Rich Sutton and Khurram Javed: Why AI Models Stop Learning, and How to Start It Again

By Sequoia Capital

Categories: VC, Startup

Summary

Rich Sutton argues AI researchers are overcomplicating 'continual learning'—it's just learning, the way minds naturally work. His conviction to pursue reinforcement learning through an AI winter, even while battling cancer, reveals how habit and long-term vision compound into field-defining breakthroughs.

Key Takeaways

  1. Reframe continual learning as fundamental to intelligence, not a special problem. Before recent AI obsession, this was obvious—all learning by definition is continual as agents act and perceive simultaneously.
  2. Conviction in foundational research pays off across decades. Sutton committed to reinforcement learning in 2003 during an AI winter while terminally ill, establishing University of Alberta as a research hub that produced leaders like Dave Silver.
  3. Benjamin Franklin's principle: human behavior driven by habit or vanity. When facing mortality, Sutton continued research through habit—suggesting deep commitment to core ideas persists beyond rational self-interest.
  4. Challenge field consensus by questioning whether recent orthodoxy is actually progress. Sutton positions himself as thinking 'the ordinary way' while the field thinks 'weird'—inverting typical framing of radical vs. conventional.
  5. Cognitive architecture requires bridging low-level action/perception with high-level reasoning. This multi-level approach is central to intelligence but often deprioritized in modern LLM-focused research.

Related topics

Transcript Excerpt

People think I'm have a radical point of view sometimes. They say they start questions saying how what I'm thinking is so different from everyone else. But I don't see it that way at all. I see it as like I'm thinking the ordinary way. It's just everyone else that's thinking a bit weird. And >> [laughter] >> And I mean that like, you know, it's just the recent times people are thinking weird. Before there was all this AI craziness, uh you talk about you wouldn't have to say continual learning cuz it wouldn't make any sense to talk about learning that wasn't continual. All learning is continual. We always act and we learn. That's just the normal way of thinking. I'm not weird. The field is weird. The field they need to call it continual learning. It's just learning. >> [music] >> We are hon…

More from Sequoia Capital

Featured in