Ask HN: Does your company require human review of AI-generated code?
Mid-size SaaS ~2k employees. We still require human review of every PR. I imagine 99% of the industry is still doing the same but was curious if there has been a meaningful shift in this balance in recent weeks/months.
We don't require it. It's tough to match 1:1, especially cross functions where builders are no longer just devs. You have people building/shipping like never before, so it would be tough for our dev team to oversee salespeople building prospecting agents.
We are having this conversation, where do we keep humans in the loop. I've built a custom PR review bot that makes comments. CTO has some changes that would let subsequent bit reviews resolve comments, but myself and tech lead prefer we keep that part human only for the time being. There is another ticket to make it easy to do the follow up corrections after humans make their comments on bot comments. It could slip into eyes free pretty easily if both are in place
Both my employers do, but it is not genuinely getting human reviewed anyway. I frequently let my Claude approve PRs that I have not even read the title of.
Code is often PR correct but business logic buggy, so how are you finding this is working out for you so far? Had many issues in production?