With Love to KDE: Take a Moment (korcenji.neocities.org)

A long‑time KDE user published an open plea asking the project to adopt a formal “AI” contributions policy (citing Servo, Asahi Linux, Bevy and others) instead of defaulting to permissive acceptance of LLM‑produced patches. The author objects to a line of reasoning that because absolute provenance is impossible, KDE should simply accept contributions that may originate from large language models; they argue KDE should evaluate whether accepting such contributions is actually desirable and plan accordingly. Technically and practically, the piece outlines why LLM‑sourced code is problematic for FLOSS projects: models trained on scraped data can’t provide full, reliable attribution (risking plagiarism and licensing violations), frequently hallucinate or misattribute authors, and produce low‑quality or inscrutable patches that increase reviewer burden and maintainer burnout. The author also highlights ecosystem harms — extractive scraping, amplification of disinformation and surveillance, and substantial energy costs of large AI datacenters — and points to KDE examples like optional Whisper integrations. The call to action is pragmatic: draft an AI contributions policy informed by peers (Servo, Krita, Asahi, Bevy) and community voices to protect provenance, maintain code quality, and set standards for acceptable AI usage.