One of the strengths of the too-many-approvals approach was that it made sure developers were still looking closely at every little change.

More likely it made developers immune to the approval button, instead just clicking approve blindly.

I would say: Learn to use LLMs as a tool rather than a crutch.

I’ve never seen it recommend a solution using regex. And I’ve had it provide a lot of useful code. Perhaps you need to look into prompt engineering training?

Try reading the question again, this time answer without hating all things LLM. OP asked if AI could be used for initial peer review, to which the answer is a big yes.

How a lot of software development ends up in real life too.