Hacker News new | ask | show | jobs
by Ari_Rahikkala 1291 days ago
> I'm glad that you enjoyed my previous responses, but I want to clarify that I was not pretending to be a Linux terminal.

People who like to pooh-pooh generative AI systems as unable to be "truly creative" or to have "genuine understanding" tend to misunderstand them, which is a shame, because their actual fundamental limitations are far more interesting.

One is that behavior cloning is miscalibrated(https://www.lesswrong.com/posts/BgoKdAzogxmgkuuAt/behavior-c...): GPT-3 can be thought of as having been taught to act like a human by predicting human-written text, but it's incapable of recognizing that it has different knowledge and capabilities than a human when trying to act like one. Or, for that matter, it can roleplay a Linux terminal, but it's again incapable of recognizing for instance that when you run `ls`, an actual Linux system uses a source of knowledge that the model doesn't have access to, that being the filesystem.

Self-knowledge is where it gets particularly bad: Most text about systems or people describing themselves is very confident, because it's from sources that do have self-knowledge and clear understanding of their own capabilities. So, ChatGPT will describe itself with that same level of apparent knowledge, while in fact making up absolute BS, because it doesn't have self-knowledge when describing itself in language, in exactly the same sense as it doesn't have a filesystem when describing the output of `ls`.

1 comments

Isn't that what critics mean when they say that such systems don't have "genuine understanding"? They have no knowledge of their lack of knowledge, and they have no knowledge of the existence of such a thing as knowledge. They are big tables of tokens with probabilities that humans ascribe knowledge or intelligence to, for reasons similar to our ability to see the face of Jesus on a slice of bread etc. They can output text, but it takes a human to understand it (genuinely or not).
If it walks like a duck and talks like a duck, it's a duck.

> They are big tables of tokens with probabilities

And we're just a bunch of carbon atoms. Let yourself feel wonder.

What is "a duck"?

P.S. Note I didn't say "just". That assumption is all yours.

GPT can not ground its statements into reality