US Government Forced Anthropic to Shut Down Its Most Powerful AI Models — for Everyone

The story that lit up Hacker News: on June 12, Anthropic received an export control directive from the US government. Authorities demanded the company cut off access to its most powerful models — Claude Fable 5 and Mythos 5 — for any foreign nationals, not just abroad but inside the US too, including its own non-American employees.

The catch: filtering out "foreigners" among hundreds of millions of users in real time is impossible. So Anthropic took the drastic step of simply shutting down both models for everyone. The Commerce Department's stated reason is a worry that a Fable 5 "jailbreak" could become a national security threat.

The company itself is stunned by the outcome. Anthropic said plainly that it disagrees with the idea that finding a narrow potential vulnerability should cause the recall of a commercial product used by hundreds of millions of people. In effect, its own cautious safety warnings backfired.

Why this matters for business: if you build products or processes around one specific AI model, this is a wake-up call. Access to a powerful tool can vanish in a day — not for technical reasons, but by a regulator's decision. The lesson is simple: don't tie critical processes to a single vendor, and keep a backup ready.

Source: https://www.anthropic.com/news/fable-mythos-access