Thread

1. nbulka · 1790025933 · focus · HN ↗

   I really wish the models were smart enough to "tail their own logs" and tell us whose work/papers/chats were the catalyst for these math insights. What makes their "internal" model so much better at math? Must be trained on a whole lot of honesty.

   1. pixl97 · 1790029074 · focus · HN ↗

      The models in use are the class of highly persistent models after Astra that caused a number of problems with hacking.

Unofficial Hacker News client; not affiliated with Y Combinator.