Codeberg recently updated its terms of service to exclude projects primarily written with generative AI. While Codeberg, as a member-driven association, has the right to make such decisions democratically, this move raises concerns about inclusivity and the future of open source development.

The policy's vagueness, particularly the term "mostly," makes enforcement difficult and open to interpretation. The author suggests that a clearer stance, either prohibiting all LLM involvement or focusing on specific issues like spam and resource abuse, would be more effective than the current middle ground which delegates policy to moderators and community norms.

The division within the open source community over LLMs is seen as a shame, despite valid concerns about copyright, labor, energy, and contribution management. The author believes the community needs to find ways to engage with these tools rather than splitting into opposing camps. LLMs, if used well, could empower developers and offer a true alternative to large corporations.

As a European project, Codeberg's decision is particularly noted. The author expresses a desire for Codeberg to be a broad, dependable alternative to GitHub, embracing the open source software of the future, rather than solely adhering to current community approval standards. The decision is seen as a missed opportunity for a more forward-looking approach.