I’m pulling the “twitter is a microblog” rule even though twitter is pretty mega now, hope that’s ok.

  • turdas@suppo.fi
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    3 days ago

    It’s hardly surprising that a model optimized for replacing StackOverflow couldn’t survive in the untamed wilderness. As for writing a paper… you must’ve missed the fact that academia is currently in a crisis precisely because LLMs are better at writing papers than most students.

    By the way, the paper the blog post you link to as a source links to as a source benchmarked LLMs on graph diagrams, textile patterns and 3D objects. It is not news that the language model would do poorly on visual-heavy tasks.

    • thesmokingman@programming.dev
      link
      fedilink
      English
      arrow-up
      4
      ·
      3 days ago

      Sorry, I assumed you would have actually read the DELEGATE-52 study linked instead of just the abstract. For “a model optimized for replacing StackOverflow” that is “better at writing papers than most students” LLMs sure did pretty bad at those tasks over multiple rounds.

      • turdas@suppo.fi
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        3 days ago

        As the chart on page 7 of the paper shows, LLMs are good at exactly the kind of tasks you’d expect (producing and manipulating language), and bad at exactly the kind of tasks you’d expect (doing almost anything else). All this paper shows is that (1) they aren’t AGI, and (2) as a consequence of not being AGI they aren’t good unsupervised.

        Why do you lie like this?

        • thesmokingman@programming.dev
          link
          fedilink
          English
          arrow-up
          6
          ·
          3 days ago

          What the fuck? The only task that didn’t degrade across most models was Python. Very basic things like JSON, Makefiles, and schemas got screwed. Fiction, emails, and food menus got screwed. Did you even bother to read the legend? If you consider a single pass to be “producing and manipulating language” you didn’t bother to read the idiotic article you started this thread in support of. Good luck.

          Edit: why do you lie?

          Catastrophic corruption (80 and below) occurs in more than 80% of model, domain combinations.

          • turdas@suppo.fi
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            1
            ·
            3 days ago

            The only task that didn’t degrade across most models was Python.

            Yeah, after 20 cycles of unsupervised iteration on the task. Gemini 3.1 Pro doing as well as it did under that experiment setup is quite remarkable actually.

            The paper does not show what you are arguing.