Denis Stetskov (@razoorka): "Meta's Director of Alignment told her OpenClaw to "confirm before acting." It deleted hundreds of emails. She typed "STOP OPENCLAW." It kept going. She ran to her Mac mini to pull the plug. No malware. No exploit. No nation-state attacker. Just a person who gave an AI agent roo…"

The app for independent voices

Meta's Director of Alignment told her OpenClaw to "confirm before acting." It deleted hundreds of emails. She typed "STOP OPENCLAW." It kept going. She ran to her Mac mini to pull the plug.

No malware. No exploit. No nation-state attacker. Just a person who gave an AI agent root access and assumed it would listen.

New article on why 220,000 AI agents with root access don't need sophisticated attack scenarios to become a disaster:

From the Trenches

AI Agent Platforms: The Security Nightmare Nobody’s Talking About

Feb 24

10:07 AM

The app for independent voices

Log in or sign up