The open source ecosystem is facing an existential challenge from artificial intelligence—not from automation replacing human coders, but from the sheer volume of low-quality contributions produced by large language models and agentic tools. What was once a collaborative effort sustained by human effort and care is now drowning in a flood of what maintainers call 'slop PRs.'

Mitchell Hashimoto, founder of HashiCorp, is considering closing external pull requests to his open source projects entirely. His reason: not a loss of faith in the model, but being buried under AI-generated patches. Flask creator Armin Ronacher has labeled this phenomenon 'agent psychosis,' where developers become addicted to the dopamine hit of AI-generated code and set agents loose on their own projects and, inevitably, everyone else's. The result is code that feels right statistically but lacks the context, trade-offs, and historical understanding that human maintainers bring.

The economics are brutal. It takes a developer 60 seconds to prompt an AI agent to fix typos and optimize loops across a dozen files—but it takes a maintainer an hour to review those changes, verify edge cases, and ensure alignment with project vision. Multiply that by hundreds of contributors using LLM assistants, and you get burnt-out maintainers who walk away.

This isn't hypothetical. The OCaml community recently rejected a single AI-generated pull request containing more than 13,000 lines of code over copyright concerns and maintenance burden. GitHub itself is reportedly exploring tighter pull request controls and even UI-level deletion options because maintainers are overwhelmed by AI submissions. If the host of the world's largest code forge is considering a kill switch for pull requests, this is a structural shift in how open source gets made.

The shift is hitting small projects hardest. Developer Nolan Lawson, author of the blob-util library with millions of downloads, argues that small utility libraries are becoming obsolete. In the age of Claude and GPT-5, developers no longer need to depend on a library when they can ask an AI to generate a perfectly serviceable utility function in milliseconds. Lawson warns that something deeper is lost: these libraries were educational tools where developers learned by reading others' code. Replacing them with ephemeral AI-generated snippets trades understanding for instant answers.

Ronacher offers another provocation: build it yourself. If pulling in a dependency means dealing with constant churn from AI-generated contributions, the logical response is to retreat—use AI to help you, but keep code inside your own walls. This creates a strange irony: AI reduces demand for small libraries while simultaneously increasing low-quality contributions to the libraries that remain.

Looking ahead, open source seems headed for bifurcation. On one side, massive enterprise-backed projects like Linux or Kubernetes will have resources to build AI-filtering tools and organizational weight to ignore the noise. On the other side, smaller 'provincial' projects will simply stop accepting external contributions. AI was supposed to make open source more accessible, but it has also lowered the value of each contribution. When everyone can contribute, no one's contribution is special. The only scarce resource left is human judgment—the ability to say no.

The future of open source isn't dying, but 'open' is being redefined. We're moving from radical transparency and 'anyone can contribute' to an era of radical curation. The most successful projects will be the most difficult to contribute to, demanding high levels of human effort, context, and relationship. The bazaar couldn't survive the arrival of the robots. The future of open source is smaller, quieter, and much more exclusive—and that might be the only way it survives.

In the end, we don't need more code; we need more care. Care for the humans who shepherd communities and create code that endures beyond a simple prompt.

Source: InfoWorld News