Codeberg members voted against the training of Large Language Models (LLMs) using platform data and imposed restrictions on vibe-coded projects during their annual assembly, reflecting a preference for data privacy and content quality over experimental coding practices.