I can't imagine how you'd restrict access to these models to USA nationals only. Every company in the USA would have to verify the id of every employee's model use. What do you do about programmatic AI workflows? How does Anthropic even continue to function internally right now on this Mythos and the future model development?