Skip to content
GOPPO

News · AI summarised to understand what matters

Back to news

Dev & Builders

Published on

For open source, AI coding tools are a mixed blessing

AI-assisted coding lowers the barrier to shipping features, but open source maintainers report a surge of low-quality submissions that waste review time and strain already-limited resources.

  • open-source,ai-coding,manutencao,qualidade,github

Summary

As AI coding tools become increasingly accessible, generating software feels cheaper and faster. For open source projects, however, the impact is not purely positive. Project leaders and experts describe a growing wave of low-quality contributions that consume reviewer time and sap maintainer motivation.

At the same time, the same tools can be genuinely helpful for experienced developers, especially when used under strong technical leadership and with a deep understanding of the codebase.

In practice

  • VideoLAN (VLC) CEO Jean-Baptiste Kempf says the quality of many merge requests from contributors unfamiliar with the codebase is “abysmal,” adding that AI tools work best “for experienced developers.”
  • Blender Foundation CEO Francesco Siddi says LLM-assisted contributions often waste reviewers’ time and affect motivation; Blender is developing an official policy and says such tools are neither mandated nor recommended.
  • To cope with the flood, developer Mitchell Hashimoto launched a system to restrict GitHub contributions to “vouched” users, arguing that AI removed the natural barrier that let OSS projects “trust by default.”
  • In security reporting, the cURL project halted its bug bounty program after being overwhelmed by reports its creator Daniel Stenberg described as “AI slop.”

Limitations

The article does not provide quantified metrics (e.g., rejection rates, review time increases, or release impacts). It also does not fully detail which governance or technical controls projects plan to adopt beyond the examples mentioned.

Why it matters

  • AI can boost contribution volume without increasing the supply of skilled maintainers.
  • Filtering mechanisms may reduce noise but can also narrow participation.
  • Maintainer overload can worsen fragmentation and technical debt.
  • The core challenge shifts from “writing code” to “managing complexity.”