I am no longer letting Claude Code add itself as Co-author in my commits

For the past year or more, the author explicitly added Claude Code as a co-author to every commit he used it for, and encouraged others to do the same, so reviewers would know an LLM helped build the code. Now he argues that using LLMs is so obvious and expected in work contexts that the attribution is no longer necessary, and he compares it to a carpenter not naming the saw they used.

His core reasoning is about responsibility: if work is under his name, he assumes full responsibility for it. Just as he would never have added Google or StackOverflow as co-authors (though he might have linked a helpful post), he will no longer credit Claude Code. He worries that adding an LLM as co-author makes it too easy to shirk blame for mistakes by saying 'the AI did it.' Bugs existed before and after LLMs, so developers must still verify their work before shipping.

He does see nuance: LLM co-authorship or even primary authorship may be acceptable when evaluating different models for comparison, or when an LLM creates a PR and a human acts as reviewer. For transparency, he suggests alternative approaches like committing with an AI-specific account or adding a disclaimer in the project's README that the code is AI-generated. But for serious work, he believes humans should take full ownership.

In an appendix, he notes that he woke up to a Hacker News post that reminded him of his own Bluesky post from earlier in the week, and that his stance has flipped: he used to think LLMs should be co-authors, but now believes it encourages shirking responsibility.