For open-source programs, AI coding tools are a mixed blessing
VLC and Blender report 'abysmal' AI-generated merge requests, forcing projects to restrict contributions.
Deep Dive
Industry analysis reveals AI coding assistants (like GitHub Copilot) create a mixed impact on open-source projects. While lowering barriers, they flood codebases like VLC and Blender with low-quality submissions, wasting reviewer time. Projects report a decline in average contribution quality. The response includes new tools, like Mitchell Hashimoto's system, to restrict contributions to 'vouched' users, challenging the open-door model.
Why It Matters
For dev teams, AI-generated code increases maintenance burdens and threatens to fragment critical software ecosystems.