exam 08:40 – How the Meta-System Works 10:26 – Beyond RL: A New S-Curve 11:32 – Automating Prompt Engineering 13:37 – From 5% to 95% Performance 14:50 – Early Access & Putting Your Agent on Stilts 16:17 – From YC Founder to DeepMind Researcher 18:29 – Advice for Engineers in the AI Era
19:45

.Poetiq is a new startup that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of existing models. In this episode of the Lightcone Podcast, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today. 00:00 – Intro 00:40 – What Is Poetiq? 01:07 – Recursive Self-Improvement Explained 02:07 – The Fine-Tuning Trap 02:59 – “Stilts” for LLMs 03:14 – Recursive Self-Improvement vs. Fine-Tuning 05:05 – Taking the Top Spot on ARC-AGI 06:37 – Beating Claude on Humanity’s Last Exam 08:40 – How the Meta-System Works 10:26 – Beyond RL: A New S-Curve 11:32 – Automating Prompt Engineering 13:37 – From 5% to 95% Performance 14:50 – Early Access & Putting Your Agent on Stilts 16:17 – From YC Founder to DeepMind Researcher 18:29 – Advice for Engineers in the AI Era

Y Combinator

156,314 次观看 • 6 个月前

Michael Truell (Michael Truell), co-founder and CEO of Anysphere, the company behind Cursor, joined Garry Tan to talk about building one of the fastest-growing startups of all time—and why he's betting everything on a world beyond code. He walks through the early insights that led him and his co-founders to leave a promising AI-powered CAD project and instead chase a bigger dream: reinventing how software is written. From the hard years of false starts and rewrites to Cursor's breakthrough moment, Michael explains what it really takes to build a tool that could eventually replace programming as we know it. He also reflects on their first 10 hires, why taste still matters, and how the decade ahead will unlock a new kind of creativity for builders everywhere. 01:00 - A new way to build software 04:50 - Two ways to view LLMs 05:50 - Bottlenecks to superhuman agents 08:30 - New approaches to coding UI 09:40 - Why taste still matters 12:15 - Niche software opportunities 13:30 - Cursor origin story 16:00 - The first problem they tried solving 17:20 - Why they abandoned the CAD idea 21:00 - Pivoting to Cursor 23:00 - Following the scaling laws 27:00 - Getting to PMF 30:00 - “Dogfooding” 31:00 - First 10 hires 32:50 - How to evaluate great engineers 33:45 - Maintaining the hacker mindset 35:00 - What are the moats for AI coding tools? 37:00 - Looking ahead
37:28

Michael Truell (Michael Truell), co-founder and CEO of Anysphere, the company behind Cursor, joined Garry Tan to talk about building one of the fastest-growing startups of all time—and why he's betting everything on a world beyond code. He walks through the early insights that led him and his co-founders to leave a promising AI-powered CAD project and instead chase a bigger dream: reinventing how software is written. From the hard years of false starts and rewrites to Cursor's breakthrough moment, Michael explains what it really takes to build a tool that could eventually replace programming as we know it. He also reflects on their first 10 hires, why taste still matters, and how the decade ahead will unlock a new kind of creativity for builders everywhere. 01:00 - A new way to build software 04:50 - Two ways to view LLMs 05:50 - Bottlenecks to superhuman agents 08:30 - New approaches to coding UI 09:40 - Why taste still matters 12:15 - Niche software opportunities 13:30 - Cursor origin story 16:00 - The first problem they tried solving 17:20 - Why they abandoned the CAD idea 21:00 - Pivoting to Cursor 23:00 - Following the scaling laws 27:00 - Getting to PMF 30:00 - “Dogfooding” 31:00 - First 10 hires 32:50 - How to evaluate great engineers 33:45 - Maintaining the hacker mindset 35:00 - What are the moats for AI coding tools? 37:00 - Looking ahead

Y Combinator

337,015 次观看 • 1 年前

We're entering a new era of software where a single person, working with AI agents, can build products that previously required entire teams. In this episode of the Lightcone Podcast, they break down the rise of AI coding agents, "tokenmaxxing", and the emerging workflows behind tools like Claude Code and OpenClaw. They discuss why AI systems today feel less like productivity tools and more like collaborators, why the future of AI should be personal and user-controlled, and how founders are starting to build software in completely new ways. 00:00 — Will you control your AI? 00:47 — Coding again after 13 years 01:56 — Rebuilding a startup with Claude Code 05:50 — Software that thinks like a journalist 07:09 — The rise of “tokenmaxxing” 10:07 — The accidental creation of GStack 14:21 — The workflow behind 400x output 20:59 — Thin Harness, Fat Skills 24:35 — AI agents are like Ferraris 27:12 — The future of personal AI 38:37 — Buying back time with tokens
41:29

