How-To Geek surveys the intensifying dispute inside KDE over what role, if any, AI should play, as a flood of generated contributions consumes maintainer time. It recounts KWin developer Vlad Zahorodnii’s mailing-list post ‘How to deal with fully LLM-generated merge requests’, which complains about verbose replies that arrive impossibly quickly, compares the influx to a DDoS attack and suggests simply closing obviously LLM-generated requests. It details contributor Nate Graham’s proposed policy, built on a human-in-the-loop principle that requires a person to make the decisions and the result to be functionally indistinguishable from unaided work, permitting LLMs for drafting and bug-finding so long as a human verifies and tests the output and allowing machine translation of one’s own words, while prohibiting copy-pasted output the submitter does not understand. The article notes the practical difficulty of detecting LLM use, community objections that generated code may be plagiarised and incompatible with KDE’s licences, KDE Eco’s concerns about AI energy and water use, the strict no-AI ‘KDE for the People’ campaign, and an Akademy proposal for an AI-native KDE with a Kadai integration offering local AI, other models or none, concluding that the fight may split the project into pro- and anti-AI factions.
KDE is fighting over whether AI belongs in Linux
How-To Geek surveys the intensifying dispute inside KDE over what role, if any, AI should play, as a flood of generated contributions consumes maintainer time. It recounts KWin developer Vlad Zahorodnii's mailing-list post 'How to deal with fully LLM-generated merge requests', which complains about verbose replies that arrive impossibly quickly, compares the influx to a DDoS attack and suggests simply closing obviously LLM-generated requests. It details contributor Nate Graham's proposed policy, built on a human-in-the-loop principle that requires a person to make the decisions and the result to be functionally indistinguishable from unaided work, permitting LLMs for drafting and bug-finding so long as a human verifies and tests the output and allowing machine translation of one's own words, while prohibiting copy-pasted output the submitter does not understand. The article notes the practical difficulty of detecting LLM use, community objections that generated code may be plagiarised and incompatible with KDE's licences, KDE Eco's concerns about AI energy and water use, the strict no-AI 'KDE for the People' campaign, and an Akademy proposal for an AI-native KDE with a Kadai integration offering local AI, other models or none, concluding that the fight may split the project into pro- and anti-AI factions.
Source: Howtogeek