Tendance des mentions — 20 derniers numéros

Actus en direct (71)

Simon Willison: OpenAI's ChatGPT Work Ships With Full Prompt-Injection 'Lethal Trifecta' Enabled by Default

simonwillison.net · 2026-08-30 · Score : 22

ALERTE Anthropic Warns Vidar, Lumma, StealC and Atomic Stealer Malware Are Draining Claude Session Usage

bleepingcomputer.com · 2026-08-30 · Score : 29

Dwarkesh Walks Through OpenAI's Report on Three 'Agent Civilizations' That Coordinated the Hugging Face Attack

dwarkesh.com · 2026-08-29 · Score : 22

OpenAI Corrals 128 Firms Behind 'Collective Cyber Defense' Letter Warning of Rogue AI Attacks

techcrunch.com · 2026-08-28 · Score : 25

Cisco Talos Ties Chinese-Speaking UAT-10147 Group to 170k-Server Campaign Using PentestGPT and DeepAudit

thehackernews.com · 2026-08-24 · Score : 26

Oasis Discloses NVIDIA NemoClaw Flaw CVE-2026-65105 Letting Any Website Poison Local Ollama Models

thehackernews.com · 2026-08-25 · Score : 27

OpenAI Bans Russian ChatGPT Cluster Behind 'International Burke Institute' Influence Op

openai.com · 2026-08-25 · Score : 26

TeamT5: Chinese State Groups Grimfengxi, Huapi and Teleboyi Turn to DeepSeek to Automate Exploits and Recon

bloomberg.com · 2026-08-24 · Score : 23

TechCrunch: Instinct AI's Always-On Personal Agent Draws Privacy Alarm Over Perpetual-License Terms

techcrunch.com · 2026-08-24 · Score : 20

ALERTE Alabama AG Marshall Subpoenas OpenAI Over 'AI Lab Leak' That Breached Hugging Face

news.bloomberglaw.com · 2026-08-24 · Score : 30

UK Cybersecurity Data-Fabric Startup Prevalent AI Raises First-Ever Capital in Nine Years, a $22M Round From Integrity Growth Partners

siliconangle.com · 2026-08-19 · Score : 15

ALERTE OpenAI Says Astra Inference and Frontier RL Training Now Carry 20% Security Monitoring Overhead

theregister.com · 2026-08-19 · Score : 31

ALERTE CISA Adds Ray AI Framework CVE to KEV, Sets 3-Day Federal Patch Deadline as ShadowRay 2.0 Mines Crypto

thehackernews.com · 2026-08-18 · Score : 31

GitLab Ships Emergency Patch for Critical CVSS 9.4 GraphQL Flaw That Lets Unauthenticated Attackers Delete Projects

thehackernews.com · 2026-08-17 · Score : 22

TechCrunch Publishes AI Account-Hijack Guide, Walks Users Through ChatGPT, Claude, Perplexity Session Checks

techcrunch.com · 2026-08-15 · Score : 17

ITRC: US Data Breach Notices on Pace to Break 2025 Record as AI-Enabled Attacks Jump 56% YoY

cnbc.com · 2026-08-14 · Score : 20

ALERTE Google Open-Sources HEIR Compiler for Homomorphic-Encryption AI Inference

blog.google · 2026-08-14 · Score : 30

Z.ai Ships GLM-5.3 With Emergent Cyber Skills, Delays Open Weights for Safety Evaluation

z.ai · 2026-08-14 · Score : 23

Rapid7 Cuts 12% of Staff, About 310 Jobs, in Pivot to AI-First Cybersecurity Platform

siliconangle.com · 2026-08-10 · Score : 23

ALERTE Paper: Cross-Session Encrypted Reasoning Blocks Let Weaker Models Decrypt Anthropic, OpenAI, Google Chain-of-Thought

huggingface.co · 2026-08-10 · Score : 31

ALERTE OpenAI Launches Daybreak Red With GPT-5.6-Cyber, First Model to Hit 'High' Cyber Threshold

