I dunno, it seems you have figured it out too, probably before LLMs?

I'd say all speakers of all languages have figured it out and your statement is quite confusing, at least to me.

Yes of course we’ve implicitly learned those rules, but we have not been able to articulate them fully ala Chomsky.

Somehow, LLMs have those rules stored within a finite set of weights.

https://slator.com/how-large-language-models-prove-chomsky-w...

We all make grammar mistakes but I’ve yet to see the main LLMs make any.