• NoiseColor @lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    23
    ·
    20 hours ago

    That’s a stupid take. We’ve always had algorithms that show some results over others. This is just a different one. There is no it, there is no being or whatever behind it.

      • NoiseColor @lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        9
        ·
        16 hours ago

        Perhaps. The take is still stupid.

        We still have the library to browse any way possible.

    • fishy@lemmy.today
      link
      fedilink
      English
      arrow-up
      8
      ·
      17 hours ago

      As the AI confidently gives me the wrong information, which it stole from a forum, talking about an only vaguely related process. It’s one thing to present info as “here, hope this helps” vs “here’s what you asked for, this is the relevant sentence highlighted.”

      If I didn’t know it was wrong, it could’ve set me down a path that wasted hours of my time. It’s happened to colleagues, and they weren’t wise enough to know better.

      • NoiseColor @lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        5
        ·
        16 hours ago

        I’m not debating the quality of LLM responses, or the ethics of using various pubic and private data for training it. I’m claiming:

        • it’s not alive and has no personal motivation
        • we still have the library, nobody took it away, we can still browse it
        • fishy@lemmy.today
          link
          fedilink
          English
          arrow-up
          4
          ·
          12 hours ago

          There absolutely is a being behind the AI. There’s bias built in, so while it doesn’t have personal motivations it is guided by the motivation of big tech. They’ve also been shown to lie, hallucinate and say essentially whatever in order to please the user.

          The library is fantastic, a tool that holds almost immeasurable value. It’s however nowhere near as fast or accessible as the Internet. While AI hasn’t killed the Internet, every bit it’s touched is worse for it.

          • NoiseColor @lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            arrow-down
            2
            ·
            5 hours ago

            You are grasping for straws, just to contradict me for some reason. None of that was implied in the op. You can go in the bias and how and why is the, but the fact of them matter is that is only a reflection of our real world. So using that as an argument is the same as saying ai sucks because we suck. It’s true, but a pointless argument.

            Btw: lies, hallucinations and “saying essentially anything…” Is the same thing. It’s how they work. Nobody says it’s ideal and literally everyone knows about it. It’s something we deal with when we use it. There are ways to mitigate it and there have been progress there, but the tech works in such a way, that it might always be present. It’s not machinations of a little evil elf. It’s a technology.

            Internet is the library. It’s there. Use it.

            Has it gotten worse? It got worse before ai. It got much worse before it. Unfortunately it’s is the responsibility of the user to find the good stuff.

    • deur@feddit.nl
      link
      fedilink
      English
      arrow-up
      15
      arrow-down
      3
      ·
      19 hours ago

      Okay dumbass, you’re very brilliant to miss the fact one algorithm curates human content and the other statistically models text in a way that appears human enough.