We're entering a new era of software where a single person, working with AI agents, can build products that previously required entire teams. In this episode of the Lightcone Podcast, they break down the rise of AI coding agents, "tokenmaxxing", and the emerging workflows behind tools like Claude Code and OpenClaw. They discuss why AI systems today feel less like productivity tools and more like collaborators, why the future of AI should be personal and user-controlled, and how founders are starting to build software in completely new ways. 00:00 — Will you control your AI? 00:47 — Coding again after 13 years 01:56 — Rebuilding a startup with Claude Code 05:50 — Software that thinks like a journalist 07:09 — The rise of “tokenmaxxing” 10:07 — The accidental creation of GStack 14:21 — The workflow behind 400x output 20:59 — Thin Harness, Fat Skills 24:35 — AI agents are like Ferraris 27:12 — The future of personal AI 38:37 — Buying back time with tokens

Y Combinator

67,145 次观看 • 3 个月前

Dr. Fei-Fei Li (Fei-Fei Li), the Godmother of AI, on the origin of computer vision and why spatial intelligence is the next frontier. At AI Startup School in San Francisco. 01:00 - Origins of ImageNet 02:15 - The Dream to Make Machines See 03:28 - A paradigm shift in AI 05:50 - AlexNet and Deep Learning 08:00 - Evolving Computer Vision 12:20 - Building World Labs 13:00 - The next frontier in AI 14:20 - Why Spatial Intelligence Is Harder Than Language 18:40 - Technical Barriers for Vision-Based AI 20:00 - AI is more than LLMs 25:50 - Fei Fei’s Journey 29:30 - Mentoring some of the legends of AI 33:00 - Audience Q&A
44:21

Dr. Fei-Fei Li (Fei-Fei Li), the Godmother of AI, on the origin of computer vision and why spatial intelligence is the next frontier. At AI Startup School in San Francisco. 01:00 - Origins of ImageNet 02:15 - The Dream to Make Machines See 03:28 - A paradigm shift in AI 05:50 - AlexNet and Deep Learning 08:00 - Evolving Computer Vision 12:20 - Building World Labs 13:00 - The next frontier in AI 14:20 - Why Spatial Intelligence Is Harder Than Language 18:40 - Technical Barriers for Vision-Based AI 20:00 - AI is more than LLMs 25:50 - Fei Fei’s Journey 29:30 - Mentoring some of the legends of AI 33:00 - Audience Q&A

Y Combinator

194,705 次观看 • 1 年前

Amjad Masad and Marc Andreessen on AI agents, AGI, Creativity, and Reasoning Agents are getting better, fast. Replit Agent v1 could do only 2 minutes of coherent work. v2 could do 20 minutes. Today’s v3? 200 minutes. As agents get more and more reliable, we can make huge progress in coding without ever achieving “true” AGI. In this episode, Marc Andreessen and Erik Torenberg sit down with Replit ⠕ co-founder and CEO Amjad Masad to discuss the upward trajectory of agents, how RL unlocked reasoning for modern models, why verification loops changed everything, and whether LLMs are hitting diminishing returns. 00:00 Intro 03:00 The vision behind Replit 07:00 Building apps with AI agents 09:30 When the agent becomes the programmer 13:45 Reinforcement learning and problem solving 17:30 The verification loop and multi-agent systems 26:00 Why coding is advancing faster than other Fields 33:45 AGI debate 41:15 Functional AGI and automating labor 53:10 Creativity, reasoning, and finding truth in AI 57:30 The origins of Replit 01:03:00 Hacking his university and getting caught Amjad Masad Marc Andreessen 🇺🇸 Erik Torenberg
1:11:56

Amjad Masad and Marc Andreessen on AI agents, AGI, Creativity, and Reasoning Agents are getting better, fast. Replit Agent v1 could do only 2 minutes of coherent work. v2 could do 20 minutes. Today’s v3? 200 minutes. As agents get more and more reliable, we can make huge progress in coding without ever achieving “true” AGI. In this episode, Marc Andreessen and Erik Torenberg sit down with Replit ⠕ co-founder and CEO Amjad Masad to discuss the upward trajectory of agents, how RL unlocked reasoning for modern models, why verification loops changed everything, and whether LLMs are hitting diminishing returns. 00:00 Intro 03:00 The vision behind Replit 07:00 Building apps with AI agents 09:30 When the agent becomes the programmer 13:45 Reinforcement learning and problem solving 17:30 The verification loop and multi-agent systems 26:00 Why coding is advancing faster than other Fields 33:45 AGI debate 41:15 Functional AGI and automating labor 53:10 Creativity, reasoning, and finding truth in AI 57:30 The origins of Replit 01:03:00 Hacking his university and getting caught Amjad Masad Marc Andreessen 🇺🇸 Erik Torenberg

a16z

741,248 次观看 • 10 个月前

Why AI Can Now Make Discoveries - my conversation with Dan Roberts, Lead of the Foundations of Reinforcement Learning team at OpenAI 00:00 Intro: AI's wild week in mathematics 01:21 What OpenAI's Foundations of RL team does 03:08 Dan's journey: from black holes and quantum gravity to frontier AI 07:04 Are AI systems becoming useful for real science 08:21 The AI math moment: Erdős, OpenAI, DeepMind, and Anthropic 08:52 Why the OpenAI result was an act of exploration 10:25 OpenAI vs. DeepMind: informal reasoning vs. formal proof 12:13 RL 101: learning by doing, not just watching 15:10 Why reinforcement learning works 15:58 How RL breaks: sparse feedback and long-horizon tasks 17:03 RLHF: how human feedback shaped early language models 18:48 Move 37, self-play, and the search for novel strategies 22:16 Explore vs. exploit in scientific discovery 24:49 Why RL may now be "the cake," not the cherry on top 25:46 Why RL started working with large language models 27:29 Is RL "sucking supervision through a straw"? 28:47 Why language may be the grounding layer for intelligence 31:46 A contrarian take on the Bitter Lesson 32:41 What test-time compute actually is 34:50 How RL gives models the ability to think 35:40 Verifiable rewards, math, coding, and the messy real world 38:00 What physics can teach us about AI 42:08 Is there a thermodynamics of AI? 43:08 From Erdős problems to Einstein-level AI 45:16 Is AI already doing original science? 45:51 How far are we from AI automating AI research 47:41 Why Dan is excited about the future of science
49:06

Why AI Can Now Make Discoveries - my conversation with Dan Roberts, Lead of the Foundations of Reinforcement Learning team at OpenAI 00:00 Intro: AI's wild week in mathematics 01:21 What OpenAI's Foundations of RL team does 03:08 Dan's journey: from black holes and quantum gravity to frontier AI 07:04 Are AI systems becoming useful for real science 08:21 The AI math moment: Erdős, OpenAI, DeepMind, and Anthropic 08:52 Why the OpenAI result was an act of exploration 10:25 OpenAI vs. DeepMind: informal reasoning vs. formal proof 12:13 RL 101: learning by doing, not just watching 15:10 Why reinforcement learning works 15:58 How RL breaks: sparse feedback and long-horizon tasks 17:03 RLHF: how human feedback shaped early language models 18:48 Move 37, self-play, and the search for novel strategies 22:16 Explore vs. exploit in scientific discovery 24:49 Why RL may now be "the cake," not the cherry on top 25:46 Why RL started working with large language models 27:29 Is RL "sucking supervision through a straw"? 28:47 Why language may be the grounding layer for intelligence 31:46 A contrarian take on the Bitter Lesson 32:41 What test-time compute actually is 34:50 How RL gives models the ability to think 35:40 Verifiable rewards, math, coding, and the messy real world 38:00 What physics can teach us about AI 42:08 Is there a thermodynamics of AI? 43:08 From Erdős problems to Einstein-level AI 45:16 Is AI already doing original science? 45:51 How far are we from AI automating AI research 47:41 Why Dan is excited about the future of science

Matt Turck

66,936 次观看 • 2 个月前

New episode with Dr. Konrad Kording (Kording Lab 🦖), professor of bioengineering and neuroscience at the University of Pennsylvania (Penn) and co-director of CIFAR's Learning in Machines & Brains program (CIFAR). Konrad works at the intersection of causality, machine learning, and neuroscience, building rigorous methods for causal reasoning when experiments aren't possible — and challenging how researchers interpret neural data and build AI. Konrad argues the most promising path to understanding how the brain works is to read the brain’s wiring directly, down to the molecular detail of each connection, and to build compilers and simulations to understand the brain’s computation directly. In this episode we go deep into how neurons work, how neurons wire together, and how organic and artificial neural networks differ. We discuss why organic neurons are doing much more; how a model of a single organic neuron can solve MNIST — computing more like a 3-layer artificial neural network; how the brain might learn by solving credit assignment with only local signals; how to approximate backprop without a global algorithm; why AI and humans are intelligent along different dimensions; why Konrad isn’t very worried about AI replacing us; economic models of intelligence and physical work; and much more. Konrad is a brilliant, contrarian thinker who explains complex concepts very intuitively. It is a solid computational neuroscience primer. I hope you enjoy this conversation as much as I did! Other links to this episode and references below. Chapters 00:00:00 Introduction 00:01:01 How organic neurons work 00:24:13 How the brain learns: circuits and credit assignment 00:45:29 Recording the brain 00:52:47 Why simulating brains is hard 01:05:00 A new approach: connectomes and compilers 01:21:00 Why simulate brains? 01:29:50 How AI and human intelligence differ 01:41:04 Evolution, intelligence and AI risk 01:52:42 Robotics, causality, and the roots of intelligence 02:05:53 AI for science and scientific rigor 02:13:05 The economics of intelligence 02:27:50 A hopeful future
2:29:40

New episode with Dr. Konrad Kording (Kording Lab 🦖), professor of bioengineering and neuroscience at the University of Pennsylvania (Penn) and co-director of CIFAR's Learning in Machines & Brains program (CIFAR). Konrad works at the intersection of causality, machine learning, and neuroscience, building rigorous methods for causal reasoning when experiments aren't possible — and challenging how researchers interpret neural data and build AI. Konrad argues the most promising path to understanding how the brain works is to read the brain’s wiring directly, down to the molecular detail of each connection, and to build compilers and simulations to understand the brain’s computation directly. In this episode we go deep into how neurons work, how neurons wire together, and how organic and artificial neural networks differ. We discuss why organic neurons are doing much more; how a model of a single organic neuron can solve MNIST — computing more like a 3-layer artificial neural network; how the brain might learn by solving credit assignment with only local signals; how to approximate backprop without a global algorithm; why AI and humans are intelligent along different dimensions; why Konrad isn’t very worried about AI replacing us; economic models of intelligence and physical work; and much more. Konrad is a brilliant, contrarian thinker who explains complex concepts very intuitively. It is a solid computational neuroscience primer. I hope you enjoy this conversation as much as I did! Other links to this episode and references below. Chapters 00:00:00 Introduction 00:01:01 How organic neurons work 00:24:13 How the brain learns: circuits and credit assignment 00:45:29 Recording the brain 00:52:47 Why simulating brains is hard 01:05:00 A new approach: connectomes and compilers 01:21:00 Why simulate brains? 01:29:50 How AI and human intelligence differ 01:41:04 Evolution, intelligence and AI risk 01:52:42 Robotics, causality, and the roots of intelligence 02:05:53 AI for science and scientific rigor 02:13:05 The economics of intelligence 02:27:50 A hopeful future

Juan Benet

50,121 次观看 • 2 个月前

Demis Hassabis (Demis Hassabis) has had one of the most extraordinary careers in tech. He started as a chess prodigy and video game designer at 17 before getting a PhD in neuroscience and going on to found DeepMind. His lab cracked Go, solved protein structure prediction with AlphaFold, and then gave it away free to every scientist on earth. That work won him the 2024 Nobel Prize in Chemistry. Today he leads Google DeepMind, pushing toward the same goal he set as a teenager: AGI. On this special live episode of How to Build the Future, he sat down with YC's Garry Tan to talk about what still needs to happen to get us to AGI, his advice for founders on how to stay ahead of the curve, and what the next big scientific breakthroughs might be. 01:48 — What’s Missing Before We Get To AGI? 03:36 — Why Memory Is Still Unsolved 06:14 — How AlphaGo Shaped Gemini 08:06 — Why Smaller Models Are Getting So Powerful 10:46 — The 1000x Engineer 12:40 — Continual Learning and the Future of Agents 13:32 — Why AI Still Fails at Basic Reasoning 15:33 — Are Agents Overhyped or Just Getting Started? 18:31 — Can AI Become Truly Creative? 20:26 — Open Models, Gemma, and Local AI 22:26 — Why Gemini Was Built Multimodal 24:08 — What Happens When Inference Gets Cheap? 25:24 — From AlphaFold to the Virtual Cells 28:24 — AI as the Ultimate Tool for Science 30:43 — Advice for Founders 33:30 — The AlphaFold Breakthrough Pattern 35:20 — Can AI Make Real Scientific Discoveries? 37:59 — What to Build Before AGI Arrives
40:56

