Inside the fight over Claude Mythos 5
TL;DR
The Trump administration ordered Anthropic on Friday to block Mythos 5 and Fable 5 for all foreign nationals, including its own foreign-national employees. Anthropic concluded the only workable path was a full product shutdown. The trigger was reportedly a Fable 5 jailbreak report. Anthropic says the issue was narrow, not unique, and that similar capability is available in OpenAI GPT-5.5.
Nauti's Take
This is the bill for high-volume safety marketing. When a lab sells a model as nearly too dangerous, it should not be surprised when politicians take that story literally, just less precisely than the lab hoped.
Still, a blanket foreign-national ban looks like a regulatory sledgehammer: it weakens trust in US AI platforms without clearly reducing the actual risk.
Briefingshow
The fight shows how quickly a frontier model can move from product launch to geopolitical risk. If US export controls start targeting model access rather than chips or infrastructure, global teams, customers, and partners face compliance rules that may be nearly impossible to operate. It also punishes labs that publicly dramatize their own risk profile.
Sources