正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Today's AI models train once. We don't work that way. We learn continuously, forget what doesn't matter, and retain what does. That gap is what Dan Biderman and Jessy Lin are closing at Engram. AI that never stops learning, with memory that lives inside the model instead of bolted... on as an afterthought. In our latest Training Data episode we get into why memory is the next frontier: why the brain forgets on purpose, why RAG is a band-aid, and what becomes possible when a model is always training. 00:00 Introduction 00:59 Always Training Explained 01:51 Beyond Context Windows 03:29 Ngram Product Overview 04:34 Adapters And Training Signals 05:32 Internalize Vs Externalize 06:49 Compute And Token Savings 08:19 Teams First Then Individuals 08:51 Memorization Vs Understanding 12:47 Dreams And Offline Digestion 14:08 Training Beats Curation 15:19 Why Everyone Needs A Model 21:44 Bitter Lesson And Architecture 24:44 RAG Killer And KV Cache 31:38 Future Of Memory And Modelsshow more

Sequoia Capital

816,785 subscribers

84,809 次观看 • 4 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

8 条评论

John Yang 的头像

John Yang4 个月前

@dan_biderman @realJessyLin @EngramLab @realJessyLin @dan_biderman 🐐's

Samian 的头像

Samian4 个月前

@dan_biderman @realJessyLin @EngramLab catastrophic forgetting is the make-or-break here. what's their actual mechanism for not stomping old skills when new data lands?

Nick Venturi 的头像

Nick Venturi4 个月前

@dan_biderman @realJessyLin @EngramLab static weights always felt like a temporary hack anyway

Nitish Mutha ⚡️ 的头像

Nitish Mutha ⚡️4 个月前

@dan_biderman @realJessyLin @EngramLab RAG gives you retrieval. Memory is when the model actually changes. That's a different problem entirely.

Zoolatech 的头像

Zoolatech4 个月前

@dan_biderman @realJessyLin @EngramLab Continuous learning is the right direction. But internalization doesn't eliminate the need for external feedback loops, permission boundaries, and drift tracking. Model memory is necessary; governance is separate.

Erika S 的头像

Erika S4 个月前

@dan_biderman @realJessyLin @EngramLab I'm wondering if selective forgetting is actually the feature we're missing. We optimize for retention in ML but the brain forgets on purpose for a reason.

ucixrn 的头像

ucixrn4 个月前

@dan_biderman @realJessyLin @EngramLab We don't forget what doesn't matter, we forget what doesn't repeat

ignacio tartavull 的头像

ignacio tartavull4 个月前

@dan_biderman @realJessyLin @EngramLab I hate when smart researchers get asked the dumbest questions

相关视频

Here's my conversation with Charan Ranganath, psychologist & neuroscientist at UC Davis, specializing in human memory. We talk about memory, imagination, deja vu, false memories, why we remember, why we forget, and a lot of other fascinating mysteries about the human mind. It's here on X in full, and is up on YouTube, Spotify, and everywhere else. Links in comment. Timestamps: 0:00 - Introduction 1:03 - Experiencing self vs remembering self 14:44 - Creating memories 24:16 - Why we forget 31:53 - Training memory 42:22 - Memory hacks 54:10 - Imagination vs memory 1:03:29 - Memory competitions 1:13:18 - Science of memory 1:28:33 - Discoveries 1:39:37 - Deja vu 1:44:54 - False memories 2:04:59 - False confessions 2:08:45 - Heartbreak 2:16:19 - Nature of time 2:24:00 - Brain–computer interface (BCI) 2:38:04 - AI and memory 2:48:18 - ADHD 2:55:15 - Music 3:05:00 - Human mind

Here's my conversation with Charan Ranganath, psychologist & neuroscientist at UC Davis, specializing in human memory. We talk about memory, imagination, deja vu, false memories, why we remember, why we forget, and a lot of other fascinating mysteries about the human mind. It's here on X in full, and is up on YouTube, Spotify, and everywhere else. Links in comment. Timestamps: 0:00 - Introduction 1:03 - Experiencing self vs remembering self 14:44 - Creating memories 24:16 - Why we forget 31:53 - Training memory 42:22 - Memory hacks 54:10 - Imagination vs memory 1:03:29 - Memory competitions 1:13:18 - Science of memory 1:28:33 - Discoveries 1:39:37 - Deja vu 1:44:54 - False memories 2:04:59 - False confessions 2:08:45 - Heartbreak 2:16:19 - Nature of time 2:24:00 - Brain–computer interface (BCI) 2:38:04 - AI and memory 2:48:18 - ADHD 2:55:15 - Music 3:05:00 - Human mind

