• KillingTimeItself
    link
    fedilink
    English
    324 months ago

    it’s only going to get worse, especially as datasets deteriorate.

    With things like reddit being overrun by AI, and also selling AI training data, i can only imagine what mess that’s going to cause.

    • @[email protected]
      link
      fedilink
      English
      54 months ago

      Hallucinations, like depression, is a multifaceted issue. Training data is only a piece of it. Quantized models, overfitted training models rely on memory at the cost of obviously correct training data. Poorly structured Inferences can confuse a model.

      Rest assured, this isn’t just training data.

      • KillingTimeItself
        link
        fedilink
        English
        24 months ago

        yeah there’s also this stuff as well, though i consider that to be a more technical challenge, rather than a hard limit.

    • @[email protected]
      link
      fedilink
      English
      44 months ago

      I think you are spot on. I tend to think the problems may begin to outnumber the potentials.

      • KillingTimeItself
        link
        fedilink
        English
        54 months ago

        and we haven’t even gotten into the problem of what happens when you have no more data to feed it, do you make more? That’s an impossible task.

        • Natanael
          link
          fedilink
          English
          14 months ago

          There’s already attempts to create synthetic data to train on