Skip to content

README upgrade#7

Closed
midnight-wonderer wants to merge 1 commit intombuesch:masterfrom
midnight-wonderer:chore/readme-upgrade
Closed

README upgrade#7
midnight-wonderer wants to merge 1 commit intombuesch:masterfrom
midnight-wonderer:chore/readme-upgrade

Conversation

@midnight-wonderer
Copy link

I know that you probably hate making a README presentable.
That is what AI is for.

I help vibe-code the README.

@mbuesch
Copy link
Owner

mbuesch commented Mar 8, 2026

I am sorry, this is pretty much the definition of AI slop.

Generating an AI slop pull request is almost zero work.
Reviewing an AI slop pull request is a huge amount of work.

I have no idea whether the facts added to the Readme are correct or not. For example I have no idea how these other tools work.

Please refrain from doing large "vibe coded" pull requests.
This hurts Open Source software.

@mbuesch mbuesch closed this Mar 8, 2026
@mbuesch mbuesch added the invalid This doesn't seem right label Mar 8, 2026
@midnight-wonderer
Copy link
Author

I respect your opinion. However, it's worth considering that by avoiding AI entirely, you might be missing out on its potential.

One thing AI cannot do, at least for now, is make decisions. The "slop" you referred to is likely caused by people assuming AI can handle that responsibility.

The README started with a significant amount of information from you; the AI was used for transformation work, not for making heavy decisions.

Also, for what it's worth, the project is discoverable thanks to AI; it suggests your utility when asked the right questions.

@mbuesch
Copy link
Owner

mbuesch commented Mar 8, 2026

Thanks for your response.
Please let me explain why I consider PRs like this not mergeable.

The primary reason is not that this is AI generated.

The reason is that this PR is a big blob rewrite of an artifact. It lacks a couple of fundamental standards of software development:

  • It is too big of a delta. This is a major effort to review. There is a huge difference between the time required to generate/post this PR and to merge/review it.
  • There is no reasoning and explanation for the changes. Why are things changed? Why is it better after the change? Why have things been added, removed?
  • What is the comparison to the other tools based on? I don't know these tools in such a detail that I could make a decision whether this is correct or not. I really don't want to have wrong facts about other applications in my readme.
  • As a reviewer I have to compare all numbers to the originals. AIs still make silly mistakes with numbers.

All in all this is way too much effort on my side for no real benefit.

I personally use AI tools on a daily basis.
The problem is not AI.
I would have rejected this PR, too, if it was written by a human. However, that doesn't happen. No human writes this amount of slop before contacting the maintainer. Humans make smaller changes and explain why they did these smaller changes.

Last but not least: If I wanted to completely rewrite the readme, I would prompt the AI by myself. This would be much easier than going over the PR route.

That said, you are welcome to contribute. Even with the help of AI tools.
But it's really important to keep the basic principles of software development.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

invalid This doesn't seem right

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants