AI-generated contributions and generated workload on maintainers
Thomas Hochstein
inn-workers at ml.th-h.de
Mon Nov 9 22:44:22 UTC 2026
Julien ÉLIE wrote:
> There are several AI agents from OpenAI and Anthropic running on INN
> codebase, submitting bug fixes and security reviews.
I don't think it is acceptable to let AI agents submit PRs or issues
unsupervised ...
> I do not want to spend my spare time and my evenings reading
> and discussing with an AI on PRs and the issue tracker. It responds
> within minutes with questions, decisions to make, proposals, etc.
> It even said yesterday "I'm not a C expert myself. Your review remains
> the real check." Well, it knows C far better than I and does not take
> any responsibility in its work...
> It announced tons of upcoming patches, concluding with "@Julien-Elie be
> prepared ?"
... and force the maintainer to "chat" with the agent via comments.
The right way - IMNSHO - is to check (!) and submit issues and PRs
through a human.
> I say "most" because I spotted weird things and inaccuracies in some
> submitted patches. And a few minutes after highlighting these, the
> responses came:
> "You are right, the reproduction steps are wrong: newsgroups: and post:
> can't be combined."
> "my patch added the field name to its new error message but indeed
> missed this one"
> "I mis-copied it"
> "I withdraw the sentence. I haven't substantiated it."
I find that unacceptable.
> So, what to do next?
>
> A proposal to discuss: couldn't we give push/commit access to the GitHub
> users running AI agents? This way, they could automatically work and
> improve the code as well as open and close PRs and tickets.
Your examples show that - at least - not all submissions would improve
the code.
I wouldn't give agents commit access, but rather block them. AI agents
can be a helpful tool, but I think the onus to check their output,
submit it and communicate with the maintainer should be on the
contributor.
Regards,
-thh
More information about the inn-workers
mailing list