Browse All Posts (152)
sbaynes describes how the Microsoft Discovery app can help scientists and data teams design an inspectable, reproducible bioinformatics workflow for phage therapy research, from profiling antibiotic resistance and bacterial defenses to searching environmental metagenomic data and ranking candidate phages with traceable evidence and validation steps.
Kristen Womack recaps the September 2026 Azure Developer CLI (azd) releases, covering extension dependency handling, versioned gRPC extension contracts, new azure.yaml project layering, per-phase concurrency controls for provisioning and deployment, and authentication host connectivity via local socket transports.
Microsoft Developer closes out the Fly with Copilot Redmond event with a recap of the main ideas and innovations, plus pointers to resources and concrete next steps attendees can use to keep learning after the event.
Microsoft Developer introduces Microsoft Copilot Studio fundamentals for building custom agents, including how instructions, knowledge, tools, skills, memory, and MCP fit together, plus an overview of the new GitHub Copilot harness for reasoning-heavy, multi-step agent scenarios.
Microsoft Developer shows how to use the GitHub Copilot App to build a personal assistant by describing requirements in natural language and collaborating with Copilot to create an application, add features, test, and refine the solution using agentic development workflows.
Dan Hellem shares recent updates to Copilot Code Review for Azure Repos, including configurable effort levels, faster comment resolution when applying suggestions, and new Azure DevOps audit log events to track enablement and configuration changes across organizations, projects, and repositories.
Allison announces that GitHub Team organizations can now start self-serve trials of GitHub Advanced Security, making it easier to evaluate GitHub Code Security and GitHub Secret Protection from common org admin entry points.
Microsoft Developer shares a short announcement for an October 8, 2026 session focused on getting started with Copilot Studio, including what to learn first and how Agent Academy can help build practical AI builder skills.
Cassidy and GPS talk with Camilla Moraes from GitHub’s Maintainer Love team about the “AI slop” problem hitting open source repos, and how GitHub is building small, high-impact features like PR limits to reduce spam while using Copilot CLI to help organize maintainer feedback.
The Visual Studio Code Team shares the VS Code 1.141 (Insiders) release notes, including a new Copilot Chat command to clean up inactive agent worktrees to help keep agent-driven workflows tidy.
Allison announces that the HydraFusion research preview is now available in Visual Studio Code and the GitHub Copilot app, bringing multi-model orchestration to more Copilot surfaces. The post explains HydraFusion’s workflow options (Single, Cascade, Critique), what’s improved, and how to enable it in VS Code and the Copilot app.
Mark Russinovich explains why trust is the real bottleneck to getting value from AI, and why humans remain accountable for what AI systems do, even when the models are highly capable.
Microsoft Security Research, Mahesh Mandava and Rajesh Kumar Natarajan break down real-world exploitation of CVE-2026-73570 in Zimbra, including the SNMP command-injection path, post-exploitation tradecraft (web shells, persistence, privilege escalation), and concrete hunting and mitigation guidance using Microsoft Defender and Defender XDR.
Allison announces a TLS compatibility change for GitHub Enterprise Cloud with data residency: starting October 7, 2026, clients that only offer X25519 for key agreement will no longer be able to connect over HTTPS. The post explains what’s changing, who might be impacted, and what to update or reconfigure.
Hidde de Smet explains why “agent-done” work can still create downstream toil, and how GitHub’s Agentic Engineering System (AES) helps teams make delegation reviewable and observable. The article proposes three practical handoffs—evidence packets, enforced stopping points, and release receipts—to reduce “handoff debt” and keep agent output aligned with team capacity.
Ben Brauer talks with Eric Imasogie about the Microsoft Learn MCP Server and how it helps AI coding assistants validate answers against current, official Microsoft documentation via the Model Context Protocol (MCP).
Robin_Lester introduces TypeSafe AI’s Jev “System One” model and shows how to use it as a structured judgement layer inside Microsoft Foundry agent workflows, including an OpenAPI tool setup and an example triage flow that returns typed decisions with probabilities and confidence.
varghesejoji introduces a local “Assessment & Optimization Workbench” that pulls together Azure billing and Databricks workspace evidence so teams can review Databricks cost drivers, utilization, job health, and selected posture checks. It walks through the UI workflow (configure, validate, run, visualize, export) and how to build and run it locally.
Wiliam Rosa explains how to use Azure Databricks ai_query to run a “decision model” (closed-set classification with probabilities) directly from SQL against Unity Catalog-governed tables, avoiding chat-style outputs and extra orchestration. The post covers required warehouse/runtime constraints, structured return types, hosting options, and governance caveats in Unity Gateway.
Obinna Nwokolo announces that Azure savings plan and reservation (including P3) recommendations can now be scoped to management groups in the Azure portal and via REST APIs, enabling aggregated, hierarchy-wide commitment recommendations across multi-subscription estates.
Microsoft Developer explains how Azure AI Foundry’s Toolbox helps AI agents use tools safely, separating read-only actions from operations that require approval. Using an MCP-governed toolset, the video shows how to centralize tool definitions, control permissions for sensitive actions like refunds, and keep token costs stable as the toolbox grows.
