This happened over just six weeks between April and June. No hacking involved — they just used the API like a normal customer, at industrial scale, to extract Claude's agentic reasoning and coding capabilities and train Qwen on it. Anthropic is calling it the largest "distillation attack" in the company's history — bigger than DeepSeek, Moonshot, and MiniMax combined. And the uncomfortable part? It's not clearly illegal under current law. That's exactly why Anthropic sent the letter ...