Browse All Artificial Intelligence Content (852)

Allison explains how the GitHub Copilot usage metrics API now separates local BYOK and enterprise custom model usage from Copilot-hosted models. The update adds user-level and aggregate reporting without changing existing totals, with important limitations around client telemetry, token counts, model names, and historical data.

GitHub introduces using GitHub Copilot through JetBrains Air, connecting an existing account to run coding and testing tasks locally with project context or in isolated cloud environments with Air Teams. The video covers plugin configuration and explains how this option differs from the native GitHub Copilot plugin.

Jason Farmer from Microsoft's digital operations team explains how an operational technology virtual coworker runs on site and keeps working when its cloud connection drops. The video highlights local execution as part of physical AI architecture, focusing on continuity when cloud connectivity is unavailable.

Allison announces Dependabot version update support for Agent Package Manager, helping keep AI agent skills, prompts, and instructions current through pull requests. The update handles supported Git dependencies and existing lockfiles, is available across all GitHub plans, and covers version updates rather than security updates.

Richard Lander explains how NuGet packages can carry version-matched agent skills alongside library binaries. He demonstrates tools for discovering and copying those skills into repositories, compares package delivery with Git-based approaches, and discusses model evaluation, token costs, and the security implications of agents consuming package documentation.

tompeplow demonstrates how OneLake supports agent-driven data engineering through hosted MCP servers, endpoint discovery, and shortcuts. Using GitHub Copilot, the examples explore mirrored databases, document permissions, combine external data in a Lakehouse, and run SQL and Spark analytics while working within existing access permissions.

Microsoft Security Research and Jiacen Xu examine a fake-download campaign that bundles proxyware, distributed-computing software, and unwanted utilities through trojanized installers. They explain its staging and callback patterns, scheduled-task persistence, and Microsoft Defender hunting queries, with guidance for correlating evidence across an entire installer session rather than individual payload attempts.

grakhra compares Microsoft Fabric integration options for Snowflake and Databricks, explaining when to use mirroring, OneLake shortcuts, shared Iceberg storage, or pipelines. The guide distinguishes replication from in-place access and outlines how Fabric data agents connect governed analytics data to custom experiences in Copilot Studio and AI Foundry.

Yun Jung Choi explains how six specialized Azure plugins extend AI-agent workflows for AKS troubleshooting, cost analysis, Advisor recommendations, graph queries, Azure Local, and Foundry IQ knowledge bases. The guide includes example prompts, installation commands for supported clients, and approval requirements for resource changes and disruptive operations.

Alex Perrault explores how Arnis turns OpenStreetMap data into Minecraft worlds and how its sole maintainer, Louis Erbkamm, uses GitHub Copilot to investigate bugs and review contributions. Copilot sub-agents help examine sensitive parts of the codebase and suggest corner cases for regression testing.

Allison announces GitHub Copilot support in JetBrains Air, enabling local and cloud coding tasks through the Agent Client Protocol. The update supports parallel sessions and in-context review using existing Copilot subscriptions, with isolated cloud execution and limitations around plan selection, usage visibility, and administrative spending controls.

Visual Studio Code presents creating images with GitHub Copilot in VS Code. The video focuses on using the coding assistant for image creation within the editor.

Emanuele Bartolesi recaps GitHub Universe 2026 through its two keynotes, connecting Copilot’s multi-repository workflows with review, security, and Git infrastructure changes. His practical takeaways focus on verification gates, contributor guardrails, shared planning, and keeping human judgment central as teams adopt coding agents.

Jason Farmer explains how industrial AI agents connect to real equipment through cloud and edge infrastructure. The architecture combines Azure IoT services, Kubernetes, Foundry Local, and Microsoft Fabric, with an on-site virtual coworker that continues operating when cloud connectivity drops.

kinfey explains how Meridian 24 separates managed AI research and editing from deterministic audio production and publication. The implementation uses Microsoft Copilot Harness, Foundry and durable Blob storage to produce bilingual bulletins, with evidence checks, recovery checkpoints, identity permissions and measured audio timing rather than relying on a single prompt.

AsithaDeSilva explains how a prototype Azure SRE Agent workflow connects billing data, resource inventory, and optimization recommendations to investigate spending changes. The walkthrough covers reusable skills, report setup, scoped access, and validation, with practical guardrails for distinguishing projections from actual costs and cleanup candidates from resources approved for deletion.

sbaynes shares how Microsoft's Copilot Cowork team used coding agents, written specifications, automated reviews, and human checkpoints to build and migrate a live service. The article explains practical controls for parallel development, security boundaries, trustworthy tests, and a Python-to-Rust migration, including measured gains and remaining limitations.

MichalBar explains how Copilot in Microsoft Fabric generates editable Real-Time Dashboards from natural-language descriptions and Eventhouse data. The preview workflow lets users refine a dashboard plan before creation, then inspect queries and customize visuals. A stadium example connects transit arrivals with gate activity and queue lengths.

Microsoft Developer highlights a browser change that can affect runtime behavior without code changes, Zstd WAL compression for Azure Database for PostgreSQL, and the Copilot Studio Conversation Harness. The update spans browser behavior, database storage savings, and tooling for building customer-facing AI agents.

John Edward outlines the generally available GitHub app for Microsoft Teams, including Copilot task tracking, multi-repository work, and MCP connections. He also explains onboarding and administrator requirements, showing how teams can turn conversations into issues and pull requests while retaining human review of AI-generated changes.

