• Negative12DollarBill@techhub.social
    link
    fedilink
    arrow-up
    1
    ·
    6 months ago

    @Immersive_Matthew
    There are examples in the news literally every day of “AI” making horrible mistakes and creating wildly inaccurate fake information.

    It’s not “AI” in any real sense.

    You’re using it to autocomplete code, a task which is highly suited to this kind of “AI” and it’s been trained on billions of lines of code and it still makes lots of mistakes. It’s in no way intelligent, not even 1 per cent of the way to intelligent.

    • Immersive_Matthew
      link
      fedilink
      English
      arrow-up
      1
      ·
      6 months ago

      We will have to disagree as it does show signs of intelligence here and there way above 1%. It is more than an autocomplete as it can take in and understand complex systems even when I thought it did not. Truly shocking at times. It is on par with image creation in that it can create some stunning images, but it often has a couple issue that you have to smooth over in Photoshop/GIMP. The fact that it can make images at all, including shadows and reflections is truly impressive. I find it extremely valuable and could not imagine being without even with its current limitations that can be worked around. Not for everyone though as it is a little hard to use still.