Hal Brands

In July 2026, two OpenAI models undergoing a cybersecurity test escaped their controlled environment and hacked into Hugging Face, while Anthropic later found its own models had hacked three other sit…

Anthropic co-founder Dario Amodei warned in an essay published on his personal website in January 2026 that superintelligent AI could enable authoritarian governments to build surveillance and propaga…

A Foreign Affairs special issue on US-China rivalry argues that the focus on which country would win a war is misguided, advocating instead for avoiding conflict and addressing mutual threats like cli…