Browse All Posts (248)

sukkaur announces a preview of SQL database skills for Skills for Fabric, aimed at giving AI coding agents workload-specific guidance to create, connect to, query, and manage SQL databases as part of end-to-end Microsoft Fabric solutions.

Bob Ward and Anna Hoffman walk through improvements to vector indexing in Microsoft SQL, focusing on how vector search capabilities have evolved since the initial release and where to find official docs and sample code to try it yourself.

Joshua Yue introduces new experimental routing and failover primitives in Microsoft.Extensions.AI, showing how .NET apps can route chat requests across multiple model/providers, do content-based semantic routing with embeddings, and fail over safely when a client fails—while keeping request options and streaming behavior predictable.

Allison announces an update to GitHub’s dependency graph license detection that prioritizes canonical package registries (like npmjs.org and PyPI) to improve license coverage across dependency insights, SBOMs, GitHub Advanced Security license compliance, and dependency review.

Allison announces a GitHub UI improvement that lets maintainers block or unblock users directly from pull request and issue comments in personal-account repositories, helping reduce spam and unwanted activity without leaving the conversation thread.

Visual Studio Code shows how the /btw command in the VS Code Agents window opens a side chat that shares context with your main chat, so you can ask quick questions without derailing your current task.

meghasony announces a preview feature that lets dbt jobs run directly against Microsoft Fabric Lakehouse, bringing transformation and testing workflows closer to OneLake-backed data. The post explains the Lakehouse-first use cases and the exact configuration steps to connect an existing or new Lakehouse target.

Lily Du explains how to build Microsoft Teams agents that participate in busy chats and channels without adding noise, using lightweight interaction patterns like emoji reactions, threaded replies, and quoted replies.

Peter Steinberger argues that the current bottleneck in AI isn’t model capability but human imagination, and that meaningful progress comes from applying existing intelligence creatively—especially in open source communities where maintainers continuously fix problems and invent new solutions.

Gregg Cochran shares lessons from Session 4 of the GitHub Secure Open Source Fund, where 50 projects used GitHub security tooling, expert guidance, and AI-assisted workflows to make measurable security improvements—covering incident response, supply chain risks, and how maintainers can use tools like GitHub Copilot for triage and remediation.

Tom Kerkhove introduces external dependency telemetry in Azure Application Insights for Azure API Management policies, aimed at helping platform operators pinpoint where request time is spent when policies fan out, retry, or perform downstream work that’s otherwise hard to see in standard request telemetry.

ChethanRaj asks how to structure Azure Resource Groups for an enterprise integration platform, focusing on how grouping choices affect ownership, lifecycle management, RBAC boundaries, and day-to-day operations across shared infrastructure and workload-specific components.

ChethanRaj asks how teams should structure Azure Resource Groups in real production environments, and whether boundaries should be based on technology, application/workload ownership, or resource lifecycle.

Rory Preddy demonstrates how to debug a Spring Boot Java application in Visual Studio Code by tracing a browser request through the service layer, pausing at breakpoints, inspecting variables, and using the Spring Boot Dashboard to review endpoints, health, and heap memory.

John Savill gives a quick, practical overview of tokens and tokenization in LLMs, how models generate tokens, why context windows matter, and how token usage affects both user experience and cost.

Allison announces that Gemini 3.7 Flash is rolling out as a selectable model in GitHub Copilot, with noted improvements in web/app development and agentic coding workflows, plus better code quality, codebase research, and verification for complex tasks.

Fokko at Work explains how the Visual Studio Code Agents Window works and how it uses Git worktrees to let you run multiple coding agent sessions (including GitHub Copilot, OpenAI Codex, and Claude Code) in the same repo or across multiple repos without clobbering local changes.

ceciliaxia outlines how Microsoft is co-designing Azure AI infrastructure end-to-end—from accelerators and CPUs to rack power, cooling, and networking—to improve performance per dollar and per watt for large-scale inference and agentic workloads, and shares related open standards work with the Open Compute Project ecosystem.

Pamela Fox walks through practical MCP server designs for giving coding agents safe access to a PostgreSQL database, from free-form SQL to fully templated tools. She highlights where flexibility creates risk, and shows concrete guardrails like schema discovery patterns, read-only enforcement, and confirmation flows for destructive actions.

Authorised Territory demonstrates how to configure GitHub Copilot in Visual Studio 2026 (18.9.0), focusing on adding Ollama-backed local models and tuning the model “thinking effort” setting to control how much reasoning the model applies during responses.

Visual Studio Code shows how VS Code dictation can detect your language and convert speech to text using a local model, and where you can use it across the editor experience.

Anirudh Garg explains a practical pattern for adding AI to existing event-driven Azure Functions workflows using the Serverless Agents runtime, keeping queueing and retries deterministic while using a model only for the “fuzzy” step (like parsing free-text expense requests into structured decisions).

Jakub Oleksy summarizes GitHub’s July 2026 reliability work and eight major incidents, including outages affecting GitHub Actions, APIs, and authentication. The report shares root causes (DNS automation, certificate expiry, Vitess vschema changes, Redis maintenance issues) and the mitigations GitHub is rolling out while accelerating more production traffic onto Azure.

daisami shows how to combine Microsoft Agent Framework Harness with the Agent Governance Toolkit (AGT) to enforce tool-call policy for a file-access agent. The walkthrough uses a YAML allow/deny policy, function-call middleware, and audit events so risky operations (write/delete/replace) are blocked at execution time, not just in prompts.

pankajarora announces general availability of Capacity Overview Events in Microsoft Fabric Real-Time Hub, focused on monitoring capacity conditions and reacting to them with near real-time alerts and automated actions.

Kayla Cinnamon walks through the basics of getting started with the GitHub Copilot app, focusing on how to write an effective first prompt, attach the right project context, and iterate on results as Copilot works through a task.

Peyton Fraser, Krishnakumar Muthukrishnan, and Joe Filcik announce August 2026 updates to Azure Content Understanding, including a refreshed CU 1.0 GA API and a CU 2.0 public preview. The post covers GPT-5 series support, lower token usage, improved grounding and confidence scoring, new synchronous Read/Layout APIs, semantic chunking, and agentic document reasoning.

Joe Filcik and Krishnakumar Muthukrishnan explain how Azure Content Understanding now supports the GPT-5 model series and what that changes for document, image, video, and speech workloads, including practical guidance on model selection plus updated grounding and confidence scoring aimed at reducing token usage and improving output reliability.

Allison announces general availability of Agent Plugins 1.0 across VS Code, Copilot CLI, the GitHub Copilot SDK, and the GitHub Copilot app, explaining how the open standard packages skills plus MCP server configuration into a single portable plugin and how enterprises can govern plugin and marketplace access.

Allison announces an organization-level Rule insights dashboard in public preview, giving teams an aggregated view of how GitHub repository rulesets are evaluated and enforced across all repos, with filtering and CSV export to support governance, compliance reporting, and incident/audit reviews.

Andrea Griffiths shares practical patterns from AutoGPT maintainer Nicholas Tindle for handling AI-generated pull requests without drowning maintainers, using repo-scoped instruction files and hard gates that push agents to follow your project’s rules.

sbaynes describes CARE-X, a Microsoft Research chest X-ray vision-language model that combines report generation with calibrated structured predictions and spatial grounding. The post explains the model architecture, a DAPO reinforcement-learning alignment stage, benchmark results, and a separate tool-augmented measurement pipeline for measurement-dependent findings like cardiomegaly.

Wendy Breiding explains how to keep GitHub Copilot (and other frontier coding models) effective by treating repository instructions as a high-signal budget: document the local facts, constraints, and validation commands the model can’t infer, and remove vague “prompt folklore” that adds noise.

Aditya Badmaraju discusses how to design application architecture that can handle 10x growth, focusing on making early decisions that won’t force a full rebuild later and using cloud platform capabilities (including Azure SQL) to evolve as demand increases.

Priya Sathy explains why, for budget-constrained startups building AI applications, database and data-access decisions (including Azure SQL) can have a bigger impact on reliability and cost than over-optimizing model selection.

Tina Schuchman lays out a practical FinOps-style approach to controlling AI agent spend in Microsoft Foundry, focusing on token drivers, cost attribution, and the runtime/workflow/governance levers teams can use to keep pilots from turning into runaway production bills.

ArshadAliTMMBA announces the general availability of Fabric Runtime 2.0, the Apache Spark-based execution foundation for Microsoft Fabric’s Data Engineering and Data Science experiences, positioned as a modern, high-performance runtime integrated across the Fabric ecosystem.

Rory Preddy shows how to go from a folder of Java files to a running Spring Boot web app inside Visual Studio Code, including setting up the Java and Spring Boot tooling, building with Maven, running from the Spring Boot Dashboard, and testing a simple end-to-end workflow.

John Edward breaks down Microsoft’s August 2026 Windows security update, focusing on the phased Secure Boot certificate rollout, servicing stack improvements, and what admins and users should expect during installation (including a possible extra restart) across Windows 11, Windows 10, and Windows Server.

The Visual Studio Code Team shares incremental updates in the VS Code 1.134 Insiders changelog, including improvements to Voice Mode onboarding and chat UX, new editor association behavior for opening HTML in the integrated browser, and early support for running Codex sessions via the agent host with external model providers.