←back to thread

The "confident idiot" problem: Why AI needs hard rules, not vibe checks

(steerlabs.substack.com)
323 points steerlabs | 2 comments | 04 Dec 25 20:48 UTC | HN request time: 0.001s | source
Show context
jqpabc123 ◴[04 Dec 25 21:38 UTC] No.46153440[source]▶
We are trying to fix probability with more probability. That is a losing game.

Thanks for pointing out the elephant in the room with LLMs.

The basic design is non-deterministic. Trying to extract "facts" or "truth" or "accuracy" is an exercise in futility.

replies(17): >>46155764 #>>46191721 #>>46191867 #>>46191871 #>>46191893 #>>46191910 #>>46191973 #>>46191987 #>>46192152 #>>46192471 #>>46192526 #>>46192557 #>>46192939 #>>46193456 #>>46194206 #>>46194503 #>>46194518 #
UniverseHacker ◴[08 Dec 25 14:17 UTC] No.46192471[source]▶
Specifically, they are capable of inductive logic but not deductive logic. In practice, this may not be a serious limitation, if they get good enough at induction to still almost always get the right answer.
replies(1): >>46192556 #
1. psychoslave ◴[08 Dec 25 14:24 UTC] No.46192556[source]▶
What about abduction though?
replies(1): >>46195484 #
2. UniverseHacker ◴[08 Dec 25 17:59 UTC] No.46195484[source]▶
You’ll have to wait for the FOOM “Fast Onset of Overwhelming Mastery” for that I’m afraid.