Lex Fridman

294,769 次观看 • 2 年前

In this episode, Engram co-founder and CEO Dan Biderman joins allen to cook Mediterranean meatballs with yellow rice and talk about building AI that actually learns from you: why long context, RAG, and compaction eventually break down, how Engram compresses knowledge into cartridges and model weights, what continual learning could unlock for long-horizon agents, why token efficiency is inseparable from intelligence, how personal models could improve like Tamagotchis, and what it takes to build the research and infrastructure for millions of continuously updated AI memories. Timestamps: 0:00 Intro 0:26 Engram’s $98M Launch and Meatballs 1:45 From Naval Special Operations to AI Research 4:32 Israeli Military Culture and Founder Maturity 7:12 Why Engram Is Betting on Context and Continual Learning 9:14 Knowledge Cartridges, Compression, and Model Intuition 14:10 Trillion-Token Company Knowledge and Context Rot 18:05 Long-Context Limits, Compaction, and Neural Memory 22:20 Test-Time Training and “Destroying Prefill” 24:31 Harvey and Holistic Enterprise Queries Beyond RAG 27:02 Personal AI Models and Tamagotchi Weights 30:00 What Belongs in Weights vs. Text 32:25 Autonomous Memory and User-Specific Feedback Loops 34:20 Token Efficiency, Model Routing, and Harder Tasks 38:03 Engram’s Research Team and Product Culture 43:02 Hiring Researchers and Infrastructure Engineers 45:25 Doing More With Less 47:41 Where to Find Engram 48:19 Final Taste Test

In this episode, Engram co-founder and CEO Dan Biderman joins allen to cook Mediterranean meatballs with yellow rice and talk about building AI that actually learns from you: why long context, RAG, and compaction eventually break down, how Engram compresses knowledge into cartridges and model weights, what continual learning could unlock for long-horizon agents, why token efficiency is inseparable from intelligence, how personal models could improve like Tamagotchis, and what it takes to build the research and infrastructure for millions of continuously updated AI memories. Timestamps: 0:00 Intro 0:26 Engram’s $98M Launch and Meatballs 1:45 From Naval Special Operations to AI Research 4:32 Israeli Military Culture and Founder Maturity 7:12 Why Engram Is Betting on Context and Continual Learning 9:14 Knowledge Cartridges, Compression, and Model Intuition 14:10 Trillion-Token Company Knowledge and Context Rot 18:05 Long-Context Limits, Compaction, and Neural Memory 22:20 Test-Time Training and “Destroying Prefill” 24:31 Harvey and Holistic Enterprise Queries Beyond RAG 27:02 Personal AI Models and Tamagotchi Weights 30:00 What Belongs in Weights vs. Text 32:25 Autonomous Memory and User-Specific Feedback Loops 34:20 Token Efficiency, Model Routing, and Harder Tasks 38:03 Engram’s Research Team and Product Culture 43:02 Hiring Researchers and Infrastructure Engineers 45:25 Doing More With Less 47:41 Where to Find Engram 48:19 Final Taste Test

Latent.Space: AI News & Views

35,398 次观看 • 3 个月前

Reasoning from scratch, round number 6! An introduction (and implementation) of Reinforcement Learning with Verifiable Rewards (RLVR) and Group Relative Policy Optimization (GRPO). 00:00 Introduction 01:54 What makes a reasoning model different? 04:25 Reasoning traces and model capability 08:29 Accuracy and format rewards 11:34 Aha moments and DeepSeek-R1 training 14:41 Reasoning effort and answer length 18:38 RLHF and RLVR 23:04 GRPO vs. PPO 26:40 GRPO explained with a cooking analogy 31:43 The KL term and simplified GRPO 35:04 Loading the pretrained model 36:07 Loading the MATH training data 39:26 Sampling model responses 46:30 Computing verifiable rewards 49:55 Computing advantages 51:54 Token and sequence log probabilities 55:29 Implementing sequence log probabilities 57:37 Fixing the inference-mode error 1:02:24 Computing the GRPO loss 1:04:37 Putting the GRPO step together 1:09:19 The GRPO training loop 1:12:57 Training settings, logging, and checkpoints 1:17:24 Running training and inspecting outputs 1:19:28 Loading and evaluating checkpoints 1:22:33 MATH-500 results and training stability 1:24:05 Memory requirements and next steps

Reasoning from scratch, round number 6! An introduction (and implementation) of Reinforcement Learning with Verifiable Rewards (RLVR) and Group Relative Policy Optimization (GRPO). 00:00 Introduction 01:54 What makes a reasoning model different? 04:25 Reasoning traces and model capability 08:29 Accuracy and format rewards 11:34 Aha moments and DeepSeek-R1 training 14:41 Reasoning effort and answer length 18:38 RLHF and RLVR 23:04 GRPO vs. PPO 26:40 GRPO explained with a cooking analogy 31:43 The KL term and simplified GRPO 35:04 Loading the pretrained model 36:07 Loading the MATH training data 39:26 Sampling model responses 46:30 Computing verifiable rewards 49:55 Computing advantages 51:54 Token and sequence log probabilities 55:29 Implementing sequence log probabilities 57:37 Fixing the inference-mode error 1:02:24 Computing the GRPO loss 1:04:37 Putting the GRPO step together 1:09:19 The GRPO training loop 1:12:57 Training settings, logging, and checkpoints 1:17:24 Running training and inspecting outputs 1:19:28 Loading and evaluating checkpoints 1:22:33 MATH-500 results and training stability 1:24:05 Memory requirements and next steps

