Hi Lyle!

Am Fr., 24. Juli 2026 um 18:01 Uhr schrieb Lyle Kopnicky <lists@qseep.net>:

I have been using Claude Code (Opus) to assist with code reviews lately. I find that it is of great assistance in doing code reviews, but in no way can be left to stand on its own as a reviewer. It will quickly find several issues. I then examine each one. As needed, I ask Claude for explanations and justifications, and consult the code myself, to make sure I understand. I find that most of the issues it finds are legitimate, but some are not actually issues - it may have misunderstood something about the code or the intent. I will then continue to review the code myself, and often will find major issues that Claude missed.

Although I think Claude is of great assistance in getting up to speed on the merge request, and with questions on the code, it is no magic bullet. There is no way to ensure that it doesn’t generate false positives, or false negatives. Thus I think that generating reviews completely automatically with LLMs is likely to cause frustration to the MR author. I’m fine with someone doing an LLM-assisted review, but agree with the Simon PJ’s proposed notion of someone needing to take responsibility for it.

I agree that an LLM-generated shouldn't replace a human review. My
current idea is to have optional, additional LLM-generated reviews
that I or other contributors could provide. However if the goal is to
reduce the load on human reviewers, I don't think that it should the
task of the "LLM-review-providers" to check that the review is
actually "correct". In fact, I don't think anyone should be obligated
to act on such a review. My hope is that these reviews will still
provide some value to the project by catching issues before a human
reviewer needs to spend time on them. If it turns out that the rate of
false positives and negatives is too high, we can try to improve how
these reviews are generated or end the experiment.