Lex Fridman Podcast
#434 – Aravind Srinivas: Perplexity CEO on Future of AI, Search & the Internet
Aravind Srinivas
And completion is based on whether the task was achieved, which will be verified by humans. So you do need to set up an RL sandbox for these agents to play and test and verify.
0
💬
0
Comments
Log in to comment.
There are no comments yet.