Protecting our FLOSS commons from LLMs

Codeberg e.V. members voted to approve two motions protecting the platform's free/open-source software commons from Large Language Models: committing to never use user or project data to train LLMs, and prohibiting projects designed specifically to feed LLM training. The organization argues that LLMs pose fundamental incompatibility with responsible open-source development while externalizing massive infrastructure and environmental costs onto society.