Anthropic's Claude Fable 5 and Claude Mythos 5 taken offline
Technology
Anthropic's latest AI models, Claude Fable 5 and Claude Mythos 5, have been taken offline after just three days.
The US government stepped in because users found ways to bypass safety features, raising big concerns about national security.
These models were meant to be safer than earlier versions, but things moved fast once the risks became clear.
White House gives Anthropic 90 minutes
The White House called the situation a "national security threat" and gave Anthropic 90 minutes to disable the models. There were worries about possible access by groups linked to China, though Anthropic says that didn't happen.
Right now, Anthropic executives are in Washington trying to sort things out with officials, but talks are still ongoing and nothing's settled yet.