Lex Fridman Podcast
#434 – Aravind Srinivas: Perplexity CEO on Future of AI, Search & the Internet
Aravind Srinivas
But this is literally the evolution. Then the focus went more into pieces outside the architecture on data, what data you're training on, what are the tokens, how deduped they are. And then the chinchilla inside, it's not just about making the model bigger, but you want to also make the data set bigger. You want to make sure the tokens are also...
0
💬
0
Comments
Log in to comment.
There are no comments yet.