I can't tell if the title is a joke or not, the article seems to give some very good reasons why you shouldn't use an LLM to review pull requests