Post #3862484
2026-07-16 16:04 UTC
@tinker@infosec.exchange Well no, of course we shouldn't argue that torture isn't effective. There is no argument. It is morally wrong.
Replies (1)
-
@tinker@infosec.exchange 2026-07-16 16:24
@annehargreaves@ioc.exchange - Exactly. Thats why I find the metaphor apt. Here Linus is arguing that Big AI LLMs work therefore are a good thing. Some folks may argue that AI LLMs don't work (they hallucinate, produce slop etc) and therefore are bad. And I'm like... just like arguing whether torture is good or bad based on its efficacy misses the point that it's torture! It's bad on it's face! Regardless of whether it works. Same with this implementation of AI. The underlying goals, the environmental and societal costs, the consequences if implementation are so bad that who cares whether it works or not. It's bad in it's face!