No? The top of human engineers are much better than any model would be, so AI models really aren't a big advantage when you're trying to develop anything that is SOTA.
Not every problem is best addressed by a top engineer.
Plenty of the tasks that keep a company running can benefit from good-enough (and better than the competition).
You can't fathom how 'top human engineers' could take advantage of exclusive access to a frontier model at ultrafast speeds and limitless token budgets to conduct research?