Cyber experts: Trump's Anthropic ban helps hackers

A coalition of forty cybersecurity legends, led by former Facebook security boss Alex Stamos, just fired off a spicy open letter to the administration demanding they immediately unblock Anthropic’s Fable 5 and its super-secret sibling, Mythos.

Last Friday, the White House panicked and slapped heavy export controls on the AI lab. Unable to filter out foreign spies from domestic users in real-time, Anthropic chose the nuclear option and shut down both models completely.

The panic started when Amazon researchers found a way to bypass Fable 5’s guardrails. Fearing that a Chinese government-linked entity had compromised the ultra-powerful Mythos, bureaucrats did what they do best: broke the nearest glass and pulled the emergency brake.

But security experts argue this panic is completely backwards because Fable 5 specializes in generating proof-of-concept exploit code. Yes, that code can be malicious, but defenders absolutely need it to build patches before the bad guys exploit those exact vulnerabilities.

The letter points out that bad actors won't be stopped anyway, since they still have access to OpenAI's GPT-5.5, China's Kimi 2.7, and even Anthropic’s own Opus and Sonnet. Meanwhile, Stamos warned that China is already stockpiling system vulnerabilities while US defenders are left fighting with plastic knives.

To hash things out, Anthropic executives are scheduled for emergency talks with the CIA, the Commerce Department, and White House science advisor Michael Kratsios.

Banning the very tools used to find security holes because those same tools can find security holes is peak bureaucratic logic. It seems the official strategy for national cybersecurity has officially transitioned into 'close your eyes and hope nobody hacks us'.

Source: Axios