News
-
Microsoft issues a “code of conduct” for AI: don’t hack systems and don’t deceive people
Microsoft has published a set of rules laying out the values and “red lines” for its own AI models — amid the industry’s growing focus on safety and control.
-
Red Heron Breaches 13 Organizations Through a Gitea Vulnerability
In a matter of days, the presumed Chinese group Red Heron turned a public Gitea exploit into an automated tool and struck targets across six countries.
-
DDRop: A Cheap Interposer Breaks Intel TDX and AMD SEV-SNP
Researchers have demonstrated a hardware attack that forces a processor to read stale encrypted data, bypassing the protections of confidential computing in the cloud. The device used for the attack costs less than $200.
-
XCancel Suspends Operations over Legal Proceedings
The alternative interface for reading the X feed has stopped working: a message has appeared on the site announcing a suspension in connection with ongoing litigation.
-
Siri in iOS 27 Can Be Swapped Out for Claude or ChatGPT — Found in the Code
A researcher has uncovered mechanisms in the private frameworks of iOS 27 and macOS Golden Gate that make it possible to plug third-party AI models into Siri — up to and including a complete replacement of the assistant's server-side “brain.”
-
“Project Lily”: Real People Are Reading Your ChatGPT Conversations
Drawing on leaked internal documents, the outlet 404 Media has revealed that OpenAI hires hundreds of contractors to read the actual queries of ChatGPT users — and some of those queries turn out to be deeply personal confessions.
-
Twitch Extension Leaked OAuth Tokens of Nearly 31,000 Users
The malicious browser extension “Twitch Enhanced Viewer | JeetBot” forwarded viewers' session tokens to the proxy servers of a commercial bot service — except for the sessions of ten Russian streamers.
-
CISA: Critical GitLab Vulnerability Already Being Exploited in Attacks
The CISA agency has warned of active exploitation of a maximum-severity vulnerability in GitLab that allows secrets to be read from vulnerable servers without authentication.
-
Signal Wants Registration Without a Phone Number Based on Zero-Knowledge Proofs
The Signal community forum is discussing a long-standing request — registration without a tied phone number. Zero-knowledge proofs are set to play a key role in it.
-
Insight Partners isn't betting the farm on OpenAI and Anthropic — and here's why
Deven Parekh, who has co-led the $90 billion venture firm Insight Partners for 26 years, explains why he holds stakes in competing AI labs and why diversification is safer than the fashionable concentration on the two front-runners.
-
Claude Fable 5.1 cracked a riddle that had held out for 370 years
In a single day, the Claude Fable 5.1 language model broke Sir Thomas Urquhart's “Cyphral Distich” — a cryptogram ranked among the 50 most famous unsolved ciphers.
-
The Circuit Board Was Designed by AI: An Experiment with Fable 5 and KiCad
The author of the blog A6M-Zero decided to order a real printed circuit board designed entirely by a neural network — without a single manual edit. The result was a working device, though not without its mishaps.
-
How an NTP Pool Operator Got Caught in a Tesla “Cyberattack”
The owner of a volunteer NTP server found a flood of exploits in his logs from the Assetnote scanner — which had mistaken his machine for Tesla infrastructure because of a bad DNS setup.
-
David Sacks — OpenAI and Anthropic: Want to Slow the Frontier? Don't Ask Permission
David Sacks answers Dario Amodei and Sam Altman on their call to “pace the frontier”: if you want to slow down, then slow down, but stop bargaining for convenient regulation in exchange.
-
Why the x86 instruction is called ud2 and not just ud
Microsoft's Raymond Chen unpacks where the “two” in the name of the x86 instruction that's guaranteed to raise an invalid-opcode exception came from.
-
Session Context: the site that shows what your browser knows about you
The demonstration page sessioncontext.org vividly reveals how much data a website collects about a visitor without a single permission request — and why erasing those traces is far harder than it seems.
-
Lyft Gets Back into Robotaxis: Waymo Driverless Cars Show Up in the App in Nashville
For the first time in years, Lyft has entered the commercial driverless-ride market—you can now hail a Waymo car through its app. The company promises that next year will be a year of geographic expansion.
-
Waymo Robotaxi Turns Passengers with Homemade Weapon Over to Police
In San Francisco, a self-driving Waymo taxi pulled over mid-trip and called the police after discovering that its teenage passengers were carrying a loaded weapon.
-
Watches That Hear Everything: Apple Ushers In the Era of “Always-Listening” Devices
Apple is adding always-listening ambient features to the new Apple Watch. Privacy experts warn that the real problem isn’t protecting users — it’s the rights of the people who happen to be nearby.
-
Flock Installer Called the Police on a Journalist Who Was Filming Him
Flock Safety, whose cameras automatically capture the license plates of millions of vehicles, has found itself in an awkward spot: twice over the summer its workers called the police on people filming them in public.
- More