Pull Requests has to be the most misguided engineering practice I’ve seen in my career. It produces the worst engineering process with horrible incentives.

PR reviews are a shift-right design & review process. It incentivize post-hoc discussion about the change.

Reviewers are incentivized to find flaws and not contribute to improving the change. It’s an ocean of stop energy and nagging

As everyone has to wait for a PR to be reviewed, we all have a dozen PRs in transit at multiple stages of interest and completion. We unlearned that more WiP is bad.

PRs are a perfect fit for explaining the sunken cost fallacy. Since you need to advocate for your change, it’s hard to give up and try something else.

PRs are a terrible teaching tool. The final result of your work tells me nothing of how you reached that point.

PRs are a misappropriation of how FOSS communities used to operate 2 decades ago.

PRs are great in an async and effort-symmetric setup. Most teams working full time from the same time zone (and previously same office) don’t match any of those criterias

To wrap up, PRs are a cargo cult practice that most teams are neither critical or aware of alternatives.