Key Moments
GitHub’s Growth: 225M Developers, 34.3B Commits — Jay Parikh, EVP of CoreAI, Microsoft
Latent Space PodcastKey Moments
GitHub's growth is staggering, with 225M developers and 34.3B commits, but the real challenge is scaling for AI agents, which are rapidly increasing and already comprise 40% of pull requests.
Key Insights
GitHub has reached 225 million developers, with one new developer joining every second for the past two years.
40% of pull requests on GitHub are now automated via bots, a number that has increased sixfold in six months.
The number of open-source contributions on GitHub has increased by two and a half times, with 2.5 billion contributions on the platform.
Copilot Fusion is designed to orchestrate multiple AI models, aiming to deliver the best quality code at a lower cost.
GitHub is in the process of migrating its entire infrastructure to Azure, with the goal of completing this in early next year.
GitHub is enhancing security with AI-powered scanning and Code Review features to detect more complex errors and vulnerabilities.
Unprecedented developer growth and platform scale
GitHub has experienced exponential growth, now boasting 225 million developers worldwide. This figure represents a massive increase from previous years, with the platform adding approximately one new developer every second for the past two consecutive years. To put this scale into perspective, the platform handled 17.9 billion software pushes, 34.3 billion software deliveries, and 1.3 billion pull requests in the past year alone. The open-source community on GitHub has also seen a significant surge, with contributions doubling to 2.5 billion on the platform. This rapid expansion underscores GitHub's central role in global software development and collaboration, presenting a constant challenge to maintain reliability and performance under such immense load.
The rise of AI agents and automated workflows
A striking shift in GitHub's usage is the dramatic increase in automation. Currently, 40% of all pull requests on GitHub are generated by bots, a figure that has multiplied sixfold in just the last six months. This indicates a fundamental change in how developers interact with the platform, leveraging AI and automation for a significant portion of their workflow. This trend is reshaping the developer experience, with AI agents becoming integral to the development process. The implications are vast, ranging from increased efficiency to new paradigms for code creation and collaboration. The challenge for GitHub now is to scale its infrastructure and services to accommodate this burgeoning agent workload effectively.
GitHub as a social network for software
Jay Parikh draws parallels between GitHub and traditional social networks, highlighting its function as a 'social network for software.' Historically, platforms like Ning and Facebook, where he has extensive experience, share core principles of connecting users and facilitating interaction. However, GitHub's unique characteristic is its focus on developers and the collaborative creation of software. The platform's evolution, particularly with the integration of AI, is expanding its accessibility beyond highly skilled software engineers. Even individuals without deep coding expertise can now leverage AI agents to contribute and share ideas, fostering a more inclusive development environment. This democratization of development, enabled by AI, means that the community can now encompass a much broader range of users, reinforcing GitHub's role as a central hub for collaborative innovation.
Adapting to evolving development paradigms
The platform's flexibility is key to accommodating diverse developer needs and evolving workflows. Parikh acknowledges that traditional concepts like pull requests may need to be re-evaluated in light of AI agents. Some developers, like those from Open Claw, push GitHub's limits, proposing changes that reflect new ways of working. GitHub's strategy is to remain adaptive, allowing for modification and customization of features based on user feedback and technological advancements. Whether a developer prefers a strict pull request workflow or a more fluid, continuous building approach without formal requests, GitHub aims to support both. This means the platform must support a multitude of concepts and enable users to assemble them in ways that best suit their individual or team needs.
The expanding definition of code and repositories
The nature of what constitutes 'code' and a 'repository' is also evolving. Parikh points to examples like a Markdown file being stored and versioned on GitHub, challenging the traditional view of code as solely executable programs. With the advent of AI, the distinction between different programming languages (e.g., Rust, Java, .NET) becomes less critical as tools can abstract away some of the complexities. The focus is shifting towards storing and managing a broader spectrum of development artifacts, including documentation, configuration files, and AI-generated assets. This expanded definition requires infrastructure that can handle diverse file types and metadata, supporting decades-old codebases alongside the latest AI-driven projects.
Copilot Fusion: Orchestrating AI models for optimal results
Copilot Fusion represents a significant advancement in GitHub's AI strategy, acting as a comprehensive abstraction layer for AI capabilities. Building on the foundation of Copilot, Fusion's core function is to maximize the potential of AI models by intelligently orchestrating them. This involves guiding models to achieve specific tasks, delivering high-quality code efficiently and at a reduced cost. The platform supports multiple models and allows developers to choose the applications they want to use, adapting to the rapidly changing landscape of AI tools and techniques. Fusion aims to alleviate the cognitive load on engineers by abstracting away the complexities of selecting and integrating different models with varying costs and performance characteristics. This capability is crucial for navigating the burgeoning AI ecosystem and ensuring developers can leverage the best available tools for their projects.