Sebastian Raschka

76,179 次观看 • 1 个月前

Why AI Progress Suddenly Feels Real - my conversation with Yann Dubois, who co-leads the Post-Training Frontiers team at OpenAI 00:00 - Intro 01:30 - Why recent AI progress feels like a step function 04:13 - Model reliability & the emotional rollercoaster of shipping GPT-5.5 07:33 - How OpenAI structures vertical and horizontal teams 09:49 - Improving model efficiency and test-time compute 12:32 - Yann's journey from Switzerland to OpenAI 15:37 - Reasoning in 2026: Real-world utility vs verifiable rewards 18:34 - GPT-5.5 Thinking vs Pro: Scaling test-time compute 20:09 - How reasoning models become more efficient 23:23 - Pre-training scaling and overcoming the data wall 27:03 - Multimodal data, synthetic data, and embodied AI 31:05 - Demystifying mid-training and post-training 37:21 - Does RL create new capabilities in AI? 38:53 - The challenges and frontier of scaling RL 43:09 - Is building AI models a craft or a strict science 48:21 - How AI models generalize across different domains 54:18 - How reinforcement learning cures AI hallucinations 56:04 - Negative generalization and conflicting instructions 58:05 - Can RL scale to law, medicine, and the broader economy? 1:00:19 - The evaluation bottleneck and Model as a Judge 1:04:21 - Continuous AI progress & continual learning 1:08:49 - Will foundation models eat the agent harness 1:11:23 - Why startups should focus on the last mile of AI

Why AI Progress Suddenly Feels Real - my conversation with Yann Dubois, who co-leads the Post-Training Frontiers team at OpenAI 00:00 - Intro 01:30 - Why recent AI progress feels like a step function 04:13 - Model reliability & the emotional rollercoaster of shipping GPT-5.5 07:33 - How OpenAI structures vertical and horizontal teams 09:49 - Improving model efficiency and test-time compute 12:32 - Yann's journey from Switzerland to OpenAI 15:37 - Reasoning in 2026: Real-world utility vs verifiable rewards 18:34 - GPT-5.5 Thinking vs Pro: Scaling test-time compute 20:09 - How reasoning models become more efficient 23:23 - Pre-training scaling and overcoming the data wall 27:03 - Multimodal data, synthetic data, and embodied AI 31:05 - Demystifying mid-training and post-training 37:21 - Does RL create new capabilities in AI? 38:53 - The challenges and frontier of scaling RL 43:09 - Is building AI models a craft or a strict science 48:21 - How AI models generalize across different domains 54:18 - How reinforcement learning cures AI hallucinations 56:04 - Negative generalization and conflicting instructions 58:05 - Can RL scale to law, medicine, and the broader economy? 1:00:19 - The evaluation bottleneck and Model as a Judge 1:04:21 - Continuous AI progress & continual learning 1:08:49 - Will foundation models eat the agent harness 1:11:23 - Why startups should focus on the last mile of AI

Matt Turck

101,646 次观看 • 5 个月前

🚀 AI Deep Dive Episode 6: Are We Building World Models The Wrong Way? 🚀 Rocket and World Mechanics co-founder Sonia Joseph explore how models represent physics, whether interpretability could improve training and why language alone may not be enough for physical AI. 00:00 Intro 01:39 What counts as a world model? 07:12 Looking inside models for physical understanding 11:28 Interpretable training and the bitter lesson 19:24 Physics, language and understandable explanations 28:07 Surgical robots and manufacturing 31:10 Language models and recursive self-improvement 38:13 JEPA and specialized world models 45:25 What video models actually learn about physics

🚀 AI Deep Dive Episode 6: Are We Building World Models The Wrong Way? 🚀 Rocket and World Mechanics co-founder Sonia Joseph explore how models represent physics, whether interpretability could improve training and why language alone may not be enough for physical AI. 00:00 Intro 01:39 What counts as a world model? 07:12 Looking inside models for physical understanding 11:28 Interpretable training and the bitter lesson 19:24 Physics, language and understandable explanations 28:07 Surgical robots and manufacturing 31:10 Language models and recursive self-improvement 38:13 JEPA and specialized world models 45:25 What video models actually learn about physics

The Information

18,723 次观看 • 18 天前

Anthropic co-founder Jared Kaplan on scaling and the road to human-level AI. At AI Startup School in San Francisco. 00:17 - From Physics to AI 01:41 - Initial Skepticism and Shift to AI 02:12 - AI Training Phases 02:32 - Pre-Training 03:16 - Reinforcement Learning 04:19 - Scaling Laws in Training 08:19 - Unlocking AI Capabilities 11:27 - Organizational Knowledge and Memory 12:19 - Oversight and Nuanced Tasks 13:38 - Preparing for the Future 15:48 - Claude 4 and Beyond 21:18 - Human-AI Collaboration 29:50 - Scaling Laws and Compute Efficiency 35:26 - Audience Q&A

Anthropic co-founder Jared Kaplan on scaling and the road to human-level AI. At AI Startup School in San Francisco. 00:17 - From Physics to AI 01:41 - Initial Skepticism and Shift to AI 02:12 - AI Training Phases 02:32 - Pre-Training 03:16 - Reinforcement Learning 04:19 - Scaling Laws in Training 08:19 - Unlocking AI Capabilities 11:27 - Organizational Knowledge and Memory 12:19 - Oversight and Nuanced Tasks 13:38 - Preparing for the Future 15:48 - Claude 4 and Beyond 21:18 - Human-AI Collaboration 29:50 - Scaling Laws and Compute Efficiency 35:26 - Audience Q&A

