I wrote a script that counts the work AI-written commits cost you afterwards. It says 13 hours on my own repo.
I build software — games, plugins, client work — and most of the code is written by a model and reviewed by me. Last night I wrote a read-only script that goes through git history and asks one question of every commit: did anyone promise anything about this, and did anyone own it? Concretely: does the message claim a test, an issue, a verification, a follow-up? And if the commit declares itself…
A software developer created a read-only script to calculate the time spent on reviewing and testing AI-generated commits in their own repository. The script analyzes the git history and identifies commits that claim to have specific tasks or responsibilities attached to them. In the case of the developer's repository, 9 out of 13 commits were identified as being written by AI, with 8 of those commits lacking any associated promises.
This led to 13 hours of potential work going unaccounted for. The developer emphasized that the tool is designed to be simple, requiring only a single file with JSON data to track the hours spent on each commit. They offered to perform the scan on other repositories at no cost, providing the number of hours as a transparent ledger for better accountability. The source code is available on GitHub for anyone to review before trusting the results.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
