It is just funny to think about—LLMs are sometimes viewed big piles of linear algebra, it would not be that surprising to hear that somebody had worked out that one model was somehow a subset of another (or something along those lines) and then claim some theoretical superiority.