No LLM Code in Dependencies

A git-annex developer spent 100 hours auditing dependencies to eliminate LLM-generated code, discovering several instances of poorly-written AI code including reverted changes and incoherent commits that raised quality and copyright concerns. The developer expresses frustration that reviewing entire dependency trees for AI-generated code has become necessary and questions whether open-source communities can adequately address this problem. They warn against casual use of LLMs in collaborative projects, noting that poor AI-generated contributions can damage both code quality and community relationships.