Use AI for reviewing code especially when the diff is huge

The article argues that reviewers should leverage AI to handle large code diffs rather than manually reviewing every line, since LLMs are now proficient at catching coding issues and nitpicking errors. Instead, human reviewers should focus on contributing their unique "out of distribution" knowledge—such as architectural decisions, codebase conventions, and design patterns—that neither the code author nor AI would know. This approach is most effective in codebases where individual lines aren't critical, though may not apply to domains like embedded systems where code scrutiny is essential.