>>200314
>LLMs aren't "intelligent". They're mathematical models that interpret and predict textual language patterns. The apparent knowledge of such is owed to training data that'd be trivial to find were our internet infrastructure not completely rotted to hell, and an infinite patience for bureaucracy. This is something of a pattern for everything LLMs are particularly good at.

It's a curious thing that people still make this argument.

"They just predict the next token"

So what? 3 years ago they could not count the R's in strawberry, right now they can generate pictures and videos indistinguishable from reality, code better than any human, solve millenium price problems in all the hard sciences, and so on and so forth.

And the rate of improvement is exponential.

If superintelligence emerges, will it matter if it's just "a token predictor"?