Demis Hassabis (Demis Hassabis) has had one of the most extraordinary careers in tech. He started as a chess prodigy and video game designer at 17 before getting a PhD in neuroscience and going on to found DeepMind. His lab cracked Go, solved protein structure prediction with AlphaFold, and then gave it away free to every scientist on earth. That work won him the 2024 Nobel Prize in Chemistry. Today he leads Google DeepMind, pushing toward the same goal he set as a teenager: AGI. On this special live episode of How to Build the Future, he sat down with YC's Garry Tan to talk about what still needs to happen to get us to AGI, his advice for founders on how to stay ahead of the curve, and what the next big scientific breakthroughs might be. 01:48 — What’s Missing Before We Get To AGI? 03:36 — Why Memory Is Still Unsolved 06:14 — How AlphaGo Shaped Gemini 08:06 — Why Smaller Models Are Getting So Powerful 10:46 — The 1000x Engineer 12:40 — Continual Learning and the Future of Agents 13:32 — Why AI Still Fails at Basic Reasoning 15:33 — Are Agents Overhyped or Just Getting Started? 18:31 — Can AI Become Truly Creative? 20:26 — Open Models, Gemma, and Local AI 22:26 — Why Gemini Was Built Multimodal 24:08 — What Happens When Inference Gets Cheap? 25:24 — From AlphaFold to the Virtual Cells 28:24 — AI as the Ultimate Tool for Science 30:43 — Advice for Founders 33:30 — The AlphaFold Breakthrough Pattern 35:20 — Can AI Make Real Scientific Discoveries? 37:59 — What to Build Before AGI Arrives

Y Combinator

360,550 次观看 • 3 个月前

When Mudith Jayasekara and I met Gabe Pereyra, we were expecting just another vanilla intro call and instead had the best yarn about research, the state of LLMs, and where intelligence is actually heading. It's rare to meet a founder this deep in the weeds who's also building for one of the most important verticals in this new age of intelligence So it was awesome to sit down with Gabe for an extended discussion on what it take to build agents that can reliably complete work over hours, days, or even longer? We talked about why agents today struggle with search and long context windows and how techniques like KV-cache compaction, synthetic data, and continual learning could help. 0:00 Introduction 0:36 Getting legal agents to review the whole data room 2:08 Data rooms larger than any context window 5:28 How far open-source models can go 7:58 Where specialist models fit in legal AI 10:59 Training legal models when client data is off-limits 13:06 Teaching a model how a law firm works 13:59 What belongs in context vs. model weights 15:36 From firm-wide AI to a model for every lawyer 18:37 What training adds beyond retrieving the right cases 20:26 Why context windows have plateaued 24:01 How models could learn continuously on the job 26:12 Can AI recursively improve AI research? 27:07 Research agents can run experiments but not choose them 30:00 Why open-ended research is hard to train 33:47 Why deployment, not intelligence, is the bottleneck 35:08 The cost of frontier intelligence 36:59 Different neolabs, different paths to intelligence 39:26 Using open datasets to compare research methods 41:13 Conclusion
41:50

When Mudith Jayasekara and I met Gabe Pereyra, we were expecting just another vanilla intro call and instead had the best yarn about research, the state of LLMs, and where intelligence is actually heading. It's rare to meet a founder this deep in the weeds who's also building for one of the most important verticals in this new age of intelligence So it was awesome to sit down with Gabe for an extended discussion on what it take to build agents that can reliably complete work over hours, days, or even longer? We talked about why agents today struggle with search and long context windows and how techniques like KV-cache compaction, synthetic data, and continual learning could help. 0:00 Introduction 0:36 Getting legal agents to review the whole data room 2:08 Data rooms larger than any context window 5:28 How far open-source models can go 7:58 Where specialist models fit in legal AI 10:59 Training legal models when client data is off-limits 13:06 Teaching a model how a law firm works 13:59 What belongs in context vs. model weights 15:36 From firm-wide AI to a model for every lawyer 18:37 What training adds beyond retrieving the right cases 20:26 Why context windows have plateaued 24:01 How models could learn continuously on the job 26:12 Can AI recursively improve AI research? 27:07 Research agents can run experiments but not choose them 30:00 Why open-ended research is hard to train 33:47 Why deployment, not intelligence, is the bottleneck 35:08 The cost of frontier intelligence 36:59 Different neolabs, different paths to intelligence 39:26 Using open datasets to compare research methods 41:13 Conclusion

