How Top AI Open Source Projects Are Managing Thousands of Contributors
Leading AI-native open-source projects are adopting new strategies to manage contributions, moving away from traditional pull requests. Vercel's AI SDK, which receives over 20 million npm downloads weekly, deployed a "software factory" with various agents to handle its backlog of over 1,000 open issues and nearly 800 pull requests. These agents now author 25-35% of merged PRs and close 70-80% of issues. Similarly, the Astro web framework uses agents for auto-triage, significantly reducing open issues. Some projects, like Flue and tldraw, are taking a more radical approach by automatically closing external pull requests and converting them into issues or discussions, relying on their own agents for implementation. This shift is driven by the ability of AI agents to perform tasks like bug reproduction, fixing, and review, leading to a reevaluation of community contribution models in open source.
Original sources
Newsletter Latent Space
Issue:PRs NOT Welcome: How Top AI Open Source Projects Are Managing Thousands of Contributors ↗
AI Radar wrote this summary from the publisher’s notes with AI, not from a transcript. How it works.