How LLMs Are Eroding the Role of Tiny Open-Source Utilities

Nolan Lawson says LLMs now replace tiny utilities like blob-util, producing instant, dependency-free snippets and undercutting their educational role. He urges a shift toward research, creativity and niche work.

How LLMs Are Eroding the Role of Tiny Open-Source Utilities

TL;DR

  • LLMs make tiny utility code instantly available, reducing incentive to add small dependencies
  • blob-util as exemplar: decade-old, still ~5+ million weekly downloads, typifies tiny educational libraries
  • Built-in platform features and Node/browser standardization already lowered demand; LLMs accelerate that shift
  • Durable open source tied to work LLMs cannot easily reproduce: original research, creative problem-solving, or highly niche expertise (e.g., performance/memory-leak investigations, exploratory libraries)
  • Ecosystem recalibration, not a eulogy; continued value in projects that push boundaries and documentation that teaches

Nolan Lawson on the shrinking role of tiny utility packages

Nolan Lawson argues that LLMs are reshaping the value proposition of small open-source libraries. His decade-old utility package blob-util — which still registers 5+ million weekly downloads — exemplifies the class of tiny, educational libraries that offered both a solution and a teaching moment. Today, an LLM such as Claude can generate equivalent utility functions instantly, removing much of the friction that used to send developers to npm.

What changed, and why it matters

  • The immediate availability of LLM-generated code reduces incentives to add a small dependency just to handle routine tasks.
  • That shift tends to erode the educational role these packages once played: many small libraries bundled tutorials and examples that helped developers learn underlying APIs and patterns.
  • Built-in platform features and Node/browser standardization have already reduced demand for some utilities; LLMs accelerate that trend by making quick, dependency-free snippets trivial to obtain.

Where open source still adds value

Lawson suggests that the most durable, worthwhile open source will be work that an LLM cannot easily reproduce — projects that require original research, creative problem-solving, or highly niche expertise. Examples include investigations into performance and memory leaks or exploratory libraries that introduce genuinely new ideas. Human-driven experimental frameworks and deep-dive tools are less likely to be supplanted by canned agent output.

A modest, pragmatic note

The post is not a eulogy; it is a recalibration. Small utilities are not vanishing because of poor design but because the ecosystem and toolchain have evolved. There remains fertile ground for contributors who focus on novel problems, thorough documentation that teaches rather than just answers, and projects that push the boundaries of current tools.

For the fuller argument and examples, read the original piece by Nolan Lawson: https://nolanlawson.com/2025/11/16/the-fate-of-small-open-source/

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community