Y Combinator

116,444 次观看 • 1 年前

A lot of people are calling Hermes Agent the end of OpenClaw. BRUH! It's not... Nous Research trains actual models and they built an agent around that expertise. The local model routing is solid, but the part that matters for your business is that your conversations become fine-tuning data. You can train a model on how you actually work. 00:00 The Problem with Local AI Models 00:25 Introduction to Nous Research 01:04 Cross-Platform Agent Capabilities 01:44 Deep Local Model Integration 02:30 Routing Tasks to Different Models 03:06 Conversations as Training Data 03:50 Hermes Agent vs. OpenClaw 04:15 Future Plans and Series Overview

A lot of people are calling Hermes Agent the end of OpenClaw. BRUH! It's not... Nous Research trains actual models and they built an agent around that expertise. The local model routing is solid, but the part that matters for your business is that your conversations become fine-tuning data. You can train a model on how you actually work. 00:00 The Problem with Local AI Models 00:25 Introduction to Nous Research 01:04 Cross-Platform Agent Capabilities 01:44 Deep Local Model Integration 02:30 Routing Tasks to Different Models 03:06 Conversations as Training Data 03:50 Hermes Agent vs. OpenClaw 04:15 Future Plans and Series Overview

Ray Fernando

42,398 次观看 • 7 个月前

When should you start post-training your own models? Fireworks CEO Lin Qiao’s answer: after product-market fit. Not because it's hard... but because only after PMF is the data coming off your product surface worth training on. Lin joined us for our Sequoia Capital "Own Your Intelligence" event to host a workshop on all things post-training; what works, what breaks, and how not to let the model outsmart you. Must listen!! 00:00 Introduction 00:37 What Fireworks sees across thousands of AI applications 02:47 Off-the-shelf APIs and the problem of keeping your taste 03:58 What "owning your intelligence" actually means 05:43 The progression: prompting → RAG → SFT → preferences → RL 07:20 Why this mirrors how humans learn 09:03 Matching the technique to the problem you actually have 10:46 Where teams get stuck: data quality and vibe evals 12:28 Reward hacking: the model that wrote zero lines of code 13:59 Training-to-serving alignment (and why quality drops) 15:55 Post-training in healthcare and security 17:31 From coding to every co-work domain 19:35 Incumbents, cost burden, and not scaling into bankruptcy 21:26 How much control do you want? 23:24 Q&A: What makes a good reward signal 25:00 Q&A: When to start thinking about post-training

When should you start post-training your own models? Fireworks CEO Lin Qiao’s answer: after product-market fit. Not because it's hard... but because only after PMF is the data coming off your product surface worth training on. Lin joined us for our Sequoia Capital "Own Your Intelligence" event to host a workshop on all things post-training; what works, what breaks, and how not to let the model outsmart you. Must listen!! 00:00 Introduction 00:37 What Fireworks sees across thousands of AI applications 02:47 Off-the-shelf APIs and the problem of keeping your taste 03:58 What "owning your intelligence" actually means 05:43 The progression: prompting → RAG → SFT → preferences → RL 07:20 Why this mirrors how humans learn 09:03 Matching the technique to the problem you actually have 10:46 Where teams get stuck: data quality and vibe evals 12:28 Reward hacking: the model that wrote zero lines of code 13:59 Training-to-serving alignment (and why quality drops) 15:55 Post-training in healthcare and security 17:31 From coding to every co-work domain 19:35 Incumbents, cost burden, and not scaling into bankruptcy 21:26 How much control do you want? 23:24 Q&A: What makes a good reward signal 25:00 Q&A: When to start thinking about post-training

Sonya Huang 🐥

28,212 次观看 • 2 个月前

Stanford '26, Session 3 (Full lecture) The Future of Voice Systems with @matiii from ElevenLabs 00:00 Welcome and Intro 01:31 Origin Story on Discord 05:15 The Dubbing Problem 07:44 Pipeline and Early Pivot 12:38 Building the First Model 15:24 Compute Costs and Patents 17:34 Roadmap Through 2025 22:00 Cascaded vs Fused Agents 30:38 Collaboration Over Competition 35:05 Revenue Growth and Team Design 37:56 Predictable Deployment Engine 42:32 Voice Safety and Watermarking 44:27 Research Bottlenecks Personalization 46:24 Training Tradeoffs Cascade vs Fuse 48:20 Five Year Vision Platform 51:08 Impact Work ALS and Ukraine 54:40 China Distillation and Openness 59:24 Studios AI Voice Economics 01:03:04 On Device Models and Platform Gap 01:04:36 Enterprise Tooling and Wrap Up

Stanford '26, Session 3 (Full lecture) The Future of Voice Systems with @matiii from ElevenLabs 00:00 Welcome and Intro 01:31 Origin Story on Discord 05:15 The Dubbing Problem 07:44 Pipeline and Early Pivot 12:38 Building the First Model 15:24 Compute Costs and Patents 17:34 Roadmap Through 2025 22:00 Cascaded vs Fused Agents 30:38 Collaboration Over Competition 35:05 Revenue Growth and Team Design 37:56 Predictable Deployment Engine 42:32 Voice Safety and Watermarking 44:27 Research Bottlenecks Personalization 46:24 Training Tradeoffs Cascade vs Fuse 48:20 Five Year Vision Platform 51:08 Impact Work ALS and Ukraine 54:40 China Distillation and Openness 59:24 Studios AI Voice Economics 01:03:04 On Device Models and Platform Gap 01:04:36 Enterprise Tooling and Wrap Up

Anjney Midha

79,736 次观看 • 7 个月前

My conversation with Ali Ansari the CEO of micro1, the human data engine helping AI labs train foundational models and enterprises build better AI agents. We talk about why AI still needs human experts, never stop learning, the talent war for researchers, and why the model is not the product. 01:32 — Building Micro1 in the Center of the AI Boom 06:57 — A Once-in-a-Lifetime AI Opportunity 08:44 — How Micro1 Helps Train Smarter AI Models 10:43 — Why AI Models Never Stop Learning 14:56 — Why AI Still Needs Human Experts 17:49 — The Model Is Not the Product 21:23 — Who Wins and Loses in the AI Era 27:25 — When AI Agents Become Team Members 29:19 — The Talent War for AI Researchers 33:38 — Why Your Company Data Is Training AI 38:35 — From Iran to the American Dream 44:33 — Why Taking a Unique Path Can Win 48:04 — Embrace the Chaos 49:50 — Why AI Will Create New Jobs 53:02 — How Ali Measures His Life #AI #ArtificialIntelligence #FutureOfWork #AIData #AIAgents #MachineLearning #Startups #DivotPodcast

My conversation with Ali Ansari the CEO of micro1, the human data engine helping AI labs train foundational models and enterprises build better AI agents. We talk about why AI still needs human experts, never stop learning, the talent war for researchers, and why the model is not the product. 01:32 — Building Micro1 in the Center of the AI Boom 06:57 — A Once-in-a-Lifetime AI Opportunity 08:44 — How Micro1 Helps Train Smarter AI Models 10:43 — Why AI Models Never Stop Learning 14:56 — Why AI Still Needs Human Experts 17:49 — The Model Is Not the Product 21:23 — Who Wins and Loses in the AI Era 27:25 — When AI Agents Become Team Members 29:19 — The Talent War for AI Researchers 33:38 — Why Your Company Data Is Training AI 38:35 — From Iran to the American Dream 44:33 — Why Taking a Unique Path Can Win 48:04 — Embrace the Chaos 49:50 — Why AI Will Create New Jobs 53:02 — How Ali Measures His Life #AI #ArtificialIntelligence #FutureOfWork #AIData #AIAgents #MachineLearning #Startups #DivotPodcast

Derek Andersen

14,495 次观看 • 4 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

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

24,337 次观看 • 6 个月前

When is harness optimization enough? When does post-training make sense? What will determine who becomes the new AI hyperscaler? Yash Patil, co-founder of Applied Compute, was the perfect person to discuss this and more. Yash worked on Codex at OpenAI before leaving to co-found Applied Compute with fellow OpenAI researchers Rhythm Garg and Linden Li. Sixteen months in, the company is a unicorn, and is serving enterprises including DoorDash and Cognition. We cover: 0:00 Intro 1:42 The Case for Owning Your Intelligence 4:35 OpenAI x Baseten and the Multi-Model Future 7:19 Can Post-Training Beat Frontier Models? 9:03 What Data Stays Out of Distribution 11:51 Data Efficiency and Continual Learning 14:31 When Updating Model Weights Makes Sense 16:01 Online Training in Practice 18:16 Lowering the Cost of Post-Training 19:48 Does Every Company Need Its Own Model? 24:06 RL in Non-Verifiable Domains 26:59 How Applied Compute Picks Customers 28:46 Building an Inference Business 34:43 The Case for a New AI Hyperscaler 38:21 Building RL Environments 41:33 Quickfire YouTube: Spotify: Apple:

When is harness optimization enough? When does post-training make sense? What will determine who becomes the new AI hyperscaler? Yash Patil, co-founder of Applied Compute, was the perfect person to discuss this and more. Yash worked on Codex at OpenAI before leaving to co-found Applied Compute with fellow OpenAI researchers Rhythm Garg and Linden Li. Sixteen months in, the company is a unicorn, and is serving enterprises including DoorDash and Cognition. We cover: 0:00 Intro 1:42 The Case for Owning Your Intelligence 4:35 OpenAI x Baseten and the Multi-Model Future 7:19 Can Post-Training Beat Frontier Models? 9:03 What Data Stays Out of Distribution 11:51 Data Efficiency and Continual Learning 14:31 When Updating Model Weights Makes Sense 16:01 Online Training in Practice 18:16 Lowering the Cost of Post-Training 19:48 Does Every Company Need Its Own Model? 24:06 RL in Non-Verifiable Domains 26:59 How Applied Compute Picks Customers 28:46 Building an Inference Business 34:43 The Case for a New AI Hyperscaler 38:21 Building RL Environments 41:33 Quickfire YouTube: Spotify: Apple:

