The Week Responsible AI Died

A story of Anthropic and the fucked up world of the US Government

On Friday evening, Sam Altman posted on X that OpenAI had reached a deal with the Pentagon. It took him three sentences. The deal had been in motion since Wednesday.

That same Friday, Anthropic had been designated a supply chain risk by the United States government. A label previously reserved for Huawei and Kaspersky. Russian and Chinese companies. Adversaries.

This is the story of how the most important week in AI governance history unfolded. And what it means for anyone who still believes responsible AI is a viable business model.