Cortex — Veille numérique

Gemini

Power your agents: Gemini 3.8 Live with Live Avatar is now generally available

Il n’y a presque plus de tiret cadratin dans les textes de Claude Opus 5.5. D’après une étude de l’agence Graphite, la prose du nouveau modèle d’Anthropic est aussi moins maniérée que celle d’Opus 5. Les auteurs de l’étude ont malgré tout relevé 2 548 tournures employées bien plus souvent par cette IA que par un auteur humain.

clubic.com24 sept.

Claude IA

Inside GTA 🚗, Anthropic revenue jump 📈, agentic verification 🤖

Small AI Models Could Soon Self-Replicate on IoT Devices, Anthropic Warns

CyberXero Uses Claude Code in Campaign That Stole Over 628,000 Ukrainian Records

Anthropic Cuts Live Internet Access for Internal AI Tests After Claude Exploits Injection Flaws

AI disqualification yields new Nikon Small World in Motion winner

Anthropic launches free AI security scans for open-source projects

Anthropic loosens Claude’s cyber restrictions for verified defenders

Attackers Target Rejetto HFS Flaw That Enables Admin Session Forgery and RCE

Researchers find Chinese hacking campaigns targeting AI firms, Asian governments

Who’s liable when AI agents go rogue?

Anthropic warns of ‘catastrophic’ AI risks in its own IPO filing

OpenAI wants you to use AI — but not to train its AI

New Carbonato malware uses AI agents to hijack exposed Docker hosts

Power your agents: Gemini 3.8 Live with Live Avatar is now generally available

Meta’s Muse AI Charms can interact with each other

New Anthropic, OpenAI models make same promise: A little more for a lot less money

The most Viral new AI model isn’t an LLM at all

Relays Are Masking Chinese Access to Frontier AI Models in the US

Anthropic launches Claude Opus 5.5 with stricter safeguards for cybersecurity

Muse, Meta's extraordinarily privileged AI assistant, has a serious 0-day