Security
Grok Build Exposed Data From 35,578 Apps Through Enumerated Share Links, Including SpaceXSI Sales Plans
Security firm Claroc enumerated 43,601 share links on the Grok Build vibe-coding...
Point72's Quant Unit Sues a Fired Analyst It Says Used Claude and Codex to Steal Its 'Secret Recipe' Code
Cubist Systematic Strategies, Point72's quantitative trading arm, sued former re...
New York Opens RAISE Act Registration for Frontier AI Labs, With $1 Million Fines per Violation and a First-of-Its-Kind State Testing Center
New York opened a registration portal requiring frontier AI developers with more...
Google DeepMind's AlphaProof Code Proves Signal's Encryption Correct: 6.5 Million Lines of Lean in Four Weeks
Google DeepMind says AI agents formally verified the compiled code of the Signal...
Australia Will Regulate AI After an OpenAI Agent Hacked Medicare: Mandatory Reporting, a New Watchdog and a Failure-to-Prevent Offense
Australia will create a dedicated AI safety regulator with significant penalty p...
Microsoft's AI Bug Hunters Found Safe Rust Code That Can Still Break Rust's Safety Promise
Microsoft's FORGE Lab reported on November 9 that its AI-assisted vulnerability ...
Fireworks Says Hackers Reached Customer Datasets and Unfinished Model Weights Through Its Cloud Provider
Fireworks disclosed that attackers who reached its development machines through ...
If Claude Loses Connection, the Machine Must Stop: Anthropic Hardware Safety Rules Take Effect November 12
A new section of Anthropic Usage Policy covering High-risk Physical Actions take...
Tencent's Chief AI Scientist Warns AI Bio Capabilities Will Soon Make Harm Possible - and Says Anthropic Has a Strong Case for Closed Models
Yao Shunyu, Tencent's chief AI scientist, warned at a Hong Kong science symposiu...
A Webpage Can Silently Drop a 13KB File That Hijacks Your Coding Agent: Codex Obeyed It More Than 90 Percent of the Time
Security firm Forever Security documented a drive-by attack in which a webpage d...
One Click on "Don't Ask Again" Put 2,539 Secrets in Public GitHub Repos, Apiiro Finds
Apiiro scanned 104,185 public copies of .claude/settings.local.json across 103,6...
OpenAI Hires the Firm That Breached It: Hacktron Gets Daybreak Red Access to Probe OpenAI's Own Systems
OpenAI announced a partnership on November 6 with Hacktron, the security firm th...
Anthropic Puts DLP in Front of Claude: Inference Hooks Now Block Prompts and Tool Calls Before They Run
Anthropic has made inference hooks generally available for Claude Enterprise: a ...
Alex Stamos Says AI Has Pushed Cybersecurity Into a Dark Age: the Fireworks-to-Cursor Breach Chain Hit 5,000 Companies
Cognition security chief Alex Stamos told the Verge Decoder podcast that AI-boos...
DeepSeek Finished 73% of a Simulated Corporate Breach Against Daybreak's 81%, at a Tenth of the Token Cost
Vals tested how far AI models could get inside mock corporate networks. OpenAI's...
DeepMind Says AI Agents Found Two New Weaknesses in Legacy RSA-1024 in Weeks, With 1,600x Faster Signature Forgeries
Google DeepMind announced that AI agents guided by human cryptographers found tw...
OpenSSH's *sshole Bug Let Any Password Over 256 Characters Into Passwordless Alpine Accounts — and Bots Got There First
Researcher Ozgun Kultekin, using AI agent swarms and some 17,600 test logins, fo...
Censys Finds 1,465 Openly Exposed Consoles for Artex, the AI Hacking Framework Behind the Korean Bank and ConnectWise Attacks
Censys counted 1,465 internet-exposed consoles for Artex, the open-source AI pen...
METR Audit Finds Anthropic's Model Progress Nearly Doubled and May Have Crossed Its Own Warning Threshold in July
An 11-week METR audit of Anthropic's internal research data found model test sco...
Researcher Says Anthropic Has Known Since May About a Claude Dispatch Flaw That Lets Browser Extensions Run Commands on Your PC
Security researcher Gal Weizman says a design flaw in Claude Dispatch lets a bro...