• KairuByte@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    15
    arrow-down
    9
    ·
    8 months ago

    maybe the whole damn thing is outsourced to ChatGPT now, who the fuck knows.

    I don’t understand why so many people assume an LLM would make glaring errors like this…

    • drislands@lemmy.world
      link
      fedilink
      English
      arrow-up
      19
      arrow-down
      6
      ·
      8 months ago

      …because they frequently do? Glaring errors are like, the main thing LLMs produce besides hype.

      • KairuByte@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        34
        arrow-down
        5
        ·
        edit-2
        8 months ago

        They make glaring errors in logic, and confidently state things that are not true. But their whole “deal” is writing proper sentences based on predictive models. They don’t make mistakes like the excerpt highlighted.

        • drislands@lemmy.world
          link
          fedilink
          English
          arrow-up
          4
          ·
          8 months ago

          Y’know what, that’s a fair point. Though I’m not the original commenter from the top, heh.

        • Garbanzo@lemmy.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          8 months ago

          I’m imagining that the first output didn’t cover everything they wanted so they tweaked it and pasted the results together and fucked it up.

        • GlitterInfection@lemmy.world
          link
          fedilink
          English
          arrow-up
          1
          ·
          8 months ago

          Pretty soon glaring errors like this will be the only way to identify human vs LLM writing.

          Then soon after that the LLMs will start producing glaring grammatical errors to match the humans.