The Open Source Ouroboros

Open source code is the primary training data for the AI models that undermine open source, creating a feedback loop with no exit within the current framework. GitHub Copilot was trained on all publicly available code on GitHub, the vast majority open source, under licenses that never contemplated this use. The models learned to replicate open source code without license obligations. The loop tightens: in March 2026, GitHub announced it would train AI models on Copilot interaction data by default, feeding not just the original repositories but the prompts developers write and the suggestions they accept back into the next generation of models. curl is the loop made personal: the project’s code trained the models, and the models generated slop bug reports about curl itself until Stenberg shut down the bug bounty program. The more code the community produces, the better the models become at eroding the legal and economic foundations that sustain the community. Producing less code diminishes the ecosystem without removing the threat. The structural trap operates regardless of the legal outcome.