this post was submitted on 24 Sep 2026
1115 points (95.9% liked)
Fuck AI
8326 readers
1134 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
As part of looking into the status of things on the self-improving AI front, somebody actually looked at the published papers on amongst other things AI in research and it turns out AI can only ever solve scientific problem for which there are result validators (which is probably why most of its "breakthroughs" are in Mathematics) and is only faster at research tasks than humans for very small research periods (days) because it simply can't handle the structuring of anything. (There's a post from yesterday here for the video though it's stupidly labeled as "they wanted to pay me to say AI was dangerous" or something like that - it's a 1h video but worth the watch if you have Science training)
So don't expect any breakthroughs from AI on anything for which there isn't already a concrete direction for the research to get to the solution directly (i.e. were there are no multiple possible options and it's not multi-step) or the solution can't be directly checked as working algorithmically or at least in computer time (i.e. for which there is no validation mathematical formula or similar) - the state of things now is that AI can't do it.
Maybe one day it will be able to, but then again maybe one day I'll be made the Queen Of Sheeba - it's an unquantifiable possibiliy with no actual evidence supporting that it will actually happen, ever.