Microsoft introduced run-assert-eval, a VS Code skill that discovers risks for a given AI agent, measures failure rates, generates runtime policy from those findings, and reruns the same evaluation to…
Microsoft's PyRIT (Python Risk Identification Tool), launched in early 2024, has matured into a production-grade open-source AI red teaming platform used to red team hundreds of generative AI products…
Microsoft's CoreAI organization, which includes Azure Developer Services and Microsoft Foundry, is addressing the challenge of making independently built portals feel like a cohesive ecosystem despite…
Microsoft released two open-source projects, ASSERT and Agent Control Specification (ACS), to help developers evaluate, control, and optimize AI agent behavior in production. In a banking support scen…
Microsoft's Azure SRE Agent, after a year in production, has been rebuilt to move safety controls outside the agent's execution environment rather than restricting the agent itself, after incidents wh…
Microsoft Technical Fellow Anders Hejlsberg discussed the TypeScript 7.0 native port to Go on The Peterman Pod with host Ryan Peterman, noting the new compiler delivers 8x to 12x speed increases on fu…
TokenOps, an open-source system from The Agent Plane, provides real-time, run-scoped cost control for AI agents, addressing the problem of runaway token spend across multi-call runs. Unlike traditiona…
Developer asklar released Agent Terrarium, an open-source desktop app that visualizes AI agents as animated characters in a simulated world, built with a Rust simulation engine and a React view layer.…
Microsoft Foundry's Toolboxes introduce a tool search capability that reduces token costs and context bloat by exposing only two meta-tools—tool_search and call_tool—instead of loading every tool on e…
Microsoft hired Taesoo Kim, former Georgia Tech professor and leader of Team Atlanta, as VP of Security Research to head the new Microsoft Security FORGE Labs after his team won DARPA's AIxCC competit…
A developer built a system called a "dev machine" that autonomously created a proof-of-concept version of Microsoft Word over 39 days, producing 350,000 lines of TypeScript across 3,706 commits. The m…
A real Node.js team using LangGraph and TypeScript needed to integrate Microsoft's Squad agent framework without rewriting their product in C#, leading to a three-layer architecture where a LangGraph …
Microsoft, Hugging Face, Meta's PyTorch team, NVIDIA, and others launched OpenEnv, an open protocol for agent learning environments that standardizes how agents practice and improve. The protocol aims…
Microsoft, in collaboration with Cisco, Databricks, GitHub, GoDaddy, Google, Hugging Face, Nvidia, Salesforce, ServiceNow, and Snowflake, introduced the Agentic Resource Discovery (ARD) specification,…
Microsoft researchers propose using information-flow control (IFC) to secure autonomous AI agents against prompt injection and data exfiltration. IFC labels data with integrity and confidentiality tag…
Microsoft Chairman and CEO Satya Nadella joined podcast hosts swyx, Sarah Guo, and Elad Gil for a crossover episode of No Priors and Latent Space following his Build keynote. The discussion focused on…
Microsoft launched Command Line, a new blog focused on how the company builds and operates AI systems, coinciding with Build 2026 in San Francisco. The blog introduces 10 new rules for software develo…
Microsoft employee Omar Shahine built a personal AI assistant called Lobster in his spare time using OpenClaw technology, which he later presented to Microsoft's AI Accelerator group, landing him a ne…
Microsoft has built an optimization loop for AI agents that automates the process of fixing quality failures without causing regressions, addressing a key operational challenge teams face when running…
Microsoft introduced Web IQ, a new grounding system for AI agents, at its Build conference, claiming sub-165ms P95 latency and industry-leading quality for connecting AI models to fresh, authoritative…