Keep pulling the thread on Hayden Field.
The U.S. government imposed export controls on Anthropic's new model, Claude Fable-5, and the underlying Mythos model it is based on.
The U.S. government's export controls on Anthropic's models restricted access for all foreign nationals, including those employed by Anthropic within the United States.
The Trump administration gave Anthropic a 90-minute ultimatum to shut down its Fable 5 model after being alerted to a potential jailbreak by Amazon CEO Andy Jassy.
According to sources, the capabilities demonstrated in the alleged Fable 5 jailbreak discovered by Amazon researchers can also be achieved by OpenAI's GPT-5.5 and other models.
In response to the U.S. government's export controls, Anthropic took both its Fable 5 and Mythos 5 models offline.
As a result of the U.S. government's actions against Anthropic, international companies are reportedly signing backup contracts with non-U.S. AI providers and deploying open-weight models on alternative hardware.
There are rumors that the Trump administration is attempting to acquire an ownership stake in both Anthropic and OpenAI.
Anthropic previously described its Mythos model as a "potential cyber weapon" and initially planned to release it only to enterprise and government clients designated as "cyber defenders."
The Department of War previously designated Anthropic a "supply chain risk" because the company refused to allow its AI models to be used for all lawful military purposes.
Hayden Field predicts the Trump administration will roll back the export control rule that bars foreign nationals from accessing certain AI models.
Anthropic's Fable 5 is a safeguarded version of the more powerful underlying Mythos 5 model.
Upon its release, security researchers found the guardrails on Anthropic's Fable 5 model were so strict they prevented research by downgrading to older models when boundaries were pushed.