I created this thread thinking that AI is definitely going to shake the open-source world, and I wanted to know how people deal with AI-slop PRs.
Thank you for sharing your experience. I am not an open-source rockstar, and I have only had to reject PRs a few times in my life. All the times I felt terrible for people who put in some effort to build something around my code and got rejected.
Even if you reject their changes remember they have their own fork, and they can keep using it locally. Sure they have to keep updating their local work if you don't include it in your repository, but that's a small price to pay. After all if you hadn't made your code available in the first place nobody would be using it, modified, or unmodified.
To be honest, I'm not a fan of the PR's general idea.
> Is it possible to merge one part and keep it disabled by default for most users?
This is exactly what I am about to do. To merge a small part of that PR into my repo. That part conducts communication between the web app and the proposed extension, so the author of the PR can use my APIs while keeping most of their code in their fork.
I am actually glad that I created this thread. Thank you for responding to it.
It is a ChatGPT wrapper behind AWS Lambda. We use gpt-3.5-turbo model.
The extension sends the text of the commit message to Lambda. Lambda asks ChatGPT to grade the text for politeness, constructiveness, and agreeableness. The score is the average of those 3.
It can be helpful in any text input. We decided to validate the idea on GH for the starter because most users can try it as a Chrome extension. If the idea gets some traction, it can be extended further.
Four years ago, I found a small niche for a tiny project. Not that there were no solutions yet, but all the existing solutions lacked accessibility aspects, so I decided to develop my own, introducing the accessibility flavor [0].
The scope and limitations I had:
- Should be framework-agnostic
- Should require the minimum effort in installing
- Should not conflict with anything on the page
So, Web Component was the perfect candidate for the lib.
The framework-less approach felt overwhelming, and lit-html smelled like Polymer. So, initially, I picked up Stencil. It is still fantastic for creating component libraries, but it happened to be overkill for a tiny component. So in a few years, I rewrote my WC without Stencil, and now it feels just right.
I developed several more Web Components professionally in the past few years. I saw them as a perfect fit for the problems I had to solve. The scopes of problems were similar:
Thank you for sharing your experience. I am not an open-source rockstar, and I have only had to reject PRs a few times in my life. All the times I felt terrible for people who put in some effort to build something around my code and got rejected.
reply