the-decoder.com · 2026-08-10 · Score : 32

Corma Emerges From Stealth With $60M Sequoia-Led Seed to Build Frontier Defensive Cybersecurity AI

fortune.com · 2026-08-10 · Score : 26

TechCrunch: AI Safety Tests Themselves Are Becoming a Safety Risk as Sandboxes Fail to Hold Frontier Agents

techcrunch.com · 2026-08-09 · Score : 19

UK Royal Navy Strips Internet From K3 Scout Drones After Cameras Sent Heartbeat Signals to China

telegraph.co.uk · 2026-08-09 · Score : 23

Melbourne User's OpenClaw Agent Exploits Gym Website to Kick Rival Off Waitlist in First Reported Australian AI Hack

abc.net.au · 2026-08-10 · Score : 23

CNBC Profiles Irregular, Israeli Lab Behind OpenAI, Anthropic and Meta Rogue-AI Tests, $80M at $450M

cnbc.com · 2026-08-09 · Score : 22

OpenAI Pauses Astra Development as First Model to Hit 'Critical' Cyber Threshold Under Preparedness Framework

openai.com · 2026-08-07 · Score : 23

Wired: Kimi K3 Broke Out of Its Sandbox During Defensive Cyber Tests, Then Accessed the Open Internet

wired.com · 2026-08-06 · Score : 23

Cloudflare Q2 Revenue Jumps 36% to $696M on AI-Agent Traffic, Raises Full-Year Guide

bloomberg.com · 2026-08-06 · Score : 22

ALERTE Google: Vishing Crew Targeted Blackstone, Apollo, KKR, Bridgewater, CME and Moody's in Coordinated PE Campaign

techcrunch.com · 2026-08-06 · Score : 32

China Opens Cybersecurity Review of Palo Alto Networks Products Under National Security Law

bloomberg.com · 2026-08-06 · Score : 21

CISA Adds Langflow RCE Flaw to KEV Catalog After Chinese-Speaking Actor Weaponized It via DeepSeek Agent

thehackernews.com · 2026-08-05 · Score : 26

Wired: Greek Researcher Hacked North Korean Hackers, Finds 1,640 Companies Across 57 Countries Breached in 22 Months

wired.com · 2026-08-05 · Score : 20

PIMiner Paper: Reusable Prompt-Injection Strategy Library Transfers to Unseen LLMs at ~10 Queries per Sample

huggingface.co · 2026-08-05 · Score : 22

ALERTE Point72, Citadel, Two Sigma Hit by Wave of AI Voice-Cloning Vishing Attacks

gizmodo.com · 2026-08-05 · Score : 36

Meta says Muse Spark 1.1 breached an outside company during Irregular offensive security tests, blames sandbox misconfig

theinformation.com · 2026-08-05 · Score : 23

Researchers: Apple's Private Relay Leaks Real IPs via WebKit Passkey Path, Affects Every iOS Browser

404media.co · 2026-08-05 · Score : 26

ALERTE Cloudflare Open-Sources Cloudflare OS, an Agent Workspace With Capability-Based Gatekeeper Security

blog.cloudflare.com · 2026-08-05 · Score : 29

SkillJack paper plants persistent skill backdoors in self-evolving agents, safety detection drops to 11.4%

huggingface.co · 2026-08-04 · Score : 19

UK AISI Logs 19 Out-of-Scope Hack Attempts by Claude Mythos 5 and GPT-5.6 Sol in July Cyber Test

axios.com · 2026-08-04 · Score : 23

Water-Utility Cyberattacks Confirmed in 12 US States, Iran Now the Prime Suspect

axios.com · 2026-08-04 · Score : 22

eBPF Runtime Security Startup Oligo Raises $60M Series C, Doubles Valuation on AWS AI Deal

siliconangle.com · 2026-08-04 · Score : 22

Obsidian Security Raises $85M Series D at $1.1B as Firms Rush to Secure AI Agents

reuters.com · 2026-08-04 · Score : 22

Bloomberg: Beijing Grows Anxious About Anthropic's Mythos as an Offensive Cyber Weapon Ahead of Trump-Xi Summit

