Episode 22: Archit Sharma, Stanford, on unsupervised and autonomous reinforcement learning
Generally Intelligent
00:00
How to Generate a General Agent
Navy: The only reason we have any hope for generalist agents is because we care about a small distribution of tasks in this entire world. And hopefully there's some structure to be exploited across those tasks, which can help us generalize to those things. "I think one simple insight, maybe I can share is that it's like a human can show you how to do something," Navy says. 'Maybe perhaps that's where you need to maximize your information'
Transcript
Play full episode
