RE: AI heist: Anthropic complaining about industrial-scale distillation attacks. Chinese models using Anthropic APIs to train their own. Anthropic and Google blocking OpenClaw users.