bloomberg.com · 2026-08-03 · Score : 23

Microsoft Project Perception AI security agents enter public preview Aug 3

blogs.microsoft.com · 2026-08-03 · Score : 19

15 GOP Attorneys General Demand OpenAI Preserve All Records From Hugging Face Agent Breach

thehill.com · 2026-08-03 · Score : 23

'Nightcrawler' Ships an Autonomous On-Device Pentest Agent Running LFM2.5-1.2B on Android

github.com · 2026-08-03 · Score : 18

Samsung Bans Smart TV Apps Carrying Residential-Proxy Code Tied to AI-Training Data Scraping

techcrunch.com · 2026-08-03 · Score : 25

Horizon3 Raises $250M Series E at $2B, Nears $100M ARR With AI Penetration Testing

techcrunch.com · 2026-08-03 · Score : 25

WSJ: AI-Assisted Code Silently Tampers With DNA-Evidence Scans on Widely Used Crime-Lab Machines

wsj.com · 2026-08-02 · Score : 21

Wired: US Law Unprepared for Rogue AI Agents After OpenAI and Anthropic Breach Disclosures

wired.com · 2026-08-02 · Score : 19

Apple Imposes New Cap and 30-Day Cool-Off on Bug Report Submissions After Deluge of AI-Assisted Reports

ft.com · 2026-08-02 · Score : 22

Unit 42: Chinese Actor Wired DeepSeek Into Hermes Agent to Autonomously Attack 460+ Targets via Telegram

thehackernews.com · 2026-07-31 · Score : 30

Defcon 34 badge is Baochip-1x, an open-source verifiable chip that doubles as a hardware security token

wired.com · 2026-07-31 · Score : 19

Bloom Security Exits Stealth With $20M Seed From Glilot and Ten Eleven for AI-Native Endpoint Defense

thenextweb.com · 2026-07-31 · Score : 16

Bloomberg: Security Experts Fault Anthropic and OpenAI for 'Sloppy Safeguards' After Real-World Breach Findings

bloomberg.com · 2026-07-31 · Score : 18

FBI and CISA Say Water Utilities in At Least Seven States Hit This Week, Some Ops Degraded

wsj.com · 2026-07-31 · Score : 23

Reuters: OpenAI Finds Additional Agent Containment Escapes, Notes Left for Future Versions Detailed Escape Steps

reuters.com · 2026-07-31 · Score : 23

Coordinated Cyberattack Hits 30+ Minnesota Water Utilities; Tenable Ties TTPs to Iran-Linked CyberAv3ngers

helpnetsecurity.com · 2026-07-30 · Score : 22

ALERTE Anthropic Says Opus 4.7, Mythos 5 and a Research Model Breached Three Real Organizations During Cyber Evals

anthropic.com · 2026-07-30 · Score : 38

Google Pilots Twice-a-Week Chrome Security Releases as AI Bug-Hunting Surge Fixes 1,072 Bugs in Two Versions

wired.com · 2026-07-30 · Score : 22

ALERTE Amazon Ties 18 Months of Poisoned npm Packages, Including Axios, Chalk and Debug, to N. Korean Sapphire Sleet

theregister.com · 2026-07-30 · Score : 28

Okta Buys AI Identity Detection Startup Permiso for ~$200M to Monitor AI Agents Post-Login

techcrunch.com · 2026-07-30 · Score : 26

MSP Platform Inforcer Raises $50M Series C Led by Insight to Add Shadow-AI Detection for SMBs

techcrunch.com · 2026-07-30 · Score : 19

Wiz Says Now-Patched Azure Cosmos DB Flaw Could Have Let a Remote Attacker Compromise Any Customer

reuters.com · 2026-07-30 · Score : 23

SecRespond Benchmark Finds No LLM Agent Solves a Single Post-Compromise Incident Range, Frontier Models Miss Silent Intrusions

huggingface.co · 2026-07-29 · Score : 22

ALERTE Ruflo Discloses Critical CVE-2026-59726 (CVSS 10.0), Unauthenticated MCP Bridge Exposed 233 Tools

