GitHub Next

What is GitHub Next?

GitHub Next is a team at GitHub dedicated to exploring the future of software development. They focus on prototyping tools and technologies that push the boundaries of what's possible, aiming to revolutionize the software development process. This includes identifying innovative strategies for creating thriving and highly efficient software engineering teams.

How does GitHub Next work?

GitHub Next operates through several key activities:

• Research: Conducting in-depth investigations into emerging trends, challenges, and opportunities in software development.
• Prototyping: Developing experimental tools and technologies that address identified needs and explore new paradigms.
• Collaboration: Engaging with the broader software development community through events, publications, and open-source contributions.

Key Projects and Research Areas

GitHub Next is actively involved in a range of research areas and projects, including:

AI for Code

• Towards Natural-Language Programming for GitHub Actions (Project Copernicus): Exploring the use of natural language to program GitHub Actions.
• Exploring LLM-powered navigation for your codebase (Project Copernicus): Utilizing LLMs to improve code navigation.
• Exploring LLM-powered automation in platform-based software collaboration (Continuous AI): Investigating the automation possibilities offered by LLMs in software collaboration.
• Agentic Setup, Build, and Testing of Repositories (Discovery Agent): Creating automated processes for setting up, building, and testing repositories.
• Can we enable anyone to create or adapt software for themselves, using AI and a fully-managed runtime? (GitHub Spark): Empowering individuals to create or adapt software with the help of AI and a managed runtime environment.
• An innovative superfamily of fonts for code (Monaspace): Designing specialized fonts to enhance code expressiveness in any editor.

Data Visualization

• Can we derive personalized design systems from sources of inspiration? (Mosaic): Aiming to generate personalized design systems using inspiration sources.
• How can we “fingerprint” a codebase to see its structure at a glance? (Visualizing a Codebase): Exploring methods to automatically visualize GitHub repositories for quick structural understanding.

Previous Research Areas

• An exploration of a new category of assists for using natural language in software development (Extract, Edit, Apply): Examining how natural language can aid in software development.
• Can we make it easy and fun to learn as we build? (Learning Sandbox): Creating personalized, interactive learning environments within the development workflow.
• Live notebooks in VS Code for JavaScript/TypeScript, web development, and AI experimentation (Vitale): Providing live notebooks for web development and AI experimentation.
• An agentic dev environment, designed for everyday tasks (Copilot Workspace): Designing an agent-assisted development environment.
• Can we improve Copilot code completion by suggesting the next logical change, wherever it is in your project? (Copilot Next Edit Suggestions): Enhancing code completion by suggesting logical changes within the project.
• Multiplayer collaboration for your whole repo (Realtime GitHub): Exploring multi-user collaboration for entire repositories.
• Can we develop software entirely in natural language, and an AI-powered toolchain manage the implementation? (SpecLang): Investigating the use of natural language for software development with AI-driven toolchains.
• How can we make LLM responses more robust and easier to understand by combining their fluid reasoning with rigid structure? (Code Atlas): Improving the reliability and interpretability of LLM responses by integrating reasoning with structure.
• An exploration of using calculation generation to improve GPT-4's capabilities for numeric reasoning (GPT-4 with Calc): Enhancing GPT-4's numeric reasoning through calculation generation.
• How would it feel to have an expert on hand at all times? (Copilot for Docs): Developing a tool to provide expert-level documentation assistance.
• Pull requests are a central part of the GitHub user experience. (Copilot for Pull Requests): Improving pull request descriptions and team reviews using Copilot.
• Ever having trouble remembering that shell command or this obscure flag? (Copilot Completions in the CLI): Integrating GitHub Copilot assistance into the terminal.
• Write code without the keyboard. (Copilot Voice): Enabling voice-based coding with GitHub Copilot.
• Can editing code feel more tactile, like painting with Photoshop brushes? (Code Brushes): Introducing tools for modifying code with a Photoshop-like brush interface.
• Reimagine repositories with custom, interactive blocks. (GitHub Blocks): Enhancing repositories with custom, interactive blocks for better documentation and workflows.
• As we increasingly work together remotely, how might we unify our workflows to enable remote collaboration for developers? (Collaborative Workspaces): Exploring ways to unify workflows for remote developer collaboration.
• GitHub Copilot Radar is a code navigation tool that shows developers the most relevant line to their position in the code. (GitHub Copilot Radar): Providing enhanced code navigation with context-aware suggestions.
• We want to let Copilot see the code in your entire repo to make suggestions even better. (GitHub Copilot for Your Codebase): Enhancing Copilot's suggestions by allowing it to access the entire codebase.
• Faster feedback about security vulnerabilities on your PRs. (Incremental CodeQL): Offering quicker security feedback on pull requests.
• TestPilot helps you craft readable unit tests based on your code and documentation. (TestPilot): Assisting in the creation of readable unit tests based on code and documentation.
• A VS Code extension for experimental applications of GitHub Copilot. (GitHub Copilot Labs): Providing a VS Code extension for GitHub Copilot experiments.
• An ongoing collaboration with Microsoft to bring React to their library Webview UI Toolkit for Visual Studio Code (React Webview UI Toolkit for VS Code): Integrating React into the Visual Studio Code Webview UI Toolkit.
• A new way to build software (GitHub Copilot): Introducing a novel approach to software development.
• Flat explores how to make it easy to work with data in git and GitHub (Flat Data): Simplifying data handling in Git and GitHub with versioning.

Who is GitHub Next for?

GitHub Next is beneficial for:

• Software Developers: Professionals seeking to enhance their development processes and workflows.
• Researchers: Individuals exploring the future of software development and AI-driven tools.
• Engineering Teams: Teams looking to optimize collaboration, productivity, and code quality.

Why choose GitHub Next?

GitHub Next stands out because it:

• Focuses on Innovation: Dedicated to exploring and prototyping cutting-edge technologies.
• Offers Practical Solutions: Provides tools and strategies to improve software development practices.
• Encourages Collaboration: Fosters a community-driven approach to innovation in software development.