WSilveira collects November’s Azure Logic Apps product and community updates, including a local CLI preview, AI agent orchestration, pricing allowances, and hybrid storage benchmarks. The newsletter also points developers to secure Foundry integrations, C# workflow tooling, Fabric real-time analytics, deployment walkthroughs, and BizTalk migration guidance.

Visual Studio Code shows how to move between the editor and Agents Window without losing context. This short focuses on navigating between coding and AI agent interactions within the development environment.

John Savill reviews Microsoft's October 2026 AI updates, covering foundation models, inference hardware, and agent capabilities. The overview also covers GitHub Copilot features, including session continuity and code review, alongside Azure API Management integration and broader changes to Microsoft's AI tooling.

John Edward explains how Microsoft Copilot Studio flags agents that have never been evaluated or whose models have changed since their last evaluation. The automatic Review panel warnings link makers to evaluation tools without blocking publishing, helping teams identify when previous test results may no longer reflect current agent behavior.

GitHub Agentic Workflows is generally available, letting developers compile natural-language Markdown into GitHub Actions workflows with read-only defaults, sandboxing, and validated outputs to constrain agent changes. Microsoft Foundry introduced the Copilot harness in public preview, while long-running hosted-agent resilience and browser automation reached general availability, though applications still need to save progress and prevent duplicate actions after restarts. Fabric introduced Project Osmos for persistent, permission-bound data engineering tasks and made the OneLake catalog Govern experience generally available, connecting agent execution with centralized governance. Evaluation and security findings reinforce the need to verify outcomes rather than trust plausible responses, with Microsoft's code-transformation work combining agents with testing and formal verification, and Prompt Shields demonstrations blocking malicious instructions in both user input and tool responses.

GitHub walks through the GitHub Universe 2026 badge’s built-in games, apps, and custom digital profiles. The video also shows how to use GitHub Copilot for setup help when adding community games from Ashley Wolf’s repository, including installing the classic Microsoft 3D Maze screen saver on the badge hardware.

John Edward covers the general availability of automation flow evaluations in Microsoft Copilot Studio. The capability extends structured quality assessment beyond agents to automation flows, helping makers assess expected outcomes and identify issues before production. The article also outlines repeatable evaluation practices and explains why no immediate administrative action is required.

Microsoft Developer demonstrates Intelligent Terminal running a 120-billion-parameter model locally on Surface hardware with NVIDIA RTX Spark N1X and 128 GB of unified memory. The demonstration covers completing workflows such as writing new agent skills with context, without sending requests to the cloud.

Kristen Womack reviews October’s Azure Developer CLI releases, covering deployment previews, Container Apps express support, environment value removal, and extension-aware pipelines. The recap also explains stable extension APIs, breaking Go API changes, reliability fixes, and new templates for deploying semantic search and streaming AI applications to Azure.

John Savill reviews this week's Azure updates, covering Container Apps logging, confidential VMs, network security, storage quotas, and cross-tenant backup restore. The Halloween edition also covers Azure AI Foundry model cost management, Azure MCP Server 3.0, new model topics, and GitHub Universe.

Michael MacKenzie, Microsoft’s VP of Product Management for Azure IoT and Digital Operations, explains how physical AI differs from traditional automation: automation repeats a task, while physical AI gets smarter each time it runs. This short introduces the distinction as part of a broader conversation on IoT and physical AI.

Microsoft Developer explores model selection using the Microsoft Foundry model catalog. By swapping only the model deployment and running the same evaluation on both options, the episode compares results alongside measured cost and latency to show when a smaller model is suitable for production and when it is not.

Visual Studio Code demonstrates how to use IntelliSense and GitHub Copilot to write, explain, and improve PostgreSQL queries without switching tools. The workflow also covers running queries and validating their results directly in the editor.

Waldek Mastykarz examines whether naming a workspace folder ‘eval’ changes GitHub Copilot CLI behavior. Across about 900 runs with four models, the name increased agents’ suspicion that they were being tested without producing statistically significant behavioral differences. The analysis explains the experiment’s controls, limitations, and practical guidance for evaluation harnesses.

John Edward covers the general availability of Microsoft Copilot Studio's tool accuracy grader, which evaluates whether AI agents select the correct tools and supply expected inputs. The feature adds execution-level checks to agent testing, while leaving response quality, downstream outcomes, and connected-system reliability as separate evaluation concerns.

Maggie Appleton, Martin Woodward, and Jocelyn Velarde close GitHub Universe 2026 with open-source perspectives and coding demos. The session explores learning with GitHub Copilot, scaling projects with AI agents, and collaborative AI planning, alongside creative experiments that highlight the craft of coding and building together.

Martin Woodward demonstrates how to install Super Mona on a GitHub Universe 2026 conference badge using GitHub Copilot. The demo shows copying repository instructions into Copilot Chat and letting the agent configure the app, then using the badge’s hidden capacitive buttons to play.

amitch and Shehab Amin explain how Sail connects Spark-compatible workloads to Microsoft OneLake without copying data. They demonstrate lakehouse transformations, Python-based embedding generation, and scoped access through LakeSail Cloud Sessions, while outlining compatibility checks and limitations of the reported performance and cost comparisons.

Microsoft Developer shows how to build a prompt agent in Microsoft Foundry using a model, instructions, and MCP tools, without managing containers or servers. The video explains when custom orchestration or a different framework calls for a hosted agent, while keeping name-based invocation consistent for callers.

Patrick LeBlanc joins Anna Hoffman to demonstrate Database Agents in Microsoft Fabric, which detect SQL performance problems before they affect users. The episode covers Database Hub, configuration requirements, issue detection, and SQL Server Management Studio integration, alongside a longer-term vision for agents that learn from DBA interactions.