• sleepy@reddthat.com
    51·
    1 year ago

    Isn’t that a part of the ai marketing though? That whole “this thing could destroy us” stuff?

    • Square Singer@feddit.de
      61·
      1 year ago

      Totally is. Because it makes the AI look and feel much better than the smoke-and-mirrors it actually is.

      • sleepy@reddthat.com
        9·
        1 year ago

        We thought we were getting Skynet but, instead we got Super Clippy and I Can’t Believe It’s Not Art Theft

      • Comment105@lemm.ee
        3·
        1 year ago

        Do you see any reason to think enough iterations of random nodes in a large enough network could result in emergent conscious intelligence?

        Or are you more of a spiritualist than a materialist when it comes to the mind?

        • Square Singer@feddit.de
          1·
          1 year ago

          I can’t say anything about the spiritualist/materialist thing, but there are two things that are clear:

          First: Same as you won’t be able to ever get a Shakespeare work by randomly stringing letters together in any reasonable time frame, you won’t be able to do the same with conciousnes. If it’s possible, the number of incorrect permutations are so massive, that just random trying will not ever be enough in any realistic amount of time.

          Second: Transformer networks and all other generative AI concepts we have today aren’t even trying to create a conciousnes. They are not the path to general AI.