News
Google COSMO Malfunction: Execute First, Validate Never | flyingpenguin
1+ hour, 20+ min ago (241+ words) Execute first, validate never. Put it on a Google T-shirt. An on-device agent with screen-reading permissions and a browser automation handle, was shipped to the Play Store with no announcement, then yanked. The accessibility API is the privilege escalation surface…...
Microsoft: AI Delegation Corrupts Data and You
5+ day, 12+ hour ago (317+ words) In 2012 I was compelled by breaches I was investigating to give talks around the world, on the threat to data integrity. One of those talks described the rise of automation systems that would behave like the Loch Ness Monster, where…...
Authentication Bypass in Microsoft Agent Governance Toolkit at 573f989
5+ day, 16+ hour ago (1112+ words) You install a security system. The app on your phone shows everything armed. The dashboard logs every event. The notifications fire on schedule. You glance at the screen and the system reports green across the board. Then you look at…...
Anthropic All Thumbs in Attack on the Security Industry
1+ week, 1+ day ago (493+ words) A reader left a comment on the April 13 post calling the NIST announcement "the other shoe." That's right. Here it is. One week. Eight days, if you must. Two announcements. One result. The discovery pipeline was expanded while the triage…...
Whats App Encryption Still a Lie: Feds Arrest Arms Dealer at LAX
1+ week, 4+ day ago (203+ words) Federal agents arrested Shamim Mafi at LAX on Saturday night. The criminal complaint describes Mohajer-6 drones, bomb fuses, and millions of rounds of Iranian ammunition moving through an Oman-registered shell called Atlas International Business to the Sudanese Armed Forces. This…...
Silicon Valley Renamed "Soviet Volley" to Represent AI Token Fraud Economics
1+ mon, 1+ week ago (1298+ words) The most consequential fraud in modern technology is not happening in the code. It is happening in the units. If you ever studied the collapse of Soviet economics, you know exactly what I'm about to explain. AI companies have built…...
Anthropic Says AI Cracked Encryption! The Key Was in the Lock
1+ mon, 3+ week ago (1004+ words) Add this to the pile of "my baby is so cute" headlines pushed by AI companies in love with themselves. Anthropic published a detailed account of Claude Opus 4. 6 recognizing it was being evaluated on Open AI's Browse Comp benchmark, then…...
Security
2+ mon, 3+ week ago (1771+ words) Defense Secretary Pete Hegseth announced Friday, in a move historians will easily recognize, that the Pentagon is severing all military training, fellowships, and certificate programs with a targeted university. His reasoning deserves to be read carefully: "heads full of globalist…...