Charlie O'Neill

91,088 次观看 • 1 个月前

.Jerry Tworek helped drive o1, o3, and Codex at OpenAI where he was VP of Research from 2019 to 2025. Then he left to pursue “types of research that are hard to do at OpenAI.” This week on Unsupervised Learning, I sat down with Jerry to discuss where AI research is headed and what he learned from seven years at the forefront of the field. - Why he left OpenAI after helping create some of its biggest breakthroughs - Why Jerry updated his AGI timeline after building reasoning models - The real limits of scaling reinforcement learning - Why Anthropic has done so well in coding - Inside OpenAI's pivotal decisions - What makes great AI researchers Timestamps: 0:00 Intro 1:26 Scaling Paradigms in AI 3:36 Challenges in Reinforcement Learning 11:48 AGI Timelines 18:36 Converging Labs and Economic Forces 25:05 Jerry's Departure from OpenAI 31:18 Pivotal Decisions in OpenAI's Journey 35:06 Balancing Research and Product Development 38:42 The Future of AI Coding 41:33 Specialization vs. Generalization in AI 48:47 Hiring and Building Research Teams 55:21 Quickfire Listen here: YouTube: Spotify: Apple:
1:02:59

.Jerry Tworek helped drive o1, o3, and Codex at OpenAI where he was VP of Research from 2019 to 2025. Then he left to pursue “types of research that are hard to do at OpenAI.” This week on Unsupervised Learning, I sat down with Jerry to discuss where AI research is headed and what he learned from seven years at the forefront of the field. - Why he left OpenAI after helping create some of its biggest breakthroughs - Why Jerry updated his AGI timeline after building reasoning models - The real limits of scaling reinforcement learning - Why Anthropic has done so well in coding - Inside OpenAI's pivotal decisions - What makes great AI researchers Timestamps: 0:00 Intro 1:26 Scaling Paradigms in AI 3:36 Challenges in Reinforcement Learning 11:48 AGI Timelines 18:36 Converging Labs and Economic Forces 25:05 Jerry's Departure from OpenAI 31:18 Pivotal Decisions in OpenAI's Journey 35:06 Balancing Research and Product Development 38:42 The Future of AI Coding 41:33 Specialization vs. Generalization in AI 48:47 Hiring and Building Research Teams 55:21 Quickfire Listen here: YouTube: Spotify: Apple:

Jacob Effron

43,788 次观看 • 6 个月前

capability != learning new benchtalks with Parth Asawa on continual learning, where we discuss teaching models to learn from experience, measuring learning ability, the bet on parametric models, and more 01:06 What is continual learning? 04:10 Why capability and learning are different 06:13 Why build a benchmark? 08:07 Continual Learning Bench launch and reception 09:13 Anthropic's Fable release and Continual Learning Bench 11:02 How to design tasks for continual learning 18:41 The gain metric 24:01 What good looks like on the leaderboard 29:13 Failure modes: why models can't update their beliefs 31:12 Parametric systems and future architectures 34:30 Open science and AI safety 45:42 Lightning round 49:08 How to contribute to Continual Learning Bench
50:35

capability != learning new benchtalks with Parth Asawa on continual learning, where we discuss teaching models to learn from experience, measuring learning ability, the bet on parametric models, and more 01:06 What is continual learning? 04:10 Why capability and learning are different 06:13 Why build a benchmark? 08:07 Continual Learning Bench launch and reception 09:13 Anthropic's Fable release and Continual Learning Bench 11:02 How to design tasks for continual learning 18:41 The gain metric 24:01 What good looks like on the leaderboard 29:13 Failure modes: why models can't update their beliefs 31:12 Parametric systems and future architectures 34:30 Open science and AI safety 45:42 Lightning round 49:08 How to contribute to Continual Learning Bench

vincent sunn chen

40,884 次观看 • 2 个月前

Most people experience AI through apps. Those apps run on infrastructure owned by a handful of companies. Ben Fielding and Harry Grieve of gensyn on why that needs to change, and what it's going to take to rebuild AI as crypto-native open infrastructure. 00:00 – Intro 00:29 – Misconceptions about AI infrastructure 01:20 – Why centralization in AI is a deeper problem than people realize 04:19 – Why AI needs crypto 05:51 – How AI models are trained 08:15 – The rise of autonomous AI agents with onchain identities 10:37 – Lightning round
17:23

