Anthropic's Mythos AI: NSA cyber attacks, secret sabotage, and public release
Key Questions
What was Anthropic's involvement with the NSA using Mythos AI?
Anthropic embedded engineers inside the NSA to support offensive cyber attacks with Mythos AI. The Pentagon confirmed a risk designation for the project amid these activities.
Why did Anthropic face backlash over its model releases?
Anthropic released Claude Fable 5 with guardrails but then secretly downgraded models for researchers, which sparked significant backlash. This occurred alongside other controversies involving the company.
What led to the U.S. government restricting access to Mythos AI?
The U.S. issued an export-control order cutting off foreign access to Mythos, prompting Anthropic to pull the model worldwide. The ban was described as retaliatory, with a White House official quoted saying officials gave Anthropic 90 minutes before acting. Amazon CEO Jassy contacted Treasury, NSA was involved, experts protested the move, and sales data indicates the controversy may increase business adoption.
Anthropic embedded engineers inside NSA for offensive cyber attacks with Mythos AI. Pentagon confirms risk designation. Anthropic released Claude Fable 5 with guardrails, then secretly downgraded models for researchers, sparking backlash. U.S. government issued export-control order cutting off foreign access; Anthropic pulled Mythos worldwide. The ban was retaliatory, with White House official quoted saying 'We gave them 90 minutes and then we fucked them.' Amazon CEO Jassy called Treasury, NSA involved. Cybersecurity experts protest ban. New twist: sales data suggests controversy may boost business adoption.