China warns of AI risks in Anthropic's Claude Code. Anthropic has the third-best AI model by July 2026 at 89% YES.