Richard Sutton Leaves Keen to Found OakLab
量子位 · wechat · 2026-07-14
Reinforcement learning pioneer and 2024 Turing Award winner Richard Sutton has left John Carmack's Keen Technologies to co-found a new lab, OakLab, with his student Khurram Javed.
Key highlights from the report:
- Sutton believes the current deep learning path is weak and inefficient, requiring new foundational ideas and restructuring rather than minor tweaks.
- OakLab aims to build agents capable of real-time learning and planning, with a long-term vision of achieving a "trillion-parameter, 20-watt" system.
- The team proposed the OaK (Options and Knowledge) architecture, emphasizing that agents should abstract verifiable, reusable skills from experience and learn online through continuous interaction.
- The article reviews Sutton's academic timeline: temporal difference learning, his RL textbook, "The Bitter Lesson," and his latest views on the "Experience Era" of intelligence.
The article concludes by mentioning that Sutton's first stop after starting the company will be WAIC in Shanghai, where he will give a presentation titled "First Principles of Reinforcement Learning: Raising Superintelligence from Experience."
Related event: Turing Award Winner Sutton Leaves Keen to Found OakLab(10 posts)→
More from Companies & People
- A simple standup question exposes who owns AI model approval in customer workflows — YvesMulkers · 2026-07-21
- Ben’s Bites roundup highlights Kimi K3, Fable 5, Cursor costs and self-driving companies — Ben's Bites · 2026-07-21
- Microsoft expands Mistral models across Azure, Foundry, Copilot Studio and Azure Local — arthurmensch · 2026-07-21
- Veridas and Fourthline merge into a global identity platform for 115 million verifications — TechNadu · 2026-07-21
- Anthropic's AI Submits 65% of Engineering PRs, Slashes Prompts by 80% — Simon Willison · 2026-07-21
- LangChain brings its Interrupt agent conference back this fall in New York and London — LangChain · 2026-07-21