February 3, 2026
When vibes write your code
How Vibe Coding Is Killing Open Source
Devs say AI chatbots are gutting the community while maintainers fight fakes
TLDR: A new paper warns AI code-writing could drain open‑source of users, feedback, and funding. Commenters split between “RIP community” and “open source is an idea,” while maintainers report polished AI forgeries, a few good contributions, and some devs now hesitating to share code at all.
A new preprint has thrown gasoline on the debate: “vibe coding” — letting AI chatbots write your code — could be starving open-source projects of visits, feedback, and a future. The paper says bots push the “most popular” tools they were trained on, while human devs skip docs and community forums in favor of quick answers. The crowd’s reaction? Spicy.
One camp is waving red flags. arjie says we’ve “lost the single meeting place”, and lists the fallout: no pressure to contribute, LLM-written “resume PRs,” and bogus issues. Maintainer jph brings receipts: “AI-forgery attacks are highly polished” with fake profiles — but admits legit contributions are up too. Meanwhile, dom96 confesses they’re now wary of open-sourcing at all, worried their code will just feed another model.
On the flip side, Flavius fires back with a rallying cry: “You cannot kill an idea.” Open source, they argue, is bigger than how code is written. jauntywundrkind adds that the “substrate” — the shared foundation we all build on — will keep absorbing innovations, even if getting traction is tougher. And yes, people dunked on the vibe: jokes about becoming “prompt DJs,” and memes about a “Spotify for code” that pays in vibes.
Between studies claiming 41% more bugs and 19% less productivity, and the GitHub Copilot era, the community’s split: doomers, true believers, and exhausted maintainers watching Stack Overflow get quieter while their inboxes get louder.
Key Points
- •A pre-print paper warns that LLM-assisted “vibe coding” may harm open source by diverting engagement and making new OSS projects harder to start.
- •LLMs influence library/tool choices toward dependencies prevalent in training data, reducing organic selection and project website visits.
- •Community engagement declines are noted, including reduced usage of forums like Stack Overflow and fewer useful bug reports to maintainers.
- •Cited findings include 41% more bugs and 19% lower productivity associated with vibe coding tools like GitHub Copilot.
- •Authors propose compensation from companies like OpenAI and Google for OSS code use, but warn outcomes may mirror Spotify’s concentrated payouts.