• ShinkanTrain@lemmy.ml
    link
    fedilink
    English
    arrow-up
    216
    arrow-down
    3
    ·
    edit-2
    5 days ago

    I did it so you don’t have to

    By the way, it took 5 tries because in 3 of them it made up a story about a different journalist, and in one of them it listed who it thinks would eat the most hotdogs.

    Why is the entire economy riding on this thing?

    • bampop@lemmy.world
      link
      fedilink
      arrow-up
      26
      arrow-down
      1
      ·
      edit-2
      4 days ago

      Is it my imagination or are LLMs actually getting less reliable as time goes on? I mean, they were never super reliable but it seems to me like the % of garbage is on the increase. I guess that’s a combination of people figuring out how to game/troll the system, and AI companies trying to monetize their output. A perfect storm of shit.

      • Nikelui@lemmy.world
        link
        fedilink
        arrow-up
        36
        ·
        4 days ago

        It was inevitable, when you need to train GPT on the entirety of the internet and the internet is becoming more and more AI hallucinations.

        • Zerush@lemmy.ml
          link
          fedilink
          arrow-up
          6
          ·
          4 days ago

          That is the point. Training an LLM on the entire internet never will be reliable, apart of the huge energy waste, not the same as training an LLM on specific tasks in science, medicine, biology, etc., with this they can turn in very usefull tools, as shown, presenting results in hours or minutes in investigations which in traditional way would have least years. AI algorrithm are very efficient in specific tasks, since the first chess computers which roasted even world champions.

          • Grandwolf319@sh.itjust.works
            link
            fedilink
            arrow-up
            1
            ·
            4 days ago

            Those MLs don’t automate anything though, they increase output but also increase cost. The AI bubble is about reducing costs by reducing head count.

      • Joeffect@lemmy.world
        link
        fedilink
        arrow-up
        7
        ·
        4 days ago

        garbage in ( text generated by other ai) garbage out ( less realiable text to train on)

        LLM are not smart they have no brain it is a prediction engine: I could see a LLM being used in a real AI to form sentences or something but I’m sure there are better ways to do it, I mean a human brain does not hold all the knowledge of humanity to be able to process thoughts and ideas… it’s a little overkill…

    • pyre@lemmy.world
      link
      fedilink
      arrow-up
      3
      ·
      3 days ago

      because money is made up and also fuck you that’s why now pay taxes so billionaires can get more money

    • HrabiaVulpes@lemmy.world
      link
      fedilink
      arrow-up
      2
      ·
      3 days ago

      There are many things around the world that are advertised as much better than they are in reality. Like life in the USA, Russian military might, AI employees or whatever the hell is dubai chocolate.

    • GreenBeanMachine@lemmy.world
      link
      fedilink
      arrow-up
      6
      ·
      edit-2
      4 days ago

      “greatest hula-hooping traffic cops” works as explained in the google ai search and “Officer Maria “The Spinner” Rodriguez” is the GOAT.