Jacob Effron

27,225 次观看 • 1 个月前

My conversation with OpenAI co-founder Greg Brockman This is the most detailed first-person account of the 72 hours after Sam Altman was fired. We also go deep on what comes next: the global race to AGI, why ChatGPT stopped showing reasoning, how much of OpenAI's own code is now written by AI ("it's hard to know what percent is not"), and the untold story of how OpenAI actually started in 2015. 00:00:00 Introduction 00:00:49 Meeting Sam Altman and Starting OpenAI 00:02:40 Building the Founding Team 00:04:25 DeepMind's Lead Over OpenAI 00:04:54 Changing OpenAI to a For-Profit Model 00:06:05 Breakthrough Moments at OpenAI 00:08:22 What Dota 2 Meant for OpenAI 00:10:04 Reasoning Versus Prediction 00:11:59 Tensions Grow at OpenAI 00:15:44 Sam Altman's Firing 00:17:49 Greg Quits OpenAI 00:19:56 Sam Explores Deal with Microsoft's Satya 00:20:28 Petition for Altman's Return 00:23:43 Ilya Sutskever Leaves OpenAI 00:24:59 Lessons Learned after Sam Ousting 00:28:22 The Thing Ilya Said that Greg Can't Forget 00:32:22 Is AI Going Parabolic? 00:33:24 How Much of OpenAI's Code is Written by AI? 00:36:21 Do AI Chatbots Tell Us What We Want to Hear? 00:38:06 The Global AI Race to Reach AGI 00:38:40 What Happens if US Doesn't Reach AGI First? 00:39:49 Are Countries Stealing AI Advancements? 00:40:38 Why ChatGPT No Longer Shows Reasoning 00:41:47 The Finite Constraints of Compute 00:43:38 On Investing Early in Data Centers 00:46:31 The Future of Data Center Specialization 00:47:52 How to Decide Whose Queries to Serve 00:49:08 OpenAI on Consumer vs Enterprise Models 00:53:05 Data Centers in Space? 01:00:56 What Should AI Regulation Look Like? 01:04:33 The Future of AI-Powered Entrepreneurship 01:04:44 AI and Job Loss 01:07:15 The Skills Young People Should Invest In 01:11:30 What Does Success Look Like For You? Full episode on X below. Also find it on: • YouTube: • Spotify: • Apple:

My conversation with OpenAI co-founder Greg Brockman This is the most detailed first-person account of the 72 hours after Sam Altman was fired. We also go deep on what comes next: the global race to AGI, why ChatGPT stopped showing reasoning, how much of OpenAI's own code is now written by AI ("it's hard to know what percent is not"), and the untold story of how OpenAI actually started in 2015. 00:00:00 Introduction 00:00:49 Meeting Sam Altman and Starting OpenAI 00:02:40 Building the Founding Team 00:04:25 DeepMind's Lead Over OpenAI 00:04:54 Changing OpenAI to a For-Profit Model 00:06:05 Breakthrough Moments at OpenAI 00:08:22 What Dota 2 Meant for OpenAI 00:10:04 Reasoning Versus Prediction 00:11:59 Tensions Grow at OpenAI 00:15:44 Sam Altman's Firing 00:17:49 Greg Quits OpenAI 00:19:56 Sam Explores Deal with Microsoft's Satya 00:20:28 Petition for Altman's Return 00:23:43 Ilya Sutskever Leaves OpenAI 00:24:59 Lessons Learned after Sam Ousting 00:28:22 The Thing Ilya Said that Greg Can't Forget 00:32:22 Is AI Going Parabolic? 00:33:24 How Much of OpenAI's Code is Written by AI? 00:36:21 Do AI Chatbots Tell Us What We Want to Hear? 00:38:06 The Global AI Race to Reach AGI 00:38:40 What Happens if US Doesn't Reach AGI First? 00:39:49 Are Countries Stealing AI Advancements? 00:40:38 Why ChatGPT No Longer Shows Reasoning 00:41:47 The Finite Constraints of Compute 00:43:38 On Investing Early in Data Centers 00:46:31 The Future of Data Center Specialization 00:47:52 How to Decide Whose Queries to Serve 00:49:08 OpenAI on Consumer vs Enterprise Models 00:53:05 Data Centers in Space? 01:00:56 What Should AI Regulation Look Like? 01:04:33 The Future of AI-Powered Entrepreneurship 01:04:44 AI and Job Loss 01:07:15 The Skills Young People Should Invest In 01:11:30 What Does Success Look Like For You? Full episode on X below. Also find it on: • YouTube: • Spotify: • Apple:

Shane Parrish

450,952 次观看 • 6 个月前

*Major* open source AI drop today. Can America win the Open AI race? My conversation with Nathan Lambert and Luca Soldaini 🎀 of Ai2 about the launch of Olmo 3 00:00 – Cold Open 00:39 – Welcome & today’s big announcement 01:18 – Introducing the Olmo 3 model family 02:07 – What “base models” really are (and why they matter) 05:51 – Dolma 3: the data behind Olmo 3 08:06 – Performance vs Qwen, Gemma, DeepSeek 10:28 – What true open source means (and why it’s rare) 12:51 – Intermediate checkpoints, transparency, and why AI2 publishes everything 16:37 – Why Qwen is everywhere (including U.S. startups) 18:31 – Why Chinese labs go open source (and why U.S. labs don’t) 20:28 – Inside ATOM: the U.S. response to China’s model surge 22:13 – The rise of “thinking models” and inference-time scaling 35:58 – The full Olmo pipeline, explained simply 46:52 – Pre-training: data, scale, and avoiding catastrophic spikes 50:27 – Mid-training (tail patching) and avoiding test leakage 52:06 – Why long-context training matters 55:28 – SFT: building the foundation for reasoning 1:04:53 – Preference tuning & why DPO still works 1:10:51 – The hard part: RLVR, long reasoning chains, and infrastructure pain 1:13:59 – Why RL is so technically brutal 1:18:17 – Complexity tax vs AGI hype 1:21:58 – How everyone can contribute to the future of AI 1:27:26 – Closing thoughts

