AI-generated pull requests strain Postgres maintainers
The Postgres core team is facing an influx of AI-generated code that is difficult to validate, creating a bottleneck for maintainers. Experts note that AI models struggle with C code, raising concerns about the quality and security of critical infrastructure projects.
The Postgres core team, consisting of fewer than 100 active contributors and roughly 50 committers, is facing an influx of AI-generated pull requests. Craig Kirsteins notes that while AI models are effective for generating user interface code in languages like TypeScript, they struggle with C code due to limited training data on high-quality open source C. This creates a challenging inflection point for maintainers who must validate the correctness of these submissions, as errors in database code can lead to data corruption.
The situation highlights a broader tension in open source where the volume of AI-generated code increases the burden on maintainers to filter low-quality submissions and ensure quality. This is particularly acute in critical infrastructure projects where the cost of failure is high. Kirsteins expresses concern that the surge in AI-generated code is overwhelming small open source communities, creating a validation bottleneck that could impact the quality and security of projects like Postgres.