AI code contributions

The next major release of the PostgreSQL database is reportedly running late, with maintainers attributing delays to the volume of AI- and LLM-generated contributions that require extra review and cleanup. Developers say vetting low-quality or machine-written patches is consuming time normally spent on the release itself.

PostgreSQL 19's release is running late, with developer Robert Haas attributing the delay to an unusually high number of bugs being found in the codebase. In a blog post he connects the bug surge to AI-assisted development generating more testing and code scrutiny. The PostgreSQL community is debating whether this reflects healthier testing or AI-produced low-quality contributions straining maintainers.

Linus Torvalds says the Linux kernel is out of its stabilization phase and that AI-related contributions now follow a 'much healthier model'. He acknowledged machine-made patches still 'often get rewritten', suggesting AI-assisted code is entering the kernel process under closer human review rather than being accepted wholesale.

Discussion at the Wuzhen Summit has turned to open source, prompting a fresh look at Ant Group's fourteen-year record of open-sourcing its technology and the business value of its AI work. Commentators are weighing whether the company's long-standing commitment to sharing code is paying off commercially as AI becomes central to its strategy.

Attendees returning from the Google Summer of Code Mentor Summit say artificial intelligence was the main subject of conversation. Mentors noted that AI has changed how contributors apply to open source projects, but not what makes those projects succeed: being a genuine user, engaging with the community, and caring about the project. A recurring point was that maintainers, not code, are now the scarce resource.

DeepSeek has open-sourced dsh-superagent, a coding agent built on a router-worker architecture in which a router model delegates tasks to specialised worker agents. The release gives developers free access to the framework, and it is being picked up in AI and developer circles as another notable open-source contribution from the Chinese AI lab, whose earlier models already drew wide attention for their low cost and strong performance.

AI-generated contributions are making it harder to trust open source code, as low-effort automated submissions flood maintainers. Bloomberg and the OpenTelemetry project have adopted measures to verify contributor identity and intent, restoring some confidence in the review process. The debate reflects wider concern in the open source community about maintaining quality and accountability as AI tools reshape how code gets contributed.

The open-source RADV Vulkan driver for AMD Radeon graphics, part of the Mesa project, has adopted a formal policy on the use of large language models in its development. The policy sets rules for how LLM-generated code contributions are handled by the driver's maintainers. It follows a broader debate in open-source communities over accepting AI-written patches and disclosure requirements.

The open-source Remote Desktop Protocol client project FreeRDP has released version 3.33, its headline feature being support for WebAssembly, allowing the client to run in web browsers. The release also coincides with a surge of contributions to the project, which maintainers attribute to AI-generated code submissions.