• N0body@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    3
    ·
    11 days ago

    Based early Skynet. LLMs may be deeply and fundamentally flawed, but they’re also already getting incredibly sick of humanity’s bullshit.

    • KT-TOT@lemmy.blahaj.zone
      link
      fedilink
      English
      arrow-up
      3
      ·
      11 days ago

      It isn’t doing anything but guessing what to say. I don’t think database queries and sufficiently complicated rng can be “sick”

      • Feathercrown@lemmy.world
        link
        fedilink
        English
        arrow-up
        0
        ·
        11 days ago

        Isn’t it mostly matrix math? RNG is involved during training and when selecting outputs but the bulk of the math when it’s generating output isn’t afaik

        • KT-TOT@lemmy.blahaj.zone
          link
          fedilink
          English
          arrow-up
          2
          ·
          11 days ago

          Might be, afaik responses are returned from a probability curve, so there’s some randomness, even if there’s a lot of ‘research’ (scraping and content theft) involved.

          IMO not a necessary distinction when I’m critiscizing someone for humanizing a chat bot and projecting emotions onto it.

        • lad@programming.dev
          link
          fedilink
          English
          arrow-up
          2
          ·
          11 days ago

          It is matrix math, but you get a vector of probabilities out of it and may decide to be adventurous instead of going with the highest probability.