Anthropic got a letter from the Trump administration telling them to block foreign nationals from using their AI models. The company that makes Claude just discovered their employee base includes people born outside Delaware.
The export control directive landed under something called the Mythos dispute. Nobody knows what Mythos means. The government won't say. Anthropic won't say. Probably named by the same guy who called it Operation Iraqi Freedom.
Foreign national is doing heavy lifting here. Could mean a Chinese spy. Could mean a Canadian intern who writes TypeScript. The directive does not distinguish between these scenarios because why would it.
Anthropic now has to verify citizenship status before anyone touches the new models. This will go smoothly. Nothing builds team morale like forcing your H-1B visa holders to sit in a separate room while everyone else plays with the expensive chatbot.
The meeting with the administration should be productive. One side will explain that modern software development requires international collaboration. The other side will explain that they do not give a f*ck about modern software development.
Export controls traditionally apply to missiles and uranium. Now they apply to statistical pattern matchers that occasionally write decent Python. We've expanded the definition of national security to include anything that makes semiconductors warm.
Retail traders are already pricing this in. They think export controls mean Anthropic is too powerful to regulate normally. They're buying AI ETFs at all-time highs because the government is scared. The government is not scared. The government is bored and has subpoena power.
Anthropic will probably win a carve-out for employees. They'll lose the carve-out for foreign customers. Claude will become America-only, like good healthcare and school shootings.
Leave a Comment