• 0 Posts
  • 9 Comments
Joined 2 years ago
cake
Cake day: April 28th, 2024

help-circle
  • algorithm that’s been modelled after the real world structure and behaviour of neurons and how they process signals

    Except the Neural Net model doesn’t actually reproduce everything real, living neurons do. A mathematician in the 70s said, “hey what if this is how brains work?” He didn’t actually study brains, he just put forward a model. It’s a useful model. But it’s also an extreme misrepresentation to say it approximates actual neurons.




  • I think what people are struggling to articulate is that, the way AI gets thrown around now, it’s basically being used as a replacement for the word “algorithm”.

    It’s obfuscating the truth that this is all (relatively) comprehensible mathematics. Even the black box stuff. Just because the programmer doesn’t know each step the end program takes, doesn’t mean they don’t know the principals behind how it was made, or didn’t make deliberate choices to shape the outcome.

    There’s some very neat mathematics, yes, and an utterly staggering amount of data and hardware. But at the end of the day its still just an (large) algorithm. Calling it AI is dubious at best, and con-artistry at worst.



  • Langton’s ant can procedurally generate things, if you set it up right. Would you call that AI?

    As for enemies in gaming, it got called that because game makers wanted to give the appearance of intelligence in enemy encounters. Aspirationally cribbing a word from sci-fi. It could just as accurately have been called “puppet behavior”… more accurately, really.

    The point is “AI” is not a useful word. A bunch of different disciplines across computing all use it to describe different things, each trying to cash in on the cultural associations of a term that comes from fiction.