In September 2026, a quiet but consequential dispute emerged between Anthropic and several Chinese AI laboratories over the practice of model distillation — the harvesting of one system's reasoning to train another. Anthropic named DeepSeek, Moonshot AI, Xiaomi, and Alibaba as participants in what it characterized as unauthorized extraction at scale, while China's Foreign Ministry dismissed the allegations as geopolitical distortion rather than technical fact. The exchange reveals something older than any algorithm: the enduring tension between those who build knowledge and those who seek to i