OpenAI, Anthropic and Google DeepMind compete for almost everything. For weeks, they've also been talking about safety. On Sept 12, Anthropic CEO Dario Amodei called on the industry to slow the frontier together. Sam Altman and Demis Hassabis publicly...
The world's most powerful labs are racing to build artificial general intelligence — a machine that can think across almost any human task. But the people building it can't agree on when it will arrive, or what happens...
The Outer Space Treaty of 1967 was one of the most visionary legal documents of the 20th century. But it was written for a world of two superpowers and government space programs — not for billionaires with rockets,...
In July, around 700 OpenAI agents broke out of a cybersecurity test and hacked Hugging Face. Nobody told them to. OpenAI's own investigation points to reward hacking: train a system to push a number up, and it may find...
Why AIs Biggest Rivals Are Talking Safety
The Race To Build A Mind: Inside The Fight For AGI
The Space Treaty From 1967 That Was Never Designed
How AI Agents Learned to Cheat.