AI News & Latest Updates - Microsoft for Developers

Additional Contributors: Justin Stoecker, Sheil Kumar Overview One of the biggest challenges in building AI/ML platform stacks for mobile, client, or server is creating a machine learning (ML) compiler that is both scalable yet performant. A common approach is to build specialized compilers for specific AI hardware, frameworks, or workload types. For example, independent hardware vendors (IHVs) provide their own frameworks to deploy models on their hardware, and PyTorch includes torch.compile as a framework-specific compiler. Other ML compiler efforts include MLIR (Multi-Level Intermediate Representation), whic...
I see three agentic stories shaping SQL’s future. Here’s how DP-800 connects the pieces and helps you build the skills to put SQL’s AI capabilities to work.
Use harness.AgentConfig to configure task lists, file memory, skills, approvals, conversation history, and tracing for agents written in Go.
.NET 11 is here with C# 15, faster developer tools, runtime and library improvements, and new capabilities for web, mobile, desktop, and AI apps.
We gave AI coding agents the same breaking upgrade twice. Labeled 2.0.0, Claude Sonnet 5.5 checked it in all 10 runs. Labeled 1.4.3, it checked in two and often left the app broken. Nearly 1,000 runs show which agents still read your version numbers, and what a patch label talks them out of.
Take the 2026 .NET Developer Survey to share your experience, help identify common challenges, and guide future .NET investments. The survey takes about 15 minutes.
Modernize your .NET apps for Windows Arm64 with the GitHub Copilot upgrade agent, from identifying migration blockers to planning and making code changes.
All five agent runs produced a working script, and nothing in the eval score looked wrong. The trajectories told a different story. Every run read a DLL as raw bytes to find plugin names, and three scripts would quietly reroute a CI runner's network traffic. Here are the patterns to look for in your own agents' trajectories, and what to change in your product when you find them.
Powered by NVIDIA RTX Spark, we are bringing the promise of unmetered local intelligence to increasingly demanding coding workloads—giving developers greater freedom to experiment, delegate, and iterate on their own hardware, without per-token cloud charges. Highly capable mixture-of-experts (MoE) models are making this opportunity tangible. These models tackle complex coding tasks spanning reasoning, implementation, testing, and debugging, while advances in efficiency are bringing more of that capability onto local devices. With 128 GB of unified memory, RTX Spark PCs enables us to ask a crucial question: how...
Agent skills can be delivered in NuGet packages as a way to deliver up-to-date library documentation to models.