Will code review eventually disappear?
My team already auto-approves many AI-written PRs. (So much faster than waiting for humans.) As models get smarter, agents reviewing each other's code will make less and less sense.
The endgame might be just one agent that writes code and proves it works.