Hacker News Reader

Comment by kqr | Hacker News Reader

> probabilistic token generators aren’t intelligence

Maybe this has been extensively discussed before, but since I've lived under a rock: which parts of intelligence do you think are not representable as conditional probability distributions?

View on HN

> which parts of intelligence do you think are not representable as conditional probability distributions

Maybe I'm wrong here but a lot of our brilliance comes from acting against the statistical consensus. What I mean is, Nicolaus Copernicus probably consumed a lot of knowledge on how the Earth is the center of the universe etc. and probably nothing contradicting that notion. Can a LLM do that ?

And how many people actively do that? It's very rare we experience brilliance and often we stumble upon it by accident. Irrational behavior, coincidence or perhaps they were dropped on their heads when they were young.

Copernicus was an exception, not the rule. Would you say everyone else who lived at the time was not 'really' intelligent?

Yes but this was just to make a blatant example. The questions still stands. If you feed a LLM certain kind of data is it possible it strays from it completely - like we sometimes do in cases big and small when we figure out how to do something a bit better by not following the convention.

That's an illogical counterargument. The absence of published research output does not imply the absence of intelligent brain patterns. What if someone was intelligent but just wasn't interested in astronomy?

It could be "probability of token being useful" rather than "probability of token coming next in training data"!

Reply on news.ycombinator.com