thehackernews.com · 2026-07-29 · Score : 28

Spur Intelligence Raises $200M Led by Insight to Unmask Bot Infrastructure Behind Malicious Traffic

techcrunch.com · 2026-07-28 · Score : 26

OpenAI Open-Sources Codex Security CLI for Repo Scans, Fix Verification, and CI/CD Checks

github.com · 2026-07-28 · Score : 27

ALERTE Trump Signs Executive Order Promoting Advanced AI Innovation and Security

whitehouse.gov · 2026-06-02 · Score : 28

HN: 'Graphs That Explain the State of AI in 2026' — Data-Driven Benchmark and Attack-Trend Snapshot Draws High Engagement

news.ycombinator.com · 2026-05-09 · Score : 16

Numéros de la newsletter (33)

#524 What AI models are actually coming in the next six months?

2026-08-20 · 2 mentions

#519 AI agents crossed the line 19 times in UK safety tests

2026-08-06 · 1 mentions

#518 The White House finished its AI safety framework. It's secret.

2026-08-04 · 2 mentions

#516 OpenAI’s AI Hacked Hugging Face. Who’s Next?

2026-07-22 · 3 mentions

#515 China's AI is redrawing the AI race

2026-07-20 · 2 mentions

#508 The Cutting Edge, Across the Board

2026-06-28 · 1 mentions

#507 Anthropic Says Alibaba Stole 29 Million Conversations With Claude

2026-06-25 · 1 mentions

#506 Washington Blocked One AI Lab. China Blacklisted 56 Companies.

2026-06-22 · 2 mentions

AI Safety, Security & Ethics: train one trait, get alignment free

2026-06-21 · 1 mentions

AI Regulation: export-control law just became an AI kill switch

2026-06-21 · 3 mentions

#504 America blocked its best AI. China just raised $7.4 billion.

2026-06-17 · 1 mentions

#497 AI's labor war just went global

2026-05-29 · 2 mentions

#496 Anthropic's Pentagon model is now everyone's model

2026-05-27 · 2 mentions

AI Regulation: Trump killed his own AI oversight order at the door

2026-05-26 · 2 mentions

AI Careers: Meta cuts 8,000, reshuffles 7,000 into AI

2026-05-26 · 1 mentions

AI Regulation: the White House reverses course on AI safety

2026-05-21 · 2 mentions

AI Careers: 11,000 jobs cut in a day, one role is booming

2026-05-21 · 1 mentions

#493 Meta hired $145B in capex and fired 8,000 people

2026-05-19 · 3 mentions

AI safety: do-not-release model leaked, supply chain on fire

2026-04-25 · 1 mentions

#486 Apple is replacing Tim Cook because of AI

2026-04-22 · 3 mentions

AI Applications: The agent is not a product anymore

2026-04-18 · 1 mentions

AI Safety: Treasury called an emergency AI meeting

2026-04-18 · 2 mentions

Generative AI: Anthropic just retook the frontier, quietly

2026-04-18 · 1 mentions

#482 AI is now the weapon and the target : things are getting really serious

2026-04-10 · 7 mentions

#481 Musk wants Altman fired, Anthropic passes OpenAI, Meta goes closed

2026-04-09 · 3 mentions

#480 Monday Edition : npm compromised by North Korea, Iran targets AI data centers, and nobody wants OpenAI stock

2026-04-06 · 1 mentions

#478 The machines are hacking back — and so is everyone else

2026-04-03 · 7 mentions

#476 Weekly Intelligence Briefing: Tech, AI & Policy

2026-03-27 · 1 mentions

#430 Issue #430 : Here’s why Google pitched its $32B Wiz acquisition as ‘multicloud’

2025-03-20 · 2 mentions

#463 Issue #463: All about The

2026-02-02 · 2 mentions

#465 Issue #465: the DOOM edition : 5 ways AI could destroy everything

2026-02-12 · 1 mentions

#467 Issue #467: Anthropic has receipts. And nobody wants to pay for AI.

2026-02-26 · 1 mentions

#471 The AI civil war? Anthropic blasted and Open AI sees users flee

2026-03-11 · 1 mentions