Codeberg bans LLM-generated contributions, FLOSS pushes back
Codeberg, the nonprofit Git hosting platform, passed a member vote to ban LLM-generated code contributions. The thread was largely supportive, framing it as a community defending the quality of its commons. Parallels were drawn to their earlier crypto ban: both were democratically voted in by active members, both represent a values stance rather than a technical one.
The 'Making' thread ran alongside this one and made a related philosophical point: there's a meaningful difference between making something and asking to have something made, even if the output is identical. HN commenters in that thread worried specifically about AI-generated Show HN submissions polluting what used to be a space for human ingenuity.
The counterpoint is that enforcement is nearly impossible. You can't reliably detect LLM-generated code, and banning it may just create a rule that's selectively applied or ignored. But the Codeberg approach at least establishes a norm, and norms shape culture even when enforcement is imperfect.
So what?
Open source communities are starting to treat LLM-generated contributions as a quality and trust problem, not a productivity feature. If you're building developer tools or contributing to open source as a business strategy, your AI-assisted workflow may need to be transparent or explicitly human-reviewed. The norm is forming now.