AI

Dashlane encrypted vaults were downloaded by threat actors exploiting API endpoints. The company has contacted the fewer than 20 affected users; others are safe.

The EU Tech Sovereignty Package introduces crisis powers over chip supply and cloud sovereignty tiers to reduce dependence on US tech for critical AI infrastructure.

Hackers tricked Meta AI support chatbot to change emails on valuable Instagram accounts. Patch is out, but MFA still protects you.

Ubisoft Ghostwriter AI pipeline restructuring reflects a wider industry move where 90% of developers use AI and Steam saw a 681% spike in AI-disclosed titles.

Memory OS is a new 6-layer local memory stack for Hermes Agent. Here's the reality check for developers and prosumers.

Red Hat NPM packages backdoored: attackers compromised official accounts to push a worm that steals credentials and spreads via npm install.

Claude Opus 4.8 adds effort control and parallel agents. We break down costs, safety, and what it means for your coding workflow.

Genesis World 1.0 delivers 400x faster policy evaluation using Nyx, Quadrants, and multi-physics, and 0.8996 correlation.

A jqwik prompt injection designed to sabotage AI coding agents ignites ethics debate after the hidden payload targeted users' existing work.

Anthropic's Milan office opens with enterprise customers Generali, Pirelli, Enel, linking its launch to Pope Leo's AI encyclical.

FBI and DHS documents reveal a new 'anti-tech extremism' threat category targeting peaceful data center protesters and AI skeptics

Walmart Sparky AI assistant is getting more autonomous, but new governance rules suggest safety guardrails like human approval for purchases.