Update CONTRIUBTING.md

[Blake-Latchford]

What is the Problem?
An abundance of AI slop has made reviewing community PRs untennable.

What is the Solution?
Propose a pathway for the community to police itself.

Is you change compaitble with nodes that don't have it? Not applicable

Checklist:
[x] Is your PR < 100 lines?
[ ] Has your PR been reviewed by at least two community members.
[ ] Have you reviewed at least 5 other community member PRs?
[ ] Have you built and flashed at least two other community member PRs?

[Blake-Latchford]

b48g55m — 8:32 AM
People will just gather friends. Look at https://github.com/ghostty-org/ghostty/blob/main/CONTRIBUTING.md for how other ways to do this.

[Blake-Latchford]

Avamander [ÕÄÖÜ], — 10:10 AM
For starters it would make reