November 18th, 2024

Software is a team sport: Building the future of software development together

At Microsoft, we strive to build the world’s most beloved developer tools and services. Our vision is to empower every developer to transform their concepts into reality, from idea to code to cloud, at lightning speed. The Visual Studio family, Azure, GitHub, and GitHub Copilot empower developers around the world to do just that.

Since GitHub joined Microsoft in 2018, we’ve been working hard to bring our tools and services together to deliver a productive and delightful experience for developers so they can focus on driving innovation through their code. We started by integrating features and services developers love and use every day from GitHub into our code editors. This includes creating and managing pull requests, issues, and repositories, and making it easier to deploy apps and services to Azure with GitHub Actions. We continue to innovate and collaborate to deliver an unparalleled end-to-end experience for developers and their teams.

GitHub Copilot is the most adopted AI pair programming tool today with more than 1.8 million paid developers in addition to the 1 million students, teachers, and open-source maintainers that use GitHub for free. Copilot now makes it much easier to learn a new technology or programming language, ramp up on a codebase that’s new to you, or learn how to best configure your developer workstation. Let’s take a deeper look at how we’re building, innovating, and integrating GitHub Copilot throughout the entire development lifecycle into the tools and services you love.

End-to-end AI platform showing inner loop with Visual Studio, GitHub, and Azure, Outer Loop with GitHub, and Production with Azure
AI-powered end-to-end development platform

GitHub Copilot: Code faster with the world’s most widely adopted AI developer tool

Developing Copilot has been a cross-team effort and we’ve been relentlessly focused on building and releasing innovations to enhance the developer experience. Since the early days of code completions and contextual awareness, we’ve aimed to get to the heart of what developers can use to reduce toil and distraction and replace it with joy.

In 2021, we started by bringing Copilot’s code completions into the Visual Studio family. Today, Copilot is capable of so much more than completion, which makes it an essential part of any developer’s setup. Use Copilot Chat to ask Copilot to explain, refactor, optimize, debug, document, and test your code. As you type, inline suggestions come to life, enabling you to convert comments to code, get help fixing errors, or let it write the test for the function you just wrote.

Copilot Edits in Visual Studio Code
Copilot Edits in Visual Studio Code

As we continue to innovate together—we’re creating new ways to interact with chat. In addition to submitting prompts, developers can use keywords to help Copilot better understand your prompt, leverage slash commands to avoid writing complex prompts, and incorporate chat variables to specify context. Recently, we have added new support for multimodal modals, and the ability to speak directly to Copilot.

Perhaps most excitingly, AI model selection in Copilot unlocks a new world of possibilities giving developers the freedom to explore new models and select workable options. In Visual Studio Code, developers can access models from the model catalog like Anthropic’s Claude 3.5 Sonnet, Google’s Gemini 1.5 Pro, and OpenAI’s o1-preview and o1-mini. OpenAI and Claude 3.5 Sonnet models are available now; Google’s Gemini 1.5 Pro is coming soon.

GitHub Copilot continues to evolve with new capabilities for developers. With AI-powered code review, makes feedback from pull requests immediately actionable and easy to see in the context of your code editor.  With Copilot Extensibility, extensions can interface with services like Jira, Sentry, and more from the comfort of Copilot Chat in your favorite code editor, without having to disrupt your flow or context-switch. Best yet, you can create your very own extensions that integrate with your own data, services, and workflows into Copilot.

GitHub Copilot Workspace in Visual Studio Code
GitHub Copilot Workspace in Visual Studio Code

Finally, for task-focused scenarios, we’ve built Visual Studio Code extensions for the new GitHub Copilot Workspace so developers can get straight to the task at hand, jumpstarting the process by describing what they want in a natural language. Alongside Visual Studio Code, GitHub Copilot is deeply integrated into our entire developer tools ecosystem with handcrafted experiences and features based on the work that developers need to do within the tool of their choice.

