A security breach at the maker of ChatGPT last year revealed internal discussions among researchers and other employees, but not the code behind OpenAI’s systems.
    • Lost_My_Mind@lemmy.worldEnglish
      161·
      5 months ago

      Nothing but upvotes for you.

      I’m imagining you as this Nordic Viking, who’s 7’2, and made exclusively of muscles, and the closest thing to a shirt you wear are those straps across the chest to keep your sword on hour back. Like He-Man has.

      And long Fabio hair. With bloodlust in your eyes…

      …but you work in an office. With people you despise, but there are no more viking raids. So you have to put up with Kathy, telling you another boring story about her cats…or was it her kids? Oh, actually it might have been her dying grandmother. Whatever, you weren’t listening anyways. Too busy imagining using your sword to cut her spine out, and twist it in a knott.

      But alas. You’re just here to make photocopies of…wait, is this a blank piece of paper??? Have you been making 500 copies of NOTHING for the past 4 minutes??? But wait…then who has the picture of your butt???

      My god, is Kathy STILL talking???

  • MonkderDritte@feddit.deEnglish
    254·
    5 months ago

    And what would happen then? China would get a mostly working LLM? Such a fearmongering title.

    • UnderpantsWeevil@lemmy.worldEnglish
      41·
      5 months ago

      A big part of the rush to sink billions into AI revolves around this fervent instance that China is creating an Evil AI and if they build theirs first, it will force everyone into eternal Super AI Communism.

      “China is going to steal our AI so we have to fight back” is just the latest riff on this aging song.

    • doodledup@lemmy.worldEnglish
      3·
      5 months ago

      It would work exactly the same. What makes you think it would be inferior? It’s not like a machine that has a manufacturing quality. This is software. And the data for training they already have.

      • MonkderDritte@feddit.deEnglish
        10·
        5 months ago

        What inferior? My “mostly working” was because they are often pushed to users as more than they are capable of. I should have formulated that better.

  • jaybone@lemmy.worldEnglish
    6·
    5 months ago

    I assumed with a name like OpenAI, a lot of their code / IP was public? But never really looked into it, so maybe not. Either way China is always stealing everyone’s IP, so why would this be any different?

    Anyway, what kind of threats are they predicting this will create now?

    • GamingChairModel@lemmy.worldEnglish
      2·
      5 months ago

      Their algorithms are mostly public. Their training data/procedure, and the trained model’s parameter weights, are not.

  • uebquauntbez@lemmy.worldEnglish
    61·
    5 months ago

    Don’t fear the foreign AI developers.

    They’re just like our AI developers.

    Oh, wait! Holy $hit!

  • yamanii@lemmy.worldEnglish
    2·
    5 months ago

    Oh no the horrors of lying with confidence! Should’ve been open already anyway, piece of shit Altman.