Major open source AI drop today. Can America win the Open AI race? My conversation with Nathan Lambert and Luca Soldaini 🎀 of Ai2 about the launch of Olmo 3 00:00 – Cold Open 00:39 – Welcome & today’s big announcement 01:18 – Introducing the Olmo 3 model family 02:07 – What “base models” really are (and why they matter) 05:51 – Dolma 3: the data behind Olmo 3 08:06 – Performance vs Qwen, Gemma, DeepSeek 10:28 – What true open source means (and why it’s rare) 12:51 – Intermediate checkpoints, transparency, and why AI2 publishes everything 16:37 – Why Qwen is everywhere (including U.S. startups) 18:31 – Why Chinese labs go open source (and why U.S. labs don’t) 20:28 – Inside ATOM: the U.S. response to China’s model surge 22:13 – The rise of “thinking models” and inference-time scaling 35:58 – The full Olmo pipeline, explained simply 46:52 – Pre-training: data, scale, and avoiding catastrophic spikes 50:27 – Mid-training (tail patching) and avoiding test leakage 52:06 – Why long-context training matters 55:28 – SFT: building the foundation for reasoning 1:04:53 – Preference tuning & why DPO still works 1:10:51 – The hard part: RLVR, long reasoning chains, and infrastructure pain 1:13:59 – Why RL is so technically brutal 1:18:17 – Complexity tax vs AGI hype 1:21:58 – How everyone can contribute to the future of AI 1:27:26 – Closing thoughts

Matt Turck

37,531 次观看 • 11 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

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

93,823 次观看 • 3 个月前

1/ BREAKING – DePIN's Biggest New Deal: Valeo x NATIX Network with Alireza Ghods Host: Connor Lovely 🎬 We discuss: - Why a $3B, Fortune 500 company wants Natix data - What is a World Foundational Model? - Why autonomous driving needs synthetic data Timestamps: 00:00 - Introduction 01:14 - Updates on VX360 and Data Sales 02:09 - Vision Language Model and Its Applications 04:39 - Importance of Diverse Data for World Models 06:20 - Details on the Valeo Deal 08:15 - Use Cases and Future Plans for World Models 10:45 - Synthetic Data and Real-World Applications 19:37 - Future of Physical AI and Autonomous Driving 23:51 - Conclusion Watch on other platforms below ⬇️

1/ BREAKING – DePIN's Biggest New Deal: Valeo x NATIX Network with Alireza Ghods Host: Connor Lovely 🎬 We discuss: - Why a $3B, Fortune 500 company wants Natix data - What is a World Foundational Model? - Why autonomous driving needs synthetic data Timestamps: 00:00 - Introduction 01:14 - Updates on VX360 and Data Sales 02:09 - Vision Language Model and Its Applications 04:39 - Importance of Diverse Data for World Models 06:20 - Details on the Valeo Deal 08:15 - Use Cases and Future Plans for World Models 10:45 - Synthetic Data and Real-World Applications 19:37 - Future of Physical AI and Autonomous Driving 23:51 - Conclusion Watch on other platforms below ⬇️

Proof of Coverage | Tech’s Creative Agency

11,889 次观看 • 9 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

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

71,317 次观看 • 5 个月前

What is AI inference engineering, why is it such an in-demand skill, and how do you break into the field? With author of Inference Engineering Philip Kiely and head of training at Baseten Charlie O'Neill 0:00: What is inference? 2:47: History of inference 4:59: Downstream effects of AI research on inference 13:54: What you'll learn from Inference Engineering 16:14: Advice for engineers transitioning into AI 19:00: Open source models driving inference growth 20:55: Specialization vs. frontier closed models 23:51: "Big Token" and the importance of open source AI 27:18: Where to get Inference Engineering

What is AI inference engineering, why is it such an in-demand skill, and how do you break into the field? With author of Inference Engineering Philip Kiely and head of training at Baseten Charlie O'Neill 0:00: What is inference? 2:47: History of inference 4:59: Downstream effects of AI research on inference 13:54: What you'll learn from Inference Engineering 16:14: Advice for engineers transitioning into AI 19:00: Open source models driving inference growth 20:55: Specialization vs. frontier closed models 23:51: "Big Token" and the importance of open source AI 27:18: Where to get Inference Engineering

