Sure but everything is semantics.

LLMs have no internal secret model, they are the model. And the model is of how different lexemes relate to each other in the source material the model was built from.

Some might choose to call that the world.

If you believe your internal model of the world is no different from a statistical model of the words you have seen, then by all means do that. But I believe a lot of humans see their view of the world differently.

I very much believe my cat’s model of the world has barely anything at all to do with language.

This path to AGI through LLM is nothing but religious dogma some Silicon Valley rich types believe.

I mean by definition CATS CANNOT TALK. Their vocabulary is probably on the order of 5 different types of meows