Thread

1. skerit · 1789503470 · focus · HN ↗

   So in theory you could feed it incomplete text, and then ask it for the probabilities of what the next character could be?

   1. vatsachak · 1789503597 · focus · HN ↗

      If you provide it an AST of the english language, yes.
   2. CompleteSkeptic · 1789503943 · focus · HN ↗

      you could, but it the model is not optimized for text

      this is complex, but generating text is highly complicated and requires mode dropping to make long cohesive text

   3. mckngbrd · 1789505644 · focus · HN ↗

      I think the joke here is getting missed

      1. zenlikethat · 1789508623 · focus · HN ↗

         Strings trigger us
   4. aghilmort · 1789519674 · focus · HN ↗

      was wondering same
   5. iforgotmypasswo · 1789526842 · focus · HN ↗

      Or a partially completed song, asking for the next note. I’m not sure if you’re joking, but using it for space constrained next token generation within a grammar sounds like a really neat use case.

      1. copperx · 1789533643 · focus · HN ↗

         Feed the generated note back into the input for the next query and you have ... autoregression?

Unofficial Hacker News client; not affiliated with Y Combinator.