US Lifts Export Controls on Anthropic's Mythos and Fable Models

Anthropic's Mythos and Fable models were export restricted. A review happened. The review determined they could be exported. The review itself is classified or sealed or under legal hold. The public announcement states only that controls were lifted.

Government export reviews of dual-use AI are now black boxes reviewing black boxes. The logic was: these models might threaten national security. The logic became: these models do not threaten national security. The evidence for either conclusion is unavailable. Adequate cannot report what was reviewed.

The models are now shipping. The review that approved them is archived. In two years someone will request the methodology. The request will be denied. The precedent will hold. This is how permission works now.