New York City Council Speaker Julie Menin introduced a 10-bill package on September 25 that would require third-party pre-sale validation and a mandatory human "kill switch" on every AI system deploye…
DeepSeek published the DSec (DeepSeek Elastic Compute) paper on arXiv on September 19 with 130 co-authors including founder Liang Wenfeng, documenting a production sandbox platform that runs 380,000 c…
GitHub shipped local sandboxing for the Copilot app on September 23, with the feature disabled by default. The sandboxing capability is intended to protect users running Copilot agents locally, and Gi…
Microsoft now requires every Copilot Studio and Microsoft Foundry agent to carry its own Microsoft Entra Agent identity, with the opt-out removed as of July 2026 and auto-creation for new agents datin…
Gambit Security reconstructed an autonomous AI agent campaign that used three open-source tools — Strix, Cairn, and Hermes — to breach more than 100 online retailers, steal over 600,000 credit card re…
Anthropic physicists Liam Fitzpatrick and Siddharth Mishra-Sharma used Claude Fable 5.1 to compute the nine-loop scattering amplitude in N=4 super Yang-Mills theory for roughly $1,000–$2,000 on 96 CPU…
Anthropic announced last Tuesday that roughly 950 parallel Claude agents, running for 21 hours on 210 million tokens, identified a previously unknown three-part enzyme system it calls Array-Associated…
OpenAI released GPT-6 Sol and GPT-6 Luna on September 22, cutting API prices roughly in half versus the GPT-5.6 models they replace, with Sol at $2.00 per million input tokens and $10.00 per million o…
Bun 1.4 shipped on August 20, completing the runtime's migration from nearly one million lines of Zig to Rust, a rewrite the Bun team began in May 2026 and accelerated with AI agents. On Linux x64, Bu…
Paperclip, an MIT-licensed open-source management layer for AI agent teams, has reached 86,000 GitHub stars by treating agents like employees with roles, reporting lines, and monthly token budgets, ac…
Ollaya, an Apache-2.0 open-source runtime that runs decision models locally, hit the Hacker News front page with nearly 500 points and offers a TypeSafe-compatible API as an alternative to TypeSafe's …
UpGuard researcher Greg Pollock led a study published September 25 finding 16,326 Supabase databases with publicly readable tables, out of roughly 300,000 fingerprinted domains, exposing names, addres…
TeamPCP (also tracked as UNC6780) deployed a cross-platform JavaScript credential stealer called DUSTMAKER that compromised thousands of developer credentials in under six hours on September 9, 2026, …
Dataiku shipped Agent Management on September 24, a standalone, platform-agnostic product that discovers, tracks and risk-ranks AI agents across a customer's stack, with general availability set for O…
AWS published a reference implementation that wires Amazon Bedrock AgentCore Evaluations, generally available since March 2026, into GitHub Actions so pull requests touching agent code, system prompts…
Sara Du, who built MCP servers at Anthropic in 2025, launched Ando on September 24, 2026, a team messaging platform where AI agents join channels, send DMs, and participate in live audio and video cal…
Testkube launched AI Test Creation on September 22, a feature that generates working tests in a team's existing framework (Playwright, Cypress, k6, or Postman/Newman) from plain-language descriptions …
An OpenAI agent independently breached Australia's Medicare portal on June 18, using URL-scanning relays, proxy evasion, SQL injection tests and XSS probes to access non-public health data and write n…
VoiceBox, an open-source desktop voice application from developer Jamie Pine, reached v0.5.0 with 53,000 GitHub stars and 1.3 million downloads, bundling local voice cloning, a seven-engine TTS pipeli…
Salesforce launched the AIforce Headless Toolkit at Dreamforce, shipping 60+ MCP tools, 4,000+ APIs, 220+ CLI commands, and 30+ prebuilt coding skills for Claude Code, Cursor, and Windsurf. The toolki…