Code reviews are also a fairly long feedback loop. There’s PRs with thousands of lines of code changed. And PRs with a few lines in one file. I’d question whether we’re doing code reviews right in the first place. Leveling it up to first principles: why do we do code reviews and is there a better way?