For a year or more, software engineer Indradhanush Gupta made a point of listing Claude Code as a co-author on every commit he used it for, and encouraged others to do the same. He has now stopped, and he explains why in a short blog post that drew a lively Hacker News thread.
His original reason was transparency: reviewers should know an LLM helped write the code. That reasoning has faded, he writes, now that using LLMs is expected, and at many jobs effectively required. “A carpenter may sign their work. But they don’t add the name of the saw they used to cut wood, do they?”
It’s about responsibility
Gupta’s main argument is about accountability. “If the work is under my name, I assume full responsibility for it,” he writes. He compares it to the pre-LLM era: he never listed Google or Stack Overflow as co-authors, though he would link to a post that helped, and he still will.
Crediting the model, he argues, offers a small escape hatch when something breaks: “oops sorry this happened because it did its thing and it was not careful.” Bugs existed before LLMs and will exist after them, so developers still need to verify the work before they ship it, whatever tool produced it.
He does see a few cases where crediting the model makes sense, such as when you’re comparing models on similar tasks or when an LLM opens the pull request and a human reviews it. For people who want to be upfront about AI-generated work, he suggests a dedicated AI account for those commits, or a clear disclaimer in the project’s README. “But for anything serious, please take full ownership of it.”
The question matters because Claude Code adds a Co-Authored-By trailer to commits by default, so many repositories now carry the attribution whether developers thought about it or not. Teams that agree with Gupta can turn it off in Claude Code’s settings and handle disclosure through their own review and documentation practices instead.
WebProNews is an iEntry Publication