First pull request used to be a coin flip (dparkmit.substack.com)

A recent analysis of 143,000 pull requests (PRs) across eighteen major open-source AI projects reveals a troubling trend: the odds of newcomers successfully getting their first PR merged have significantly declined since 2026. Initially, these odds ranged from favorable to approximately 50%, but they have worsened, with many projects now accepting first-time contributions at rates lower than 50%. This presents a challenge for aspiring contributors, highlighting that gaining initial acceptance into these communities is becoming increasingly difficult. The data indicates that varying project structures impact these rates, with some repositories closing considerably to external contributions. For instance, LangChain's merge rate plummeted from 55.8% to 7.1%, while open-webui's dropped to 4.0%. The study underscores the importance of understanding the dynamics of open-source contributions and the barriers faced by those unfamiliar with specific projects. As the open-source AI landscape evolves, it necessitates a reevaluation of strategies for newcomers seeking to engage meaningfully with these vital technical communities.