Tailwind Labs CEO Adam Wathan recently blamed AI for forcing him to lay off three workers.
Tailwind Labs oversees the development of the open source Tailwind CSS framework. And according to Wathan, AI coding tools came between the company and its customers, reducing traffic to the website, which in turn hit product exposure.
“Traffic to our docs is down about 40 percent from early 2023 despite Tailwind being more popular than ever,” he wrote in a GitHub Issues post earlier this month. “The docs are the only way people find out about our commercial products, and without customers we can’t afford to maintain the framework.”
A recent pre-print paper cites the situation at Tailwind Labs as evidence that the increasing adoption of AI coding tools represents a breaking change for the open source community.
The paper, titled Vibe Coding Kills Open Source, was written by Miklós Koren (Central European University, KRTK, CEPR and CESifo), Gábor Békés (Central European University, KRTK, and CEPR), Julian Hinz (Bielefeld University and Kiel Institute for the World Economy), and Aaron Lohmann (Kiel Institute for the World Economy).
Essentially, the paper argues that AI tools install open source dependencies in a way that comes between software developers and project maintainers, undermining interactions that potentially return value to those doing the work of software maintenance.
“Vibe coding raises productivity by lowering the cost of using and building on existing code, but it also weakens the user engagement through which many maintainers earn returns,” the authors argue. “When OSS is monetized only through direct user engagement, greater adoption of vibe coding lowers entry and sharing, reduces the availability and quality of OSS, and reduces welfare despite higher productivity.”
Co-author Koren, a professor of economics at Central European University in Vienna, Austria, told The Register in an email that evidence of AI tools weakening engagement with the open source community is mostly circumstantial.
“There is a documented decline in Stack Overflow questions after the launch of ChatGPT; and it is faster than in countries where ChatGPT was not accessible,” he said. “There are many similar anecdotes circulating on social media. In fact, we started working on the paper before Tailwind’s case became publicized.”
Koren said the paper’s findings represent an extrapolation about what he and his co-authors know given an economic model of the open source software ecosystem.
“We know that developers have adopted vibe coding very fast,” he said. “[Anthropic CEO] Dario Amodei has famously said at the Axios AI+ Summit in September 2025 that ’70, 80, 90 percent of the code written in Anthropic is written by Claude.’ So users of vibe coding find it easy to switch to this mode of building software. This implies that human attention towards producers of OSS is shrinking.”
The impact of that attention shift is not measured only in terms of revenue. Rather, it’s assessed as an amalgamation of the rewards available to open source developers, such as community recognition, reputation, and job prospects, Koren said, pointing to a recent paper [PDF] that found “only about 0.1 percent of the total value created by OSS developers is captured by them.”
Koren said that the impact of the adoption of AI tools differs depending on the size and governance of the project.
“High-quality projects can still thrive,” he said. “OSS developers need visibility of their project to gather useful feedback from users, recruit new developers and maintainers, and to garner kudos from the community. We don’t think that large OSS projects will disappear overnight. But it will be harder to get beyond the ‘cold start problem’ and get an otherwise promising project off the ground. Or maintainers of marginally successful projects may lose their motivation and stop contributing. The proverbial ‘random person in Nebraska’ may give up.”
Koren said this is a systematic issue that needs industry collective action.
“Traditionally, OSS work is not directly compensated by users not because they don’t value OSS, but because of the frictions involved,” he explained. “I may use dozens of libraries for any given project. I will not look up each individual maintainer and give them a fraction of a cent. There are ongoing funding drives at GitHub or npm, but these do not solve the problem because they still need the user to pay attention and to open their wallet.”
But AI companies could help, Koren argues, noting that most LLM inference is done by a handful of large providers like OpenAI, Anthropic, or third-party providers like OpenRouter and Groq.
“Metering the usage of each OSS library would be technically simple to do,” he said. “This could serve as the basis for a revenue-sharing deal. Much like Spotify pays artists based on playtime, OSS developers could share some of the LLM revenue based on actual usage.”
Armin Ronacher, creator of Flask and an experienced open source developer, told The Register in an email that while AI has altered open source, he’s reserving judgment.
“AI is definitely changing a lot about the dynamics of open source,” Ronacher said. “In particular, it makes code cheaper and it changes the associated calculations. Open source right now sees a lot more low-quality contributions, but on the other hand, it might strengthen some key projects where trust comes from strong maintainers with a track record.
“I think it’s too early to say where this will land. This is a massive shift and we will see the effects in a few years. It will take some rewiring and it’s very hard to come to conclusions until it all settles again.
“In general I don’t try to think much about alarmism on any of this right now. All the energy put into meta discussions right now seems wasted until we find a new normal.” ®
