• joaomarrom [he/him, comrade/them]@hexbear.net
    link
    fedilink
    English
    arrow-up
    13
    ·
    edit-2
    6 days ago

    that’s actually kinda cute, it’s such a quaint prediction, coming straight from early 2023, when lots of people were impressed and had some degree of hope that AI would be a transformative, positive tool, rather than the nightmare factory churning out piss-colored garbage that it currently is

    • neo [he/him]@hexbear.net
      link
      fedilink
      English
      arrow-up
      9
      ·
      6 days ago

      Oh, don’t worry. Many CTOs at mid-sized companies throughout the country are getting in on LLMs right now. OpenAI, Microsoft, Anthropic, and Alphabet’s sales reps are all laughing their way to the bank.

    • woodenghost [comrade/them]@hexbear.net
      link
      fedilink
      English
      arrow-up
      6
      ·
      edit-2
      6 days ago

      btw do we know yet, why is all yellow? Just training bias from human made art (seen this claim) or actually reinforcing itself by more and more ai art in the training data? Probably the second one, right? But is there proof?

      • joaomarrom [he/him, comrade/them]@hexbear.net
        link
        fedilink
        English
        arrow-up
        4
        ·
        6 days ago

        I don’t know, but the theory I find to be most likely is that it’s due to inbred training data, since it started right after the whole Ghibli thing. Habsburg ass AI. Most Post-Ghibli AI slop acquired that ubiquitous piss filter look, it’s hilarious

    • peeonyou [he/him]@hexbear.net
      link
      fedilink
      English
      arrow-up
      4
      ·
      6 days ago

      as is the case with absolutely every new technology that comes about… it always holds so much promise and then a bunch of fuckwads figure out how to turn it into a dystopian tool from hell to make money from it