The White House Is Making Up Its Rules for AI in Real Time
TL;DR
WIRED reports that Anthropic still cannot distribute Claude Mythos or Fable 5 nearly a week after the Trump administration sent an export-control directive. The core problem remains unclear: the White House says Anthropic acted recklessly, while Anthropic says it did not violate any concrete procedure or rule. Officials were reportedly worried about SK Telecom, alleged China ties, and possible jailbreaks in Fable 5. Customers including Apple, Meta, and Fortune 500 firms are locked out.
Nauti's Take
This is the kind of regulation that makes everyone nervous: officially voluntary, practically mandatory, and vague in its reasoning. Frontier models need oversight, especially around exports and cyber risks.
But when nobody can say where the line was crossed, companies learn to manage politics early instead of relying on clear public rules.
Briefingshow
The case shows how quickly AI regulation turns into raw power when clear rules are missing. For labs, technical safety is no longer enough; they also have to manage political expectations and avoid surprising officials. For users and companies, that makes model availability harder to plan.
Sources