2026-10-05 15:51 UTC
Replies (3)
-
@jasongorman@mstdn.business 2026-10-05 16:21
Actual reasoning doesn't suffer from probability collapse. A chess player can look at the positions of pieces on the board and reason about the best next move. It doesn't matter what sequence of moves left the board in that state. To an LLM, it matters. It's actually everything that matters.
-
@tesmaia@mastodon.gamedev.place 2026-10-05 20:26
@jasongorman@mstdn.business The entire concept of "reasoning" models was built on top of the discovery that by prompting the damn things to explain a "chain of thought", it gets some better results. What they're doing is fishing for context that then shift the probability distributions around. It's not deeper than "oh hey this works, lets do more of it and build the models around it".
-
@CountHoldem@mastodon.social 2026-10-08 14:11
@jasongorman@mstdn.business Perhaps "not" and "are", are a bit too 'removed' from the evolving equation, at this power level. This means: they've already 'got you'.