• mesamuneOP
    link
    fedilink
    English
    144 months ago

    Most “AI” is just a series of if/then statements and/or LLMs that try to predict the next word.

    It was part of my research back in college so its crazy what companies actually say AI is. Im probably preaching to the choir around here but man it grinds my gears.

    • @deranger@sh.itjust.works
      link
      fedilink
      54 months ago

      I hate calling it “AI” (hence the scare quotes) because it’s certainly not what comes to mind when you think of the sci-fi AI many expect. ML also seems to miss the mark, I’m never really sure what to call it. LLM works for chatGPT but what about the music and image generators?

      • mesamuneOP
        link
        fedilink
        English
        44 months ago

        Yeah im not sure what to call suno and some of the others. AI isnt a good way to describe them because theres no real “intelligence” happening.

    • macniel
      link
      fedilink
      34 months ago

      Im probably preaching to the choir around here but man it grinds my gears.

      only partially. But yeah I, as a software developer, agree with your assessment.

    • I acknowledge what the current “AI” is capable of, and it frustrates me all to hell that it is being shoved everywhere except where it would really shine for the common person: video games. Predictive text based RPGs would be dope. The game could have a story that not only takes beats from what you, the player, do, but could go on indefinitely.

      • mesamuneOP
        link
        fedilink
        English
        34 months ago

        Doug Doug did a funny cast where some LLMs were the players in a very simplified DND like game. It was very funny.

        • 🇰 🌀 🇱 🇦 🇳 🇦 🇰 ℹ️
          link
          fedilink
          English
          3
          edit-2
          4 months ago

          Really the best use I’ve personally had with ChatGPT is having it take on the role of DM and 3 other players so I can play Shadowrun and D&D by myself because my group hasn’t defeated the BBG that is “scheduling issues.”

          Even the idiosyncrasies of the LLM being prone to being wrong works in its favor this way, as it makes it seem like a real group of players who bicker about the rules.