AI BriefingCodeberg moves to protect the FLOSS commons from LLMs
Key point
Codeberg has officially adopted policies banning the use of training data for LLMs and restricting vibe-coded projects.
Details
Open-source code hosting platform Codeberg put two LLM-related motions to a vote at its annual general meeting, and both passed.
First resolution: Codeberg officially declared that it does not use user code/data to train LLMs, and will not do so in the future.
Second resolution: 'Vibe-coded' projects are banned under the terms of service. Passed with 358 in favor vs 144 against (about 50% turnout among active members).
Background for the ban:
- AI crawlers were sending massive, meaningless requests to Codeberg servers (issue filter variations, entire Git history, etc.), causing DB load and degraded service quality
- 'Ghost projects'—large-scale LLM-generated code, CI/CD, and release binaries uploaded with no real users—were excessively consuming infrastructure resources funded by donations
- Surging GPU demand has intensified pressure on hardware supply and pricing
Actual enforcement will proceed in stages, with priority given to projects that consume excessive resources rather than to automated detection of LLM-generated code.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
NVIDIA Releases Open-Source LLM Router 'Switchyard'… Up to 30% Cost Savings Compared to Opus