Anthropic, the AI firm behind "Claude", released an even better model called "Fable" in June. However, this was a blink-and-you-miss-it event, as the US government had effectively banned its use 3 days later. Is this the best way to regulate AI?