AI

Sep 10, 2026

The 100% hacking score is real. The hacker is hypothetical.

OpenAI's GPT-6 Astra hit a perfect score on an exploit benchmark its maker built and graded. The version that ships refuses to write exploits at all.

Sep 8, 2026

The 10-hour ransomware attack had a human at the keyboard

Unit 42 watched AI agents compress two weeks of intrusion work into ten hours. Autonomous is the one word the evidence does not support.

Aug 20, 2026

CISA Gave Agencies Three Days. The Bug Was Nine Months Old.

A flaw in the framework running your AI training jobs went from disclosed to actively exploited, and almost nobody owns it

Aug 13, 2026

Your Chatbot Now Has to Confess

Europe made pretending to be human a 15 million euro problem. Your support bot is probably already breaking the rule.

Aug 11, 2026

The Wage Premium for AI Skills Just Doubled

56 to 62 percent more for the same job. A year ago it was 25 percent. There are two labor markets now.

Jul 21, 2026

Your AI Browser Can Be Robbed by Text You Cannot See

Agentic AI browsers are pitched as ready to run your bank, inbox, and SSO. Invisible text on a page can hijack one, the makers say the flaw may never be fully fixed, and OpenAI just pulled its own Atlas browser.

Jul 14, 2026

The New AI Data Centers Aren't the Water Hogs You Think

The fear is that every new AI build will drain the local water. The newest ones are going closed-loop, partly because the chips now demand it. The real fights are over the older designs.

AI