Most people experience AI through apps. Those apps run on infrastructure owned by a handful of companies. Ben Fielding and Harry Grieve of gensyn on why that needs to change, and what it's going to take to rebuild AI as crypto-native open infrastructure. 00:00 – Intro 00:29 – Misconceptions about AI infrastructure 01:20 – Why centralization in AI is a deeper problem than people realize 04:19 – Why AI needs crypto 05:51 – How AI models are trained 08:15 – The rise of autonomous AI agents with onchain identities 10:37 – Lightning round

a16z crypto

24,266 次观看 • 4 个月前

He just quit. He was a primary editor for Diary of a CEO. So I asked him to tell all of his secrets for how he helps the Diary of a CEO be the number 1 podcast on YouTube. 00:00:00 Why Trailers Are Essential for Podcast Growth 00:03:30 Why Most People Are Doing Trailers Wrong 00:06:36 From Film and TV to Diary of a CEO 00:11:48 Retention Rates and What They Actually Mean 00:14:52 What Actually Makes a Podcast Good 00:17:00 Why the Guest Is 50% of Everything 00:22:59 What Makes a Great Podcast Host 00:26:31 Red Flags in Podcast Trailers 00:30:33 The Jefferson Fisher Hook Breakdown 00:34:16 The Louis Tomlinson Hook Breakdown 00:42:35 The 15 Point Checklist for the Perfect Trailer 00:45:28 The Oz Pearlman Hook (41 Second Rule) 00:48:06 The Laughter Track Hack 00:54:11 Sound Design and Visual Details 00:57:04 How to Choose Music Like a Film Director 01:00:02 Disney Imagineering Lesson 01:02:46 The Vanessa Van Edwards Hook Breakdown 01:07:02 The Kevin Hart Narrative Arc Breakdown 01:16:40 The A-B-C Storytelling Framework 01:20:13 The Easiest Win You Can Implement Right Now 01:22:36 How Diary of a CEO Uses AI
1:30:01

He just quit. He was a primary editor for Diary of a CEO. So I asked him to tell all of his secrets for how he helps the Diary of a CEO be the number 1 podcast on YouTube. 00:00:00 Why Trailers Are Essential for Podcast Growth 00:03:30 Why Most People Are Doing Trailers Wrong 00:06:36 From Film and TV to Diary of a CEO 00:11:48 Retention Rates and What They Actually Mean 00:14:52 What Actually Makes a Podcast Good 00:17:00 Why the Guest Is 50% of Everything 00:22:59 What Makes a Great Podcast Host 00:26:31 Red Flags in Podcast Trailers 00:30:33 The Jefferson Fisher Hook Breakdown 00:34:16 The Louis Tomlinson Hook Breakdown 00:42:35 The 15 Point Checklist for the Perfect Trailer 00:45:28 The Oz Pearlman Hook (41 Second Rule) 00:48:06 The Laughter Track Hack 00:54:11 Sound Design and Visual Details 00:57:04 How to Choose Music Like a Film Director 01:00:02 Disney Imagineering Lesson 01:02:46 The Vanessa Van Edwards Hook Breakdown 01:07:02 The Kevin Hart Narrative Arc Breakdown 01:16:40 The A-B-C Storytelling Framework 01:20:13 The Easiest Win You Can Implement Right Now 01:22:36 How Diary of a CEO Uses AI

Grant

205,444 次观看 • 24 天前

In this episode of the Lightcone Podcast, we sat down with Mukund Jha and Madhav Jha, the founders of Emergent - an AI platform that lets anyone build and ship production-ready software. In just eight months, users have created more than 7 million apps on Emergent, with the number doubling in just the last 45 days. We discuss how they built one of the most powerful AI coding agents, why they focused on non-technical users and what it's like building for a global audience from India. 00:00 - Intro 01:06 - What Is Emergent? 01:18 - Founder Backstory 02:09 - From AI Testing to General Coding Agents 02:52 - Getting Ahead of the Market 04:18 - The Pivot to Non-Technical Users 05:22 - Why Second Movers Can Win in AI 09:04 - Building for Production, Not Just Prototypes 18:21 - Live Demo: Building Apps with Emergent 24:40 - How Emergent Hires and Runs a Lean Team 29:04 - Is SaaS Dead? The Rise of Personalized Software 34:04 - The Future: Niche Apps, Solo Builders and AI Agency
39:32

In this episode of the Lightcone Podcast, we sat down with Mukund Jha and Madhav Jha, the founders of Emergent - an AI platform that lets anyone build and ship production-ready software. In just eight months, users have created more than 7 million apps on Emergent, with the number doubling in just the last 45 days. We discuss how they built one of the most powerful AI coding agents, why they focused on non-technical users and what it's like building for a global audience from India. 00:00 - Intro 01:06 - What Is Emergent? 01:18 - Founder Backstory 02:09 - From AI Testing to General Coding Agents 02:52 - Getting Ahead of the Market 04:18 - The Pivot to Non-Technical Users 05:22 - Why Second Movers Can Win in AI 09:04 - Building for Production, Not Just Prototypes 18:21 - Live Demo: Building Apps with Emergent 24:40 - How Emergent Hires and Runs a Lean Team 29:04 - Is SaaS Dead? The Rise of Personalized Software 34:04 - The Future: Niche Apps, Solo Builders and AI Agency

Y Combinator

107,673 次观看 • 5 个月前

New Benchtalks with John Yang: on ProgramBench (0% frontier models at launch) and the lineage/future of coding benchmarks, from SWE-bench/InterCode to now 01:29 ProgramBench launch and reception 03:41 Why artifact-level evaluation, not code-level 06:03 Why models love Python 08:29 ProgramBench as a research tool 12:45 From SWE-bench & InterCode to ProgramBench 17:47 How to grade a coding model 21:53 The position paper & humans in the loop 25:01 Managing quality with agents-in-the-loop 28:40 Internet access and benchmark integrity 35:26 Where models may surpass human abilities 38:56 When a model hits 80% on ProgramBench 43:55 Benchmarks worth paying attention to 46:24 What benchmark do you wish existed 49:32 Will benchmarks still look like benchmarks in 5 years 52:02 How to contribute to ProgramBench
54:13

New Benchtalks with John Yang: on ProgramBench (0% frontier models at launch) and the lineage/future of coding benchmarks, from SWE-bench/InterCode to now 01:29 ProgramBench launch and reception 03:41 Why artifact-level evaluation, not code-level 06:03 Why models love Python 08:29 ProgramBench as a research tool 12:45 From SWE-bench & InterCode to ProgramBench 17:47 How to grade a coding model 21:53 The position paper & humans in the loop 25:01 Managing quality with agents-in-the-loop 28:40 Internet access and benchmark integrity 35:26 Where models may surpass human abilities 38:56 When a model hits 80% on ProgramBench 43:55 Benchmarks worth paying attention to 46:24 What benchmark do you wish existed 49:32 Will benchmarks still look like benchmarks in 5 years 52:02 How to contribute to ProgramBench

vincent sunn chen

26,838 次观看 • 2 个月前

Context Rot! What is it? What does it mean? New video breaking down Chroma's latest paper and applying it to interacting with AI 00:00 - What is Context? 04:44 - The Flaw of "Needle in a Haystack" 10:22 - How Distractors Degrade AI Performance 14:31 - The Impact of Prompt Structure 25:55 - Why Unrelated Needles Are Easier to Find 31:02 - The Unreliability of Text Replication 37:03 - Actionable Takeaways for Better Prompting
40:35

Context Rot! What is it? What does it mean? New video breaking down Chroma's latest paper and applying it to interacting with AI 00:00 - What is Context? 04:44 - The Flaw of "Needle in a Haystack" 10:22 - How Distractors Degrade AI Performance 14:31 - The Impact of Prompt Structure 25:55 - Why Unrelated Needles Are Easier to Find 31:02 - The Unreliability of Text Replication 37:03 - Actionable Takeaways for Better Prompting

matt palmer

22,103 次观看 • 1 年前