AI Generated

2025-09-27

I have been maintaining open source projects for more than a decade at this point and have recently started noticing that some people who interact with my projects (especially for the first time) are sending text which is generated largely or even completely by an AI. If you are reading this, chances are you were sent a link to this page as a part of a terse response to something you pushed to me.

As a maintainer, I think it is great that LLMs have made it easier for many people to try new things or quickly get a handle on new projects. This includes both if your experience doesn't cover writing code at all or just doesn't involve directly working on this project.

However, all of our time is limited and everyone has to be selective about what they decide to spend their time doing. Open Source maintainers frequently dedicate an inordinate amount of time to making their projects better, solving bugs and implementing features requested by their users. Many of us do this purely out of an enjoyment of helping people and it makes our day or even our week to solve some stranger on the internet's problem.

Anything which kills that joy must be avoided. If maintaining a project turns from a fun side project into a painful drudge, I will stop doing it.

As you might have guessed by now, being sent large amounts of AI generated text that I have to slog through in order to uncover the actual need of the human behind it fits exactly into the type of drudgery that will lead to the decision that it isn't worth my time to interact with it.

There are a multitude of reasons for this, but in the interest of respecting your time I'll limit this post to what I am asking of you:

If making a bug/feature request, explain your problem.
LLMs can probably come up with a dozen adjacent features and explanations to make them sound critical, but as a human in the business of building things for humans, I'm interested in hearing what you want, not what some obfuscating LLM interpreted the original request as.
If making a pull request, make sure that you understand everything it contains.
Before I merge anything, I make sure to read through it and understand what it does and why. A LLM might be able to help you more quickly come up with a solution, and if it does, great! However, you, the human, needs to understand why what you are submitting works and why each portion of it is necessary.

AI output tends to include a ridiculously large amount of noise for any useful information produced. You are welcome to use AI to assist in contributing to any of my projects, but if you do, I expect you to do the work to distill that output down into something a human would actually write. I will not do that work for you, and will ignore/close any submission which attempts to foist that work onto me.