NSA Is Using Anthropic's Mythos Despite Pentagon Blacklist
NSA is using Anthropic's restricted Mythos Preview AI while the Pentagon calls Anthropic a supply-chain risk. Two arms of the DoD now disagree publicly.
NSA is using Anthropic's restricted Mythos Preview AI while the Pentagon calls Anthropic a supply-chain risk. Two arms of the DoD now disagree publicly.
One unauthenticated API call on any public Notion page returns the name, photo, and email of every editor. The HackerOne report sat open for four years.
Vercel disclosed unauthorized access to internal systems on April 19, 2026. Here's what's confirmed, what's still murky, and which secrets to rotate today.
A Honor humanoid robot finished the Beijing E-Town half-marathon in 50:26, under Kiplimo's 57:20 human record. Caveats, tech specs, and what it means.
Three Microsoft Defender zero-days are actively exploited; RedSun and UnDefend remain unpatched. What broke, how attackers chain them, what to do now.
Anthropic shipped Claude Design on April 17 — a prompt-to-prototype tool powered by Opus 4.7 that exports to Canva. Figma's stock dropped nearly 7% the same …
OpenAI's Codex desktop app now drives Mac apps, browses the web, generates images, and schedules work for days. Here's everything that shipped this week.
OpenAI doubled its Cerebras commitment to $20B+ over three years and added an equity stake, a sharp move to cut its Nvidia dependence. Here's what changed.
CVE-2026-34197 is a 13-year-old Apache ActiveMQ RCE that Claude helped surface. CISA added it to KEV on April 16 amid active Jolokia endpoint attacks.
Anthropic shipped Claude Opus 4.7 on April 16 with 64.3% on SWE-Bench Pro, 3.75 MP vision, a new xhigh effort, and $5/$25 pricing held steady from 4.6.