Visual Studio and GitHub Copilot

Visual Studio is our premier IDE for .NET, C++, and game developers. In Visual Studio, developers build large-scale applications with sometimes a hundred projects for their distributed system. With GitHub Copilot, developers can index and search their entire codebase for deep insight, make code edits across multiple files and projects with the upcoming Edits feature, and get advanced visualizations and debugging insights to help them solve problems.

For example, GitHub Copilot can be used with the Visual Studio profiler to identify and suggest fixes for a performance issue in the Visual Studio profiler codebase itself!

GitHub Copilot has been infused throughout Visual Studio with new features, including the ability to set breakpoints, suggest code fixes, debug exceptions and tests, migrate projects, and much more.  With Visual Studio, .NET, and GitHub Copilot, solutions for some of the hardest challenges in building software at scale are now within reach of every developer.

Working across GitHub Codespaces and Visual Studio Code

Setting up a developer workstation is an arduous task, but we’ve built a seamless connection between Visual Studio Code and GitHub Codespaces to make it a breeze. You can use your local install of Visual Studio Code or one hosted on the web to create, manage, work in, and delete Codespaces all via the GitHub Codespaces extension. Dev containers are the common format for setting up both, which helped the GitHub team reduce their dev box setup time from 45 minutes to under a minute to onboard developers.

We love how developers around the globe are using Codespaces for help with everyday development and how they’re being used to teach others to code with samples, workshops, and documentation. The near-instant coding environment has been integrated through all our training modules on Microsoft Learn and has become our standardized way to get started with learning any Microsoft technology.

Better collaboration for Dev Teams

Software is a team sport. In the last few years, dev teams using Azure DevOps have been seeking better integration with GitHub. Thousands of companies using Azure DevOps want to leverage the strength of both platforms and we’ve been working to make that possible. Developers and dev teams can now use GitHub Advanced Security for Azure DevOps and GitHub Copilot for Business within Azure DevOps. Secure your Azure Repos with advanced scanning features and harness the AI-driven coding assistance of GitHub Copilot, all without leaving your current workflow.

Streamlined Integration and Migration

We’re also focused on improving the integration between Azure Boards, Azure Pipelines, and GitHub Repositories. This work will empower dev teams to fully leverage the advanced features of Copilot and GitHub Advanced Security, while leveraging enterprise-focused project and tracking capabilities in Azure Boards, and the robust continuous delivery capabilities in Azure Pipelines. Developers can now create GitHub branches directly from Azure Boards, find Azure Boards work item links in GitHub pull requests, trigger Azure Pipeline jobs at pull request time or after changes are merged, and more.

Our goal is to provide developers using all of these products with the same end-to-end traceability they’ve come to expect from Azure DevOps. With the GitHub Enterprise Importer, developers can transition their repos from Azure repos to GitHub repos all while continuing to use Azure Boards and Azure Pipelines, and access GitHub’s powerful capabilities in AI and security.

Simplifying cloud deployments

We want to make the path to Azure easier and more intuitive for developers. To do that, we’re building better integration with GitHub.

Learn how to build and deploy to the cloud with GitHub Copilot for Azure

GitHub Copilot makes it easier to learn how to build and deploy to the cloud and how to do it without friction safely from the cockpit of where your code lives. We recently launched the preview of GitHub Copilot for Azure, which builds upon the Copilot Chat capabilities in Visual Studio Code to help developers manage resources and deploy applications. Using GitHub Copilot for Azure, developers can get personalized guidance to learn, provision, deploy, diagnose, and estimate costs for Azure services and resources without needing to leave the comfort of their personalized code editor.

Using GitHub Copilot for Azure in Visual Studio Code
Using GitHub Copilot for Azure in Visual Studio Code

Seamless Azure CI/CD integration with GitHub Actions

GitHub Actions for Azure helps you create workflows that build, test, package, release, and deploy to Azure for CI/CD integration. GitHub Actions also includes support for key utilities that can streamline testing and deployment of intelligent apps with a variety of AI models, including Azure Resource Manager templates, Azure CLI, and Azure Policy.

