Users of AI coding tools are flooding open-source projects with low-quality contributions, overwhelming maintainers and potentially eroding community engagement
Our obsession with AI code-writing tools is overwhelming the web's unsung human caretakers. If you are reading the digital version …
Financial TimesSam Learner
Context & Ripple Effects
Related coverage had already identified declining average contribution quality at projects including VLC and Blender as AI coding tools lowered the barrier to producing submissions. This report advances that pattern from a quality concern to an operating burden for the people responsible for reviewing and maintaining shared software.
The same coverage arc has shown AI-assisted code generation creating review and security bottlenecks inside companies. Open-source maintainers face a comparable mismatch: code can be produced faster than trusted humans can assess it.
First-order effects
Maintainers must spend more time triaging, reviewing, and rejecting weak or poorly understood submissions, reducing time available for roadmap work and established contributors.
Contributors using AI coding tools encounter more scrutiny and a higher likelihood that low-quality submissions will not be accepted, while project communities risk friction between maintainers and newcomers.
Second-order effects
Projects are likely to tighten contribution rules and emphasize evidence of testing, issue discussion, and maintainer review, raising the effective cost of accepting outside code even as producing it becomes cheaper.
Organizations that depend on open-source software may face slower upstream maintenance and need to devote more internal effort to reviewing AI-generated code before it reaches production.
Third-order effects
If AI-assisted submission volume continues to outpace human review capacity, open source could shift toward more gated, maintainer-controlled contribution models rather than the lower-friction participation model that built many projects.
The durable constraint in software development is increasingly code verification, security assessment, and stewardship—not code generation—potentially concentrating influence with projects and organizations able to fund that work.
The trend: AI coding tools are turning software creation into an abundant resource while making trusted review and maintenance the scarce infrastructure layer.
is prompting an LLM “just another step on the ramp of abstraction?” Thank you @sam_learner for this remarkably informative and thought-provoking text https://www.ft.com/...
this article opens with 10 paragraphs about Daniel Stenberg and cURL's slop problem as of January without mentioning that in April he said: “The slop situation is not a problem anymore” and AI reports “are mostly very high quality” — daniel.haxx.se/blog/2026/04...
This piece on the “load-bearing internet people” and how AI-code writing is overwhelming the web's unheralded maintainers by @samlearner.bsky.social is absolutely terrific: brilliantly accessible without being patronising:
Humans don't ask humans questions anymore - great piece on vibe-coding & its consequences by @samlearner.bsky.social www.ft.com/content/cec8... [image]
“Our findings frame AI slop as a tragedy of the commons,” the researchers concluded, “where individual productivity gains externalise costs on to reviewers, maintainers and the broader community.” — www.ft.com/content/cec8...
I reposted a link to this earlier and it really is a terrific article that is well worth your time. Includes comments from Guido van Rossum (Python) and Rich Harris (Svelte), among others. — bsky.app/profile/saml... [embedded post]
wrote for the magazine about the open source software that underpins our digital lives, how it is being upended by AI code tools, and about maintainers — as.ft.com/r/b7f62212-9...
The current top story on the FT Magazine is pay-walled but features some mentions of #curl and yours truly... (the picture shows the start of it) — https://www.ft.com/... [image]