• Windex007@lemmy.world
    link
    fedilink
    arrow-up
    15
    arrow-down
    5
    ·
    3 days ago

    I mean, in this case it’s what it’s been fed, which is roughly the internet, which is roughly a giant heap of garbage.

    We, collectively, are the programmers, and we fucking suck.

    • gibmiser@lemmy.world
      link
      fedilink
      arrow-up
      18
      ·
      3 days ago

      You miss the point. The programmers control the output to some degree by limiting it or instructing it to prioritize certain information or narratives.

      • Bonesince1997@lemmy.world
        link
        fedilink
        arrow-up
        3
        ·
        2 days ago

        I assume this is where you would hear claims of a lack of conservative views and bias, when really, this is the output that we get. Every time.

    • brucethemoose@lemmy.world
      link
      fedilink
      arrow-up
      2
      ·
      edit-2
      3 days ago

      Traning data is curated and continous.

      In other words, one (for example, Musk) can finetune the big language model on a small pattern of data (for example, antisemetic content) to ‘steer’ the LLM’s outputs towards that.

      You could bias it towards fluffy bunny discussions, then turn around and send it the other direction.

      Each round of finetuning does “lobotomize” the model to some extent though, making it forget stuff, overuses common phrases, reducing its ability to generalize, ‘erasing’ careful anti-reptition tuning and stuff like that. In other words, if Elon is telling his engineers “I don’t like these responses. Make the AI less woke, right now,” he’s basically sabotaging their work. They’d have to start over with the pretrain and sprinkle that data into months(?) of retraining to keep it from dumbing down or going off the rails.

      There are ways around this outlined in research papers (and some open source projects), but Big Tech is kinda dumb and ‘lazy’ since they’re so flush with cash, so they don’t use them. Shrug.