That's a really weird argument. Copyright is a legal system created by the Constitution and statutes and administrative rules. It cares about whether you are "copying", and it cares about whether you're creating things that compete with the works of the original authors. It doesn't care about potential output spaces.

In this context, I don't see a principled difference between the model weights and really good compression. If I send you a gzipped copy of the latest bestseller book it's still copyright infringement. And it would still be infringement if I shipped it inside a software program that can _also_ reshuffle the words in a bajillion different ways, if there's a "copy" of the original work in there.