Ask HN: How are you reviewing code at work these days?

[curiousgal]

With everyone using AI coding assistants, I am curious to know what policies people have implemented to deal with the influx of AI generated code? I am maintaining an internal library at work and as more people on the team gain access to Claude Code, I am wondering whether we should be highlighting/labelling AI code or not.

[azihsoyn]

We review irreversible decisions first (e.g. schema changes). Once there's agreement on that, tools like CodeRabbit or Claude review the code, and humans focus mostly on whether the behavior is correct.

We also try to build guardrails so bad code is hard to write in the first place: CI checks, custom lint rules, repo guidelines (e.g. CLAUDE.md), etc.