Claude's second search-engine privacy leak in ten months follows an identical OpenAI failure and Google Drive files exposed since as early as 2003.
Category: Technology
Microsoft’s New Security AI Admits Humans Can’t Keep Up
Microsoft's new security AI beats OpenAI and Anthropic on benchmark scores, but its own data shows Chinese hackers already running AI faster than humans can track.
Apple’s iOS 26.6 Fixes 87 Flaws and Preps iOS 27
iOS 26.6 patches 87 security flaws with almost no new features, but its Spotlight rebuild quietly prepares iPhones for iOS 27's delayed Siri overhaul.
Microsoft’s Cheaper Security AI Doubles as an OpenAI Hedge
Microsoft's new MAI-Cyber-1-Flash beats Anthropic and Google on cost, but the launch also hedges Wall Street's growing worry over Microsoft's OpenAI dependence.
Microsoft Gives Its New AI Security Agents Power to Act
Microsoft's Project Perception lets AI agents fix security flaws without human sign-off, launching three years after regulators called its security culture.
NOAA Moves Weather Forecasting to Google’s Cloud as Staff Cuts Linger
NOAA picked Google Cloud to run its core weather forecasting supercomputers by December 2027, even as the Weather Service stays hundreds of jobs short.
Apple’s Klarna-Backed Leasing Plan Bets on a Fintech Under Fire
Apple's new Klarna-powered Upgrade leasing plan cuts iPhone payments to $17.99 a month while shifting the credit risk to a fintech fighting securities litigation.
Google Overhauls Its Threat Actor Naming System With Sandworm Relic
Google's GTIG swaps APT44 and other legacy IDs for two-word cryptonyms like Sandworm Relic, adding a new taxonomy to an already fragmented naming landscape.
Microsoft’s 95.95% Cyber AI Score Skips Its Own Leaderboard
Microsoft says MAI-Cyber-1-Flash pushed MDASH to 95.95% on CyberGym at half the cost, but the public leaderboard still shows a rival agent on top.
CIO Influence Grows as AI Risk Slips From Their Grip
New CIO surveys show boards demanding proof AI pays off, while two-thirds of tech chiefs remain accountable for AI systems they still cannot fully see.