Clean AI code can still take longer to review
In one firsthand work example, a junior developer made a messy CSV parsing utility and it was approved in five minutes. The function was too long and the naming could have been better, but the reviewer already understood the work because the developer had discussed the vendor format at lunch and struggled with type errors during standup. Later the same day, duced a similar-sized utility for another feature.
That version looked cleaner, had better structure, and included the junior code did not have. It still took more than 20 minutes to review, including checks with , Bugbot, and Claude, two reads of the diff, and manual tests for that normally would not have been tested. The hard part was that the AI code arrived fully formed, with no visible history of what it tried, rejected, or misunderstood.
The reviewer could not tell whether the extra caution was sensible risk control or bias against .
Key points
- The junior developer's rough code was approved quickly because the reviewer had seen the work unfold.
- duced cleaner code with better , but it took much longer to review.
- The main issue was missing history: the reviewer could not see what the AI tried or rejected.
- The AI code was checked with , Bugbot, Claude, repeated diff review, and manual edge case tests.
- AI coding workflows should include reasoning notes and test notes, not only generated code.