I don't see why a strong model wouldn't significantly outperform any human in this sort of low level optimization work. We don't hand optimize assembly either.

It's about the competitive edge. If a public model can do it, everyone has access to it so nobody has advantage.

If you want LLMs to be your advantage you can train your own, that's completely valid.

Let's say you want to have a company that runs inference 5% faster, if everyone can do it your business model is worthless.