AI’s voracious need for computing power is threatening to overwhelm energy sources, requiring the industry to change its approach to the technology, according to Arm Holdings Plc Chief Executive Officer Rene Haas.

  • whoreticulture
    link
    fedilink
    English
    arrow-up
    35
    ·
    5 months ago

    I can’t think of a single thing AI does that is worth the amount of energy consumption.

    • Ibuthyr@lemmy.wtf
      link
      fedilink
      English
      arrow-up
      4
      ·
      5 months ago

      The only really useful AI thing is the denoiser in Adobe Lightroom. I can shoot pictures in pitch black darkness with the highest ISO settings. Obviously it is a grainy mess. The denoiser manages to clean that up while retaining all of the details. It’s really fucking great!

      Anything else is just novelty bullshit.

      • whoreticulture
        link
        fedilink
        English
        arrow-up
        4
        ·
        5 months ago

        Sounds useful, but not at all worth the amount of energy being used to produce AI. You could just use that energy to feed/house people who could do the labor of denoising.

        • Ibuthyr@lemmy.wtf
          link
          fedilink
          English
          arrow-up
          3
          ·
          5 months ago

          I know what you mean, but it’s not really possible to manually denoise a picture the way the AI denoiser does. Let alone within 10 seconds. Plus, it’s more of a niche usage. I don’t think it consumes all that much energy.

          Generating shitty images, creating deepfakes, prompting all kinds of bullshit… now that is a waste of energy as it really just makes the world worse. AI generated articles are popping up all over the internet. They aren’t even reviewed anymore. Enshittification of the internet took some gigantic strides since the AI boom.

          • frezik@midwest.social
            link
            fedilink
            English
            arrow-up
            2
            ·
            5 months ago

            Do you know if the model is running locally or some cloud shit? If locally, the actual energy usage may be modest.

            Energy spent training the model initially may have been prohibitive, though.