Unified Copilot experience across platforms
A key innovation is the consistent availability of Copilot across all platforms, including mobile, desktop applications, and the GitHub.com web interface. This unified experience is built on shared infrastructure, allowing for seamless transitions between devices. For instance, a developer can start a task on their phone, continue it on their desktop, and have all their context—repositories, issues, security findings, preferences—carried over. This persistent, personalized AI engine ensures that developers don't lose their work or context when switching environments or tools. This cross-platform integration is crucial for maintaining productivity and enabling developers to work efficiently, regardless of their location or device.
Strengthening security in an AI-driven development world
Security remains a paramount concern, especially with the increasing reliance on AI and open-source components. GitHub is investing heavily in features like Code Review Max to identify complex and subtle bugs, and AI-powered scanning within GitHub Advanced Security to detect vulnerabilities. A demonstration showed how AI could prevent a secret from being accidentally committed. Deeper scanning capabilities are being developed to go beyond surface-level checks. Furthermore, there's significant collaboration between Microsoft and GitHub security teams, and engagement with the open-source community, to reduce risks across the software supply chain. Efforts are underway to address vulnerabilities in package managers like NPM and to fortify the security of open-source projects, reinforcing the commitment to making open-source software safer.
The future roadmap: Azure migration, infrastructure evolution, and cost management
Looking ahead, GitHub's strategic priorities include completing its migration to Azure in early next year, a massive undertaking to support its continued scaling. The company plans to share more insights into this infrastructure evolution, highlighting the engineering challenges and solutions involved. They will continue to enhance platform services, enabling developers to leverage core building blocks and assemble them according to their needs. Security will remain a top focus, alongside features for cost management and governance. For enterprises, GitHub will offer controls to monitor spending, optimize model usage, and set spending limits. The goal is to provide developers with tools that not only boost productivity and collaboration but also offer visibility and control over resource utilization and associated costs, ensuring a more efficient and secure development future.
Mentioned in This Episode
●Products
●Software & Apps
●Companies
●Organizations
●People Referenced
GitHub Copilot and Development Best Practices
Practical takeaways from this episode
Do This
Avoid This
Common Questions
As of the interview, GitHub has reached 225 million developers worldwide. For the second consecutive year, a new developer has joined the platform every second, highlighting its rapid and sustained growth.
Topics
Mentioned in this video
An AI-powered programming assistant developed by GitHub, aiming to enhance developer productivity by suggesting code and automating tasks.
A suite of security features offered by GitHub to help organizations find and fix vulnerabilities in their code.
A software framework whose code is actively being updated and maintained on GitHub.
A popular source-code editor developed by Microsoft, integrated with GitHub's services and Copilot.
A programming language mentioned in the context of repository language statistics.
Microsoft's cloud computing platform, which GitHub is in the process of migrating to.
A cloud security company where Jay Parikh served as CEO.
An organization whose representative, Pete Steinberger, has pushed GitHub's limits and suggested innovative ideas like ending pull requests.
A social networking platform where Jay Parikh previously served as VP of Engineering.
A system mentioned in conjunction with Fusion, suggesting a potential integration for streamlining developer workflows.
A programming language whose code is actively being updated and maintained on GitHub.
A package manager for JavaScript, mentioned in the context of security challenges and delays in updates.
A social media company where Jay Parikh previously served as Head of Engineering.
A content delivery network and cloud services company where Jay Parikh previously served as VP of Engineering.
A technology company and the parent organization of GitHub. They are investing heavily in AI and cloud services, including Azure.
A system or platform mentioned as integral to maximizing AI model capabilities, particularly in guiding models and integrating with other tools like Harness.
EVP of CoreAI at Microsoft, leading many core AI services including GitHub. He has extensive experience in infrastructure and security from previous roles at Facebook, Lacework, Ning, and Akamai.
Representative from Open Claw, recognized for pushing GitHub's capabilities and suggesting innovative ideas.
Chief Operating Officer at GitHub, previously interviewed on the podcast.
CEO of Microsoft, who was a guest on the podcast to discuss the growing importance of AI model selection.
More from Latent Space
View all 277 summaries
103 minDeep Dive: Anthropic S-1 Myths and Mythos — Jamin Ball, Altimeter
51 minCooking with Reducto’s CEO: Why PDFs Still Break AI — Adit Abraham
51 minMini Brains, Beating Heart Organoids, and the AI That Predicts Them — Fabian Theis, Helmholtz Munich
73 minBehind Latent Space, ft AIEi Shanghai preview — with swyx and Galen Gao
Ask anything from this episode.
Save it, chat with it, and connect it to Claude or ChatGPT. Get cited answers from the actual content — and build your own knowledge base of every podcast and video you care about.
Get Started Free