Codeberg Divides
Codeberg, a non-profit Git hosting platform, has amended its terms to ban projects that mostly consist of AI‑generated code and to exclude cryptocurrency-related projects that could harm its reputation. Commenters are split between seeing this as a principled stand to protect limited resources, software freedom, and community values, and viewing it as an arbitrary, hard-to-enforce policy that undermines trust and predictability for open source maintainers. The broader debate centers on whether infrastructure platforms should remain neutral about tools and project types, or deliberately curate against “AI slop” and crypto even at the cost of becoming a smaller, more ideologically defined niche.
Scope of Codeberg’s New Rules
- Codeberg updated its ToS to:
- Ban projects “mostly consisting” of generative-AI-written code.
- Class cryptocurrency-related projects as content that can harm Codeberg’s reputation.
- Reserve broad discretion to suspend accounts for activities that harm Codeberg, even if not explicitly listed.
- Some commenters clarify that crypto isn’t categorically banned in all forms, but the official wording and enforcement remain somewhat unclear.
Concerns: Arbitrariness, Trust, and “Slippery Slope”
- “Mostly AI-written” is seen as vague and hard to objectively measure, especially in active, evolving codebases.
- Fear that value- and vibe-based rules enable biased or politicized enforcement and future rule creep.
- Retroactive application and surprise policy shifts are described as breaking trust; several people report migrating projects away as risk management.
- Worry that more open-ended rules lead to internal politics, gatekeeping, and eventual community fragmentation.
Defenses of Codeberg’s Approach
- Supporters argue Codeberg is a member-driven non-profit, not a neutral utility, and is entitled to pursue a smaller, value-aligned community.
- Open-text rules are seen as a feature: they allow moderators to act on judgment rather than be “loophole-lawyered.”
- Some welcome a curated space with less “AI slop,” crypto projects, or resource-heavy LLM spam on limited infrastructure.
- Comparisons are made to other communities (HN, Wikipedia, EU law) that use high-level principles and discretion rather than exhaustive rulebooks.
AI, Open Source, and Ethics
- One camp sees AI-assisted coding as the inevitable “software of tomorrow,” and views bans as self-marginalizing or “digital Amish.”
- Another camp argues communities can and should say no to tools they find unethical (copyright, labor, environment, “slop”), even if widespread.
- Some propose pragmatic alternatives (quotas, CI/disk limits, focusing on spam/low-value projects) instead of morality-driven bans.
- There is debate over legal interpretations of AI-generated code and GPL compatibility; several commenters say current “tainted code” rhetoric misreads copyright law.
Alternatives and Next Steps
- Many suggest self-hosting (Forgejo, GitLab, Gitea, Radicle) or returning to GitHub.
- Some think this will push more experimentation with federation and true decentralization.
- Overall split: Codeberg is praised for principled curation by some, and avoided as an unstable, ideological host by others.