How did the government decide OpenAI’s frontier model was safe to release? "Exactly what that dialog looked like between the government and Anthropic and OpenAI is unclear."