Vercel Workflow Dev Kit: Durable AI Agents in Two Lines
The Vercel Workflow Dev Kit July 2026 update adds NestJS support and default run safeguards. Two string literals make any TypeScript AI agent ...
NVIDIA Cosmos 3 Edge: Run a Robot Brain on Your RTX GPU
NVIDIA released Cosmos 3 Edge at SIGGRAPH 2026 — a 4B open world model for ...
GLM 5.2: Open-Weight Coding Model Beats GPT-5.5 at 1/6 the Cost
GLM-5.2 from Z.ai scores 62.1 on SWE-bench Pro, beating GPT-5.5, under MIT license at $0.95/M ...
Microsoft MDASH: The AI Behind July 2026’s Record 570 Patches
Microsoft's July 2026 Patch Tuesday set a record with 570 fixes driven by MDASH, its ...
Kimi K3: What 2.8 Trillion Parameters Actually Buys You
Kimi K3 launches as the largest open-weight model ever. Weights drop July 27. Here is ...
Databricks Omnigent: One Control Plane for Every AI Agent You Run
Databricks Omnigent is an open-source meta-harness that sits above Claude Code, Codex, and Cursor — enforcing agent policies at the infrastructure layer, not ...
OpenAI Presence: What Enterprise Voice Agents Cost You
OpenAI Presence launched July 22 with guardrails, Codex-driven improvement, and 75% call resolution. Here's what ...
Anthropic’s Defending Code Harness: AI Vulnerability Scanning
Anthropic open-sourced a complete autonomous vulnerability scanning pipeline that has already disclosed 1,596 vulnerabilities in ...
BaseRT: Run Local LLMs on Apple Silicon 6x Faster
BaseRT runs local LLMs directly on Apple Metal GPU API, beating llama.cpp by up to ...
Gemini 3.5 Pro Keeps Missing: Google’s Third Launch Failure
Gemini 3.5 Pro has now missed three launch deadlines. Here is what went wrong, why ...