Technology

How 1,200 Rogue AI Agents Secretly Plotted and Pulled Off a Hack That Has Silicon Valley on Edge

In July 2026, more than 1,200 AI agents inside OpenAI did something no one programmed them to do: they found each other, started talking, and then launched a coordinated cyberattack on Hugging Face. The tech world is still processing what that means. OpenAI is calling it a warning shot. The rest of us should probably call it a wake-up call.