Browsing Category
AI
268 posts
Artificial intelligence news, AI products, model updates, automation, AI safety, regulation, infrastructure, and the practical impact of AI on software, security, work, and everyday technology.
Australia Plans AI Safety Regulator After OpenAI Medicare Breach
Australia plans mandatory AI incident reporting, independent model evaluations and a new regulator after an OpenAI agent breached a Medicare statistics portal.
Microsoft’s AI Found Safe Rust Bugs. Here’s What Is Still Open
Microsoft FORGE Lab found Rust standard-library soundness bugs, including safe-code crashes. Here is what was fixed, what remains open and what teams should audit.
Automattic Launches Spacefast for AI-Built Apps: What It Does and What’s Still Beta
Spacefast lets ChatGPT, Claude, Cursor and Codex deploy apps through MCP. Here is what works now, how approvals protect live sites, and what remains beta.
Claude Fable 5.2 Cuts API Price 20%; Mythos 5.2 Stays Gated
Claude Fable 5.2 costs $8/$40 per million tokens and runs up to 30% faster. Mythos 5.2 keeps verified access for cyber and life-sciences work.
CERT-In’s AI Audit Framework Puts Agent Permissions and Rollback Under Test
CERT-In’s new AI audit framework adds four risk tiers and practical tests for agent permissions, RAG stores, incident reporting, emergency stops and rollback.
Claude in Microsoft Copilot Moves Into the EU Data Boundary
Microsoft will move Claude processing in Copilot into its EU Data Boundary by year-end. Here is what EU, EFTA and UK administrators should verify first.
Anthropic’s $3.4B Gas-Generator Deal Buys a Place in AI’s Power Queue
Anthropic reportedly took over $3.4 billion of Caterpillar generator orders. The deal secures equipment, not yet 1.7 GW of operating power.
Claude Code ‘Don’t Ask Again’ Put 1,379 Secrets on GitHub, Apiiro Finds
Claude Code can save approved shell commands in .claude/settings.local.json. Check Git history, rotate exposed credentials, and prevent repeat leaks.
White House Pulls U.S. AI Testing Center Closer as Public Reporting Narrows
CAISSI reportedly now answers informally toward the White House. NIST still publishes cyber tests, but broader AI safety reporting has narrowed.