He used AI to find a 15-year-old bug and got rejected for it — while a Linux kernel maintainer steps back because half his review queue is AI patches

"Another day, another disappointment," developer Vittorio Romeo posted on Nov. 13 after finding a 15-plus-year-old memory corruption bug in QuakeSpasm, a popular Quake source port, with AI help. The post drew 2,061 likes and 68,000 views.

Romeo said he spent his free time on an issue and two pull requests, but "the maintainer dismissed them due to AI usage and decided to copy my fix." He posted a side-by-side showing his patch carried over with a few lines changed, and said the issue and PRs were later deleted, leaving Internet Archive copies. He reported it because derivative ports share the bug, and asked people not to harass the maintainer: the point, he wrote, is "a systematic problem in the open-source community with anti-AI hysteria." The maintainer has not responded publicly.

The same week the Linux kernel showed the other side. The Register reported that Arm principal engineer Lorenzo Stoakes told the kernel mailing list that "given the current throughput in the kernel resulting from AI-generated series," he can't review outside the areas he maintains. He keeps his four maintainerships, covering most of the userland memory-management API, but said AI patches went from "once a week" to "daily" to "half of what I see in review." That came a day after Linus Torvalds said the AI patch tide was "starting to turn."

Stable-kernel maintainer Greg Kroah-Hartman said at last month's Open Source Summit: "The onslaught of the LLM-generated patches… we don't have enough people to review them all… All of the maintainers are burnt out." He wants to cut long-term-support releases from six to two, "four, maximum."