Vibe coding’s transformative impact on developer productivity is having adverse consequences for the open-source ecosystem, potentially exposing enterprises reliant on open-source software to risk.
A paper from the Central European University (CEU) in January found that vibe coding – driven by AI coding assistants such as Claude Code, Cursor and Lovable – decreases the depth to which developers engage with code, documentation, libraries and other developers.
“Traditionally, a developer selects packages, reads documentation and interacts with maintainers and other users. Under vibe coding, an AI agent can select, compose and modify packages end to end, and the human developer may not know which upstream components were used,” said the paper, authored by Miklos Koren, an economics professor at CEU, and his colleagues.
Even though open-source software does not have to be paid for, its value, especially to enterprises, is hard to deny. A 2024 Harvard Business School study estimates that firms would have to spend US$8.8-trillion globally to replace open-source software with proprietary builds, and that companies would spend 3.5 times more on software development if open-source software did not exist.
Calwyn Baldwin, automation team lead at Johannesburg-based enterprise open-source solutions provider Obsidian Systems, said that like any other tool, AI coding assistants tend to reflect the skill and care of the developer using them. Developers with good habits get good results; those with bad habits inevitably produce bad code.
Different skill set
“AI coding is certainly a different skill set. It is very easy to tell the AI to go and do something, test if it works and then never look at the code again. If you do that over the long term you are going to have a bunch of developers that don’t know the code base very well.
“While that is a fair concern, I don’t think that is true for every developer because some are still going to get stuck in and try to understand how it works. Any developer worth his salt wants to develop code that is maintainable, that they understand and can explain to someone else,” he said.
Baldwin also disagrees with the paper’s argument that vibe coding diminishes social engagement in open-source projects, arguing that AI tools have, to the contrary, raised the need for higher levels of engagement, especially between experienced developers and their more junior counterparts.
Read: The problem with vibe coding
He said the most significant impact of the AI coding paradigm is likely to be that junior programmers will not have the opportunity to develop the strong, high-level architectural skills that their senior counterparts possess, in part because AI tools remove some of the cognitive load associated with traditional software development. The antidote is to ensure juniors and seniors engage more frequently so that skills are transferred.
Bennie Kahler-Venter, a senior automation engineer at Obsidian, said it is important that senior staff are the first to engage with AI coding tools so they can identify their strengths, weaknesses and where they will be most useful in a project’s development pipeline.
Kahler-Venter also draws a distinction between vibe coding and higher-level, architectural approaches such as spec-driven development and agentic coding. He argues that AI coding tools are more powerful and produce less “slop” when approached from an architectural perspective with contexts, guardrails and style guides clearly defined.
“Vibe coding is just the start; it is your intern level. That’s what the Linux kernel guys did and they got a lot of criticism for it because not everyone wants to code using AI agents. They then started putting down rules for AI agents so they follow style guides and testing protocols. All those guardrails make good-quality software repeatable,” he said.
Read: Inside GitHub’s plan to foster a billion developers
Julian Gericke, chief technology officer at LSD Open – which also provides open-source solutions to enterprises – said it is too early to tell what the effects of AI coding tools on the open-source community will be. But he sees a likely outcome: the velocity at which code can now be developed will flood the market with vibe-coded open-source projects, making it difficult for enterprises to discern which are well written, well maintained and backed by a good team.
With many more projects to choose from, the need for reliable, mature open-source projects will rise, creating an incentive for enterprises to invest in their success, he said.
Gericke said another often overlooked point is that many AI projects rely on open-source repositories not only for training but for the execution of tasks, too. This creates an incentive for AI coding companies to ensure those open-source projects are maintained, even if it means creating AI tools to do so.
“Where there are highly valuable open-source projects out there, a lot of their revenue stream comes from an enterprise model where they sell enterprise features. Or enterprises deem that those projects are so valuable that they hire in their core developers. A lot of the market has a vested interest in the open-source world thriving and not imploding,” he said. – © 2026 NewsCentral Media
Get breaking news from TechCentral on WhatsApp. Sign up here.