Microsoft Security Research, Parasharan Raghavan, Deva Kanna Kannan and Sai Chakri break down a July 2026 phishing campaign that delivered a legitimate MSP360 RMM installer under deceptive filenames, then used it to deploy ScreenConnect for redundant remote access. The post includes detection coverage, mitigations, and KQL hunting queries.
Chu Lahlou and Cha Zhang explain why enterprise GenAI systems still need a dedicated content-extraction layer, even with stronger LLMs. They break down the operational gaps of “LLM-only” extraction and position Azure Document Intelligence and Azure Content Understanding as complementary tools for grounded, auditable, multimodal workflows.
Justin Bettencourt rounds up the September 2026 Azure SDK releases across .NET, Java, JavaScript/TypeScript, Python, Go, Rust, and C++, calling out notable updates like Azure AI Search preview API support, a stable Rust SAS crate, and new Entra ID authentication libraries for PostgreSQL.
Dalibor Kovacevic explains how Azure SRE Agent plugin marketplaces can now be backed by Azure DevOps Repos, so teams can keep runbooks and tooling in their existing Azure DevOps Git repositories instead of mirroring to GitHub, using managed identity or PAT-based access.
Allison announces new repository-level controls for where Dependabot version and security update jobs run, including runner type selection plus optional custom labels and runner groups for self-hosted or larger GitHub-hosted runners.
Allison announces a public preview feature that lets enterprises sync repository business context from an external system of record into GitHub using external custom properties, keeping GitHub values read-only and continuously updated via APIs.
Matt Hernandez introduces Azure Canvases for GitHub Copilot, a shared workspace that pairs Copilot chat with interactive dashboards and guided workflows for common Azure tasks like deploying Functions, querying resources across subscriptions, and reviewing cost and governance signals.
emiller635 compares three practical ways to evaluate and observe AI agents: self-hosted Langfuse on an Azure VM, a Microsoft Foundry hosted agent, and a standalone agent on Azure Container Apps. The post breaks down what each option gives you for tracing, scoring, and dashboards, and what you still have to build and operate yourself.
Microsoft Developer shares an Azure Essentials Show conversation on estimating and controlling the cost of AI workloads on Azure, including what drives spend (models, compute, storage, and usage patterns) and how to use Azure pricing and governance tools to avoid surprises as you scale.
sarah-zhou announces the general availability of Azure Lasv5 and Laosv5 storage-optimized virtual machines, highlighting major gains in local NVMe capacity, IOPS, and networking versus the prior Lasv4/Laosv4 generation, plus sizing guidance and regional availability for data-intensive workloads.
Allison announces that OpenAI’s GPT-6.1 Sol is generally available and rolling out in GitHub Copilot, highlighting improved multistep coding performance and more efficient token usage. The post also covers where the model can be selected, how admins can control access, and how it’s billed under usage-based pricing.
Raj Pochiraju announces general availability of SQL Server on Azure Local for both connected and disconnected environments, aimed at sovereign, regulated, and edge scenarios. The post explains deployment options with Azure Arc management, and how Foundry Local can run AI inference next to SQL Server while keeping data and processing on-premises.
Fernando de Oliveira, Fernanda Lasmar, and Denise Mignoli explain how to package and distribute shared AI-assisted engineering workflows using GitHub Copilot plugins, so teams can reuse agent instructions, skills, and MCP tool connections while still keeping application-specific context and decisions inside each repo.
James Montemagno hosts a VS Code Live release recap with members of the VS Code and GitHub Copilot teams, highlighting recent feature updates and what they mean for day-to-day developer workflows in the editor.
sbaynes introduces Quine, a Microsoft Research system that combines a multimodal “world model” of biology with an interactive harness that connects models, scientific tools, literature, and wet-lab workflows to speed up biological discovery and hypothesis testing.
Mark Downie outlines the September Visual Studio update, including Bring Your Own Model (BYOM) for Copilot-powered agent mode, one-click fixes for NuGet Audit vulnerability warnings from the Error List, a Git agent for pull request exploration, improved C# debugging hints for compound if conditions, and Podman container attach support.
GeertVanTeylingen shares updated, independently published SPECstorage Solution 2020 EDA_BLENDED benchmark results showing higher single-volume performance for Azure NetApp Files large volume breakthrough mode, plus what the now generally available cache volumes change for hybrid EDA workflows where data remains on-premises.
Microsoft Defender Experts Cybersecurity Incident Response (DART) breaks down how threat actor Storm-3068 used a compromised identity to pivot into Azure DevOps, tamper with trusted pipelines, and harvest Kubernetes credentials, plus practical steps to harden identity controls and DevOps pipelines against similar attacks.
Microsoft Threat Intelligence breaks down how Star Blizzard’s 2026 campaigns evolved toward higher-volume phishing and a new “RedFlick” delivery chain, including scheduled-task persistence, VHDX/LNK/MSI stages, and CosmicPulse backdoor deployment, plus concrete mitigations, Defender detections, hunting queries, and indicators of compromise.