Reality check please

[RlySkiz]

I'm fairly new to actually contributing bug reports to a bigger project like this.
It is seemingly moving really fast and issues/bugs get responded to just as quick including implementations/fixes.
Are there even people looking over things or is everything just Ai's responding to the reports and other Ai dealing with implementation etc.?
I feel like i have too much power by just stating an issue or suggestion and it immediately gets implemented if so.
Like.. things gets just accepted as facts, "no problem to implement", "sure thing"..
But yea as i said i just never dealt with a project like this before.

[CoolJohn-lab]

Pewds already started appointing volunteers to help him read the bugs and issues and respond to them. Yes, he has an AI agent doing the work as well but a lot of it is people manually reviewing things now. Also as of today, you can no longer just request things to get merged into main ;D There's a dev branch and everything will have to get reviewed at least twice before being accepted, and won't just be released straight away.

[Michiel-VandeVelde]

That’s a fair concern, and I get why it can feel a bit chaotic from the outside.

The repo has been moving really fast, so some of us have started helping with triage/review to bring a bit more structure to it. I’ve spent around 6 hours today going through discussions, routing things, closing duplicates, and trying to move bugs/features into Issues where they’re easier to track.

There are actual people looking over reports and PRs. We’re just still catching up with the pace. The idea now is to keep things a bit more controlled: bugs and feature requests should be tracked properly, PRs should go through dev, and main should stay more curated and properly tested.

Reports and suggestions are very welcome, but they aren't treated as instant facts or automatic approval to implement something. They still need to be reviewed, and tested.

So please keep reporting things, especially with clear details. Just don’t feel like you have “too much power”, reports help us decide what to look at, they don’t automatically decide what gets merged or made :)

[Michiel-VandeVelde]

Yeah, that’s fair. If some things were marked as fixed but the behavior is still broken, that’s definitely something we’ll need to re-check.

Things moved very fast at the start, and the project scaled way quicker than expected. Pewds has acknowledged that and asked for help from people with the right domain knowledge to improve things, so some fixes may need a second pass now that more people are helping with triage and review.

There’s a dedicated maintainer group starting to work through this more carefully, so please give us a bit of time to sort through the ripples :)

[RlySkiz]

Speaking of pewds, i don't even watch him but i did follow him dipping his toes into linux + well, this project.
From an outside perspective while you say all that, it would be nice seeing more posts of him on the forefront of what is happening/going to happen in a direction sense. As all you really see from him is a few commits here and there which are also claude-assisted. Of course a thing that happens here, but when you put something like this out it would be appreciated seeing him spearhead this visibly enough for people that everyone knows whats going on. I've seen y'all conversing about potential restructuring etc. but you almost see nothing of him. I'm normally not invested into any github projects and i don't even watch him really, its just something on principle that irks me about many other projects/games etc. Communication.

[CoolJohn-lab]

Ok well it’s been 5 days since the project launched and he has been very active on this GitHub project the whole time, go check his comment history. That just isn’t valid criticism I am afraid. He’s very much communicating, albeit mainly with contributors directly. I imagine we will see another YouTube video before too long with an update also.

[bitboody]

@Michiel-VandeVelde While there are many real people putting real work into this project, I think there should be some rules for AI generated PRs to maintain architectural integrity.

[Michiel-VandeVelde]

That’s a fair point.
That said, I don’t think the real distinction is simply “AI-generated vs human-written.” alot of AI-assisted PRs can still be useful if the contributor understands the problem, responds to feedback, tests properly, and keeps the change scoped.

The bigger issue is whether the contributor actually understands what they are changing. A lot of low-quality PRs are easy to spot because they ignore the contribution guidelines, bundle unrelated changes, do not explain the reasoning, or cannot respond meaningfully to reviewers when asked for clarification.

So I’d support rules that focus on quality and accountability: clear problem statement, scoped changes, testing notes, no gigantic bulk PRs, and contributors being able to explain and maintain what they submit.