I understand the parallel to peer-based code reviews, but not static code analysis. Can someone elaborate?

This is largely based on my experience (which may be incomplete), but static code analysis tools in our CI pipeline don’t add comments to a PR, rather generate a report and will sometimes annotate the affected coded.

Annotations are not tracked historically so any new ci run will regenerate the annotations. I thought this was better for chat gpt reviews since they’re non-deterministic and each CI run could result in different types of feedback.