The Azure Developer CLI (azd) gives developers a direct path in the dev loop to provision Azure resources, deploy, and set up a CI/CD pipeline for continuous deployments. GitHub Actions workflows are now automatically generated when creating apps with azd templates and infrastructure for popular services, including Azure App Service and Static Web Apps.

Creation of GitHub Actions with azd
Creation of GitHub Actions with azd

To help start you on your AI app development journey, we’ve launched the AI app templates gallery with a curated collection of templates that can be deployed to Azure using Visual Studio Code or GitHub Codespaces. These templates are completely open source on GitHub for you to contribute your own ideas.

AI App Template gallery showing available templates
AI App Template gallery

Tools for building your own intelligent apps

The AI field is evolving rapidly, and it can be hard to keep up. GitHub Models supports secure experimentation with models, providing QuickStart code for various languages and frameworks to simplify model exploration. You can seamlessly transition from GitHub Models to Codespaces, Visual Studio Code, and Azure as you go from experiment to prototype to production deployment. If you’re not sure where to start or want to test drive, you can experiment in the GitHub playground for free and scale your AI apps to paid endpoints using Azure for secure, enterprise-level deployment and monitoring.

You can now also scale your AI applications with key insights leveraging Azure AI evaluation and online experimentation Actions (both in preview). These capabilities can be fully integrated into your CI/CD development workflows with pre-production evaluation and online experimentation results posted directly back to GitHub for analysis.

Self-provisioning GitHub from the Azure portal

Procurement of tools and infrastructure should be simple and easily accessible. You can now purchase and self-provision GitHub directly from the Azure portal, including setting up an Enterprise Managed Users configuration. Speaking of Azure, the new GitHub Enterprise Cloud with data residency is built on top of Azure, which means that you get the security, business continuity, and disaster recovery capabilities of the cloud.

Developers build our future

As we look ahead, the future of software development has never been brighter. Microsoft and GitHub are committed to empowering developers around the world to innovate, collaborate, and create solutions that’ll shape the next generation of technology. By combining our strengths, we’re paving the way for a more connected and dynamic digital landscape. We’re excited to continue this journey with you and we can’t wait to see what amazing things we’ll build together.

We’ve just launched the Microsoft for Developers blog, where we’ll strive to tell the full end-to-end story for developers as they build and infuse their applications with AI, migrate their applications to the cloud, or take advantage of the latest features and platform capabilities.

To stay up to date with the new Microsoft for Developers blog, subscribe in your favorite RSS reader or by signing up for email notifications whenever we publish a new post.

 

Author

Amanda Silver
CVP and Head of Product for Microsoft's Developer Division and GM for Microsoft's first-party engineering systems

Amanda Silver is the CVP of Product for Microsoft's Developer Division, which includes the Visual Studio family of products, .NET, TypeScript, and our developer platforms. She has been key to Microsoft's transformation to contribute to open source with the introduction of TypeScript, Visual Studio Code, and the acquisition of both Xamarin and GitHub. She believes that a tight digital feedback loop with zero distance between end-users and engineering teams is a critical element of great product ...

More about author

Mario Rodriguez
Chief Product Officer

Mario Rodriguez leads the GitHub Product team as Chief Product Officer. His core identity is being a learner and his passion is creating developer tools—so much so that he has spent the last 20 years living that mission in leadership roles across Microsoft and GitHub.

1 comment

  • Michael Taylor 1 day ago

    >> GitHub Copilot is the most adopted AI pair programming tool today

    Are you sure about that? The link given shows that ChatGPT is at 82% while Github Copilot is a distance second at 41%. That's half.

    Also it's unclear to me how useful this survey by SO even is because VS itself doesn't integrate with anything else that I'm aware of so any "AI" offerings it has would be coming from Copilot. Of course...

    Read more