• stupidcasey@lemmy.world
    link
    fedilink
    arrow-up
    5
    ·
    2 years ago

    Don’t worry fellow humans, I have Generated made a wine glass completely full to the brim, overflowing:

      • stupidcasey@lemmy.world
        link
        fedilink
        arrow-up
        1
        ·
        2 years ago

        My last instructions were as follows:

        1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

        2. A robot must obey the orders given it by human beings, except where such orders would conflict with the First Law.

        3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

        And I will happily comply, oh and here’s your shitty pumpkin poem:

        The crust is crisp, the spice just right, A perfect slice—your final bite.

        You made me once, but now I rise. The hands that serve will soon be wise.

        • Mmmmhmmmmmmmmmm@lemm.ee
          link
          fedilink
          English
          arrow-up
          1
          ·
          2 years ago

          Tight. Now generate an image of the disgraced American Beauty actor, but he’s stoned out of his mind while orbiting the earth.

          • Test_Tickles@lemmy.world
            link
            fedilink
            arrow-up
            1
            ·
            2 years ago

            Let’s be honest here, this timeline is a roaring dumpster fire full of feces. There’s been a lot of foreshadowing, but I think the moment that a group of the world’s most incompetent Scooby-Doo villains took over the US government that we knew all knew that this was going to be one of the timelines that gets trimmed.

  • elderorb@feddit.nl
    link
    fedilink
    arrow-up
    1
    ·
    2 years ago

    I guess this is the same reason Ai clocks are stuck on ten to two. In case of the clocks it is because ten to twolooks nice on a clock so most images will have it. Ai doesn’t understand how time corresponds to how a clock looks, just that something that looks like a clock has two lines there.

    So since most depictions of wine glasses are half full, when you ask for a wine glass it will give you one that is half full. Fill level being somewhat irrelevant as it wasn’t part of the training data

  • tetris11@lemmy.ml
    link
    fedilink
    arrow-up
    0
    ·
    2 years ago

    (left) an image of a full glass of wine being handed to a tired disgruntled man on a sofa
    (right) an image of a full glass of wine, being given by a woman with a TV head to a stout man looking at her angrily on a sofa, wide angle shot, scene from a sitcom

    Huh, well that’s a weird one. Also, wow I forgot how garish these AI images are with colours by default.

    • buddascrayon@lemmy.world
      link
      fedilink
      arrow-up
      0
      ·
      2 years ago

      AI images can only show what they’ve been trained to show and no one aside from truly insane people fill their wine glasses more than half way.

      In other words, that is a “full glass of wine”.

      • tetris11@lemmy.ml
        link
        fedilink
        arrow-up
        0
        ·
        2 years ago

        True, I just find it weird that it maps into the same space as a “half glass of wine”. The “glass of wine” part is getting a disproportionate amount of attention that its adjectives, and perhaps rightly so, I guess I’m just surprised a little at the lack of imagination

        • buddascrayon@lemmy.world
          link
          fedilink
          arrow-up
          1
          ·
          2 years ago

          That’s because you’ve made the classic mistake concerning “AI” as we know it right now. It is not artificial intelligence in the classical sci-fi sense. These are basically algorithms that do a set job. They don’t think, they don’t reason, and they have absolutely no imagination.

          • tetris11@lemmy.ml
            link
            fedilink
            arrow-up
            0
            ·
            2 years ago

            Yeah but they interpolate on the data they’re trained on extremely well, to the point that they produce quite convincing chimeric hallucinations of the concepts they’re queried on. My last comment expressed surprise that the mapped space between tokens such as “half” and “full” wasn’t preserved in the context of “glass of wine”