What happens to a review culture when the code under review was written by no one? What happens to the slow, generous work of onboarding a contributor when the fast answer is always one prompt away? Who gets credit, who gets blamed, and who quietly stops showing up? What does it mean for a community's memory and its politics when the machine that generates the code was trained on that same community's freely given labour, and gives nothing back?

This is not a talk about whether you should use these tools. It is a talk about being honest with ourselves about what we are trading, and who decides. I want to make the case that if we keep measuring only output, we will optimize away the very thing that made open source worth defending in the first place. And then I want to ask what it would look like to choose differently.