Frankly, most AI generated code is often easier to review, thanks to a combination of standardized practices (LLMs regress to the mean by design) and a somewhat overly enthusiastic approach to commenting and segmented layouts.