GitHub recently published this year's Octoverse, the state of the open-source ecosystem.
Below are my key takeaways.
AI doesn’t replace developers—it brings more people into the ecosystem. A new developer joined GitHub every second in 2025. Top 5 developer populations: 1. United States, 2. India, 3. China, 4. Brazil, 5. United Kingdom.
Open source remains the foundation. Fastest-growing OSS projects by contributors include zen-browser, vscode, and AI-focused continue.
TypeScript is now the most used language on GitHub, overtook both Python and JavaScript. The AI effect? 80% of new repositories used just six languages: Python, JavaScript, TypeScript, Java, C++, and C#.
Generative AI is now standard in development. 1.1M public repositories now use an LLM SDK.
Agents are here. Coding agents created 1M+ pull requests (PR) in the last 6 months, and it's just getting started.
See full report here - Any results that struck you?
Top comments (3)
This is such a valuable overview of the open source landscape. Your analysis of the shift toward modular architectures was particularly eye-opening. I'd love to hear more about which communities you think are leading this charge. Also, your perspective on AI's role in open-source development is spot on : any projects you're excited about in this space?
thank you!
I have a bias towards Ultracite, an opinionated, zero-config code linter and formatter, and Kilo, a popular open-source AI coding agent
hope it inspires
Thank you!
I'm seeing formatting conflicts in our ML pipelines where Kilo's suggestions need manual tweaking to pass linting. How are you handling Kilo's output when it conflicts with Ultracite's rules?