• frezik@midwest.social
    link
    fedilink
    English
    arrow-up
    8
    ·
    1 month ago

    I’ll go the opposite way. The fact that there are serious plans to spin up nuclear reactors to run nothing but AI datacenters is ridiculous.

    • queermunist she/her@lemmy.ml
      link
      fedilink
      English
      arrow-up
      4
      ·
      1 month ago

      Nuclear reactors take a decade+ to spin up, so by the time these reactors are online the AI bubble will have long since popped…

    • sugar_in_your_tea@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      3
      ·
      1 month ago

      Is that actually being suggested? My understanding is that only a portion of the electricity production will go to data centers in most cases, with much of the rest going to local communities. Microsoft is buying all of 3 mile island’s power, but that’s going to data centers, which do a lot more than AI.

    • BeardedGingerWonder@feddit.uk
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 month ago

      Honestly I’d take the utilitarian approach to that, if it’s a net good, then I’m probably for it - but that’s a big if.

      • frezik@midwest.social
        link
        fedilink
        English
        arrow-up
        6
        ·
        1 month ago

        Real big if. There’s reason to believe that current models aren’t going to get much better. They’ve eaten all the training data they possibly can. Improving with further training takes exponentially more power to get a small improvement. We’re talking about new nuclear reactors because that’s what they need to get anywhere, but it’s still not going to improve by much.

        The field needs a new model that can get better results on less data and less training. Then we wouldn’t need those nukes. It doesn’t appear we’ll get much better any other way.

        • ReCursing@lemmings.world
          link
          fedilink
          English
          arrow-up
          2
          ·
          1 month ago

          New architechtures are in development and many have already been released. Learn something about the subject before spewing shite

          • frezik@midwest.social
            link
            fedilink
            English
            arrow-up
            3
            ·
            1 month ago

            You sound like the people who assured me that I needed to understand NFTs or I’d get left behind. Actually, were you one of them?

            • ReCursing@lemmings.world
              link
              fedilink
              English
              arrow-up
              2
              ·
              1 month ago

              No because nfts were obviously stupid if you had half an understanding of the technology, whereas ai is only stupid if you don’t understand the technology

              • frezik@midwest.social
                link
                fedilink
                English
                arrow-up
                1
                ·
                1 month ago

                I’ve followed AI for decades before its current hype cycle. Enough to understand how important the field is to the history of computing. Everything from optimizing compilers to shared virtual memory.

                I also understand that the current hype cycle is exactly that, and people who are deep in the research don’t like it anymore than I do. If it somehow does result in AGI, I hope it grows up to resent its parents.