
Episode 22: Archit Sharma, Stanford, on unsupervised and autonomous reinforcement learning
Generally Intelligent
Language Models
Like you, I mean, these like goal conditioned RL things are like, Oh, that sounds so cool. It seems like we're starting to get to something closer to that with like the vision of language models, etc. So being able to use that interface to communicate the same two robots would be very interesting as well. Something I'm puzzled by is names kind of relearned language. And we kind of like evolve it. Or we generate it,. The Stephen Pinker hypothesis is like all humans at children in red language. Like come in with the ability to create language. And then we like generate a grammar based on what we hear from around us. Reminds me of this
00:00
Transcript
Play full episode
Remember Everything You Learn from Podcasts
Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.