From the archive
Saturday, 21 November
Folded from 19 sources that day.
Sorted by quality and by how many sources covered it, the same for everyone.
-
AI Agents Found a Zero-Day in MONAI Label and Gave Attackers Root on a University Server
Amazon Threat Intelligence says suspected China-linked hackers used autonomous AI agents to find an unknown flaw in the MONAI Label imaging server. It lets unauthenticated users write files outside the storage folder.
-
Signal Blocks AI Agents From Reading Private Chats via Screenshots on Android and iOS
Signal will restrict screenshots by default to keep AI agents from reading private chats on iOS 27 and some Android versions. It admits the protection is incomplete.
-
Riemann Hypothesis proved twice independently
Anthropic's Levent Alpöge released a Claude-driven proof as about 1.3GB of Lean code, while Alessandro Masullo published a 940-page human proof. Both claims will take years to verify.
-
"It is a rollercoaster ride": Angie Jones of the Agentic AI Foundation on keeping up with AI agents
In an interview, Angie Jones argues that using AI agents and building agents others can depend on are different skills, and she discusses keeping pace with the field.
-
GPT-6.2 Sol vs. Claude Opus 5.5: Not as accurate, but still might be the better buy
In three hands-on agent, coding and data analysis tests, Sol is about half the list price of Opus 5.5 and far cheaper per task, though Opus 5.5 is slightly more accurate.