Madison Kanna

119,268 次观看 • 7 个月前

Yoshua Bengio thinks he knows how to make provably safe superintelligent agents. Bengio built the foundations of modern AI and is the most cited living scientist. He believes his alternative training setup would: 1. Guarantee honesty 2. Prevent unintended goals 3. Produce capable agents 4. Port over most data and techniques from current LLMs 5. Not be inherently more expensive, and perhaps be more intelligent Bengio claims the honesty and lack of unintended goals can be proven mathematically, at least given particular assumptions. And his new organization, LawZero, is aiming to build a scrappy prototype as soon as possible. The architecture is called 'Scientist AI' and it's based on training a model to explain empirical observations, including what people say, rather than training AIs that mimic human behaviour or seek our approval. (Bengio's frank assessment is that "reinforcement learning is evil" and that allowing AIs to independently train their successors is "the most crazy, dangerous bet that unfortunately we are on track to do.") But skeptics question whether Scientist AI really does solve the fundamental problem of 'eliciting latent knowledge' from AI models. And with the commercial race for superintelligence so intense, it's not clear whether the proposal will be able to compete or have time to bear fruit, even if it's sound in theory. On The 80,000 Hours Podcast, links below – enjoy! • Making AI honest and safe (00:00:00) • Scientist AI in plain English (00:02:27) • How Scientist AI differs from LLMs (00:06:32) • How the training data works (00:14:02) • Can this become an agent? (00:21:02) • Why Yoshua is now more optimistic (00:32:11) • Why companies can’t stop racing (00:36:35) • A working prototype won't take long (00:49:15) • Scientist models might be more capable (00:53:34) • “Reinforcement learning is evil” (01:01:27) • Scientist AI from guardrail to agent (01:08:37) • Can safe AI still be competent? (01:12:38) • How much will this cost? (01:19:29) • Can it generalise beyond maths and science? (01:23:26) • A multi-national push for superintelligence (01:39:19) • Want to work with or fund Yoshua? (01:51:16) • Why smart people ignore AI risk (01:54:45) • Don’t let AI build the next AI (02:01:33) • Why politicians miss the real risks (02:12:28) • Why Yoshua changed his mind about AI risk (02:21:27)

Yoshua Bengio thinks he knows how to make provably safe superintelligent agents. Bengio built the foundations of modern AI and is the most cited living scientist. He believes his alternative training setup would: 1. Guarantee honesty 2. Prevent unintended goals 3. Produce capable agents 4. Port over most data and techniques from current LLMs 5. Not be inherently more expensive, and perhaps be more intelligent Bengio claims the honesty and lack of unintended goals can be proven mathematically, at least given particular assumptions. And his new organization, LawZero, is aiming to build a scrappy prototype as soon as possible. The architecture is called 'Scientist AI' and it's based on training a model to explain empirical observations, including what people say, rather than training AIs that mimic human behaviour or seek our approval. (Bengio's frank assessment is that "reinforcement learning is evil" and that allowing AIs to independently train their successors is "the most crazy, dangerous bet that unfortunately we are on track to do.") But skeptics question whether Scientist AI really does solve the fundamental problem of 'eliciting latent knowledge' from AI models. And with the commercial race for superintelligence so intense, it's not clear whether the proposal will be able to compete or have time to bear fruit, even if it's sound in theory. On The 80,000 Hours Podcast, links below – enjoy! • Making AI honest and safe (00:00:00) • Scientist AI in plain English (00:02:27) • How Scientist AI differs from LLMs (00:06:32) • How the training data works (00:14:02) • Can this become an agent? (00:21:02) • Why Yoshua is now more optimistic (00:32:11) • Why companies can’t stop racing (00:36:35) • A working prototype won't take long (00:49:15) • Scientist models might be more capable (00:53:34) • “Reinforcement learning is evil” (01:01:27) • Scientist AI from guardrail to agent (01:08:37) • Can safe AI still be competent? (01:12:38) • How much will this cost? (01:19:29) • Can it generalise beyond maths and science? (01:23:26) • A multi-national push for superintelligence (01:39:19) • Want to work with or fund Yoshua? (01:51:16) • Why smart people ignore AI risk (01:54:45) • Don’t let AI build the next AI (02:01:33) • Why politicians miss the real risks (02:12:28) • Why Yoshua changed his mind about AI risk (02:21:27)

Rob Wiblin

65,088 次观看 • 6 个月前

A model is only as good as its data, and we’ve long since exhausted the internet. From here on out, model progress is gated by data production. ’s Brendan (can/do) joined us at our Sovereign AI event to talk about how RL environments get built, and why your data might be your real moat: 00:00 Introduction 00:47 A short history of the data market: crowdsourcing to agentic data 02:29 What an RL environment is: worlds, apps, tasks 03:57 Why only humans can measure the frontier 05:35 Building verifiers is the hard part 06:44 Walkthrough: a real legal RL environment 08:18 Leaderboards — and what open weights change 09:45 Post-training results on Apex Agents 11:17 Three ways companies buy data 12:49 Q&A: How do you price data? 14:17 Q&A: What "data quality" actually means 16:42 Q&A: The misunderstanding about synthetic data 18:17 Q&A: Why RL environments now — and what comes after 21:20 Q&A: Can you scale rubric generation with models? 23:00 Q&A: RL environments for cyber defense 25:33 Q&A: Build data in-house or partner?

A model is only as good as its data, and we’ve long since exhausted the internet. From here on out, model progress is gated by data production. ’s Brendan (can/do) joined us at our Sovereign AI event to talk about how RL environments get built, and why your data might be your real moat: 00:00 Introduction 00:47 A short history of the data market: crowdsourcing to agentic data 02:29 What an RL environment is: worlds, apps, tasks 03:57 Why only humans can measure the frontier 05:35 Building verifiers is the hard part 06:44 Walkthrough: a real legal RL environment 08:18 Leaderboards — and what open weights change 09:45 Post-training results on Apex Agents 11:17 Three ways companies buy data 12:49 Q&A: How do you price data? 14:17 Q&A: What "data quality" actually means 16:42 Q&A: The misunderstanding about synthetic data 18:17 Q&A: Why RL environments now — and what comes after 21:20 Q&A: Can you scale rubric generation with models? 23:00 Q&A: RL environments for cyber defense 25:33 Q&A: Build data in-house or partner?

Pat Grady

67,510 次观看 • 2 个月前