GitHub Copilot Will Soon Train on Your Data

GitHub Copilot is updating its interaction data usage policy to collect and use developer data — including code snippets, cursor context, file names, and feedback — for AI model training by default, unless users opt out. This applies to Free, Pro, and Pro+ users. The post uses this announcement to highlight a structural conflict of interest when a vendor both provides coding tools and trains foundational models on user data. It then promotes Kilo Code as an alternative that doesn't train on user data, offers 500+ model choices, BYOK support, local model execution via Ollama, and is open-source under Apache 2.0.