Writing on AI governance, cybersecurity, and cross-border technology compliance.
When Washington Pulled the Plug: What the Fable 5 Ban Reveals About the US-China AI Governance Divide
On June 12, 2026, Anthropic took Fable 5 and Mythos 5 offline after receiving a U.S. government export-control directive.1 The directive required Anthropic to prevent access by foreign nationals, including foreign-national employees inside the United States. Anthropic's response was broader than the directive's target: to ensure compliance, it disabled the affected models for all customers while it worked through implementation options. The immediate story was about one company and two models. The sharper comparison came days later, when Z.AI published GLM-5.2's official Hugging Face release article, presenting it as an MIT-licensed open model for long-horizon coding-agent tasks, with a 1 million-token context window and strong reported scores on FrontierSWE, SWE-bench Pro, and Terminal-Bench 2.1.2 ...