I have a much shorter and optimistic version, which is that these models break a theory of mind that underpins a lot of bad stuff.

Basic idea is, since we have shown that we can simulate all consistent artifacts of language in an LLM, it's a proof that language can and (therefore does) exist on a substrate. That substrate for us is the real, and what kids who encounter AI now will understand intuitively is that the artifacts of language and the real are necessarily separate things. It's the end of subjectivity and everything stacked on top of it, thank literal God. I think people can be forgiven for having been decieved by some clever linguistic hacks that unmoored them from reality, and now we can look at the uncanny artifacts of the logic of ideas over language using sound models for what they are.

Welcome to the desert of the real!