• EmoDuck@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      8
      ·
      edit-2
      4 months ago

      No need for an AI. I’ve programmed an algorithm that can scan a random image for lost artworks and has a 99.9% success chance:

      def containsLostArt(pictute):
          return false
      
      • Deebster
        link
        fedilink
        English
        arrow-up
        2
        ·
        4 months ago

        What do you mean by this? I’d say you definitely could, although it would need an expert to sort through the candidates to find the ones with following up on.

        AI seems to have hit its limits, but this kind of stuff is what AI is great at.

          • Deebster
            link
            fedilink
            English
            arrow-up
            1
            ·
            4 months ago

            Not at all - you’d have to train a model but then it could be run locally. You could even have something like SETI at Home, and run it on volunteers’ computers.

            I suppose some social media charges for access and it might not be a good idea to scrape for something like this (or at least admit it).

            • Kairos@lemmy.today
              link
              fedilink
              English
              arrow-up
              1
              arrow-down
              1
              ·
              4 months ago

              Yeah lemme just download many petabytes a month of data on my home Internet connection

              • Deebster
                link
                fedilink
                English
                arrow-up
                1
                ·
                4 months ago

                In a world where you can stream 4k video, you think a few images are going to be a problem?

                  • Deebster
                    link
                    fedilink
                    English
                    arrow-up
                    1
                    ·
                    4 months ago

                    You’re just making up big numbers and ignoring what I’m saying.

                    Either it’s done centrally, in which case it’s feasible if they have funding, or it’s done SETI-style and users share the load (and could have data limits, etc).