Published September 13, 2026 at 5:22 PM Comment 0 𝕏 Print © Anthropic Seven Chinese labs, nearly 200 million diverted exchanges, and Kimi or DeepSeek users who unknowingly spoke to an American competitor’s model. Anthropic’s report reads like a slanted industrial thriller. Distillation, in AI jargon, refers to a perfectly legitimate training method: a large “teacher” model is made to answer thousands of questions, and then a smaller “student” model is trai