Open Source and the agentic wave
Source: Dev.to
Why the policy matters
While we recognize that AI‑assisted tools (such as Copilot, ChatGPT, or Claude) can be powerful aids for development, they also facilitate “low‑effort” or “random” contributions that increase the burden on maintainers without adding proportional value.
Reading out AI‑generated PRs for fun and profit
Open‑sourcing a project (or donating it to a foundation) is usually good for its longevity and sustainability. If the problem you’re solving is relevant and contributors can find your project, you’ll likely benefit from more eyeballs and external perspectives.
But with great visibility comes… the agentic wave: AI‑generated PRs and bots in your community channels.
About CloudNativePG
- What it is – a Kubernetes operator for PostgreSQL.
- Open‑sourced – April 2022 under the Apache 2.0 license.
- Stars – ~8 000 on GitHub, indicating strong attention.
- CNCF status – Joined the Cloud Native Computing Foundation as a Sandbox project in January 2025. (CNCF maturity levels: sandbox → incubating → graduated; each step brings more usage and visibility.)
“Kubernetes Maintainers Read Mean Comments”
In the spirit of the entertaining talk Kubernetes Maintainers Read Mean Comments – where speakers read the nastiest GitHub comments – we (the folks on the front line of CloudNativePG triage) decided to read out the worst AI drivel we’ve encountered masquerading as contributions, for fun and profit.
Below are some of the “gems” we found.
1. Copilot PR
“I used Copilot to implement this fix as I do not have any Go knowledge.”
No comment.
2. Massive PR
One PR changing 100 files and adding almost 23 000 lines? No thanks.
3. Missing context for an issue reference
AI trying to include a reference to an issue without the correct link and without the context that it should be managed by Renovate.
4. Complaining about another AI’s version pin
AI complaining about how another AI is pinning a version.
5. Redundant test suggestion
AI suggesting to add a test that verifies a function already has unit tests, while ignoring the need for an E2E test that was discussed in the issue comments.
Take‑aways
- Signal vs. noise: AI can produce a flood of low‑value PRs that waste maintainer time.
- Policy matters: A clear AI‑content policy helps set expectations and reduces the maintenance burden.
- Human review is still essential: Even sophisticated models lack the contextual awareness that seasoned contributors bring.
If you’ve encountered similar AI‑generated noise in your own projects, feel free to share your stories in the comments!
AI created a PR with basically the same content as the PR on the left—submitted by one of the maintainers—only four hours later.
Did the AI know about the first PR and just reuse it without crediting the original contributor? Or did the contributor fail to check whether the issue had already been solved before submitting their own solution?
Also: why are there so many spaces?
These examples may have inspired a few laughs, but some pull requests look great until unexpected behavior shows up in testing, and the author can’t explain what’s happening because they didn’t write the code.
Even more disappointing has been the surfacing of fake applications in the LFX mentorship program. This initiative exists to empower actual human beings, providing them with the tools they need to grow their expertise and climb the contributor ladder. Instead, we’ve wasted time interacting with applicants who were wholly unqualified.
Read on if you’re human
If you would like to contribute to CloudNativePG, here are some ways to get involved:
- Browse our GitHub organization and pick up issues that match your interests and skill set.
- Join the CNCF Slack; all channels prefixed with
cloudnativepgare ours. - Check out the calendar for Office Hours and Developer Meetings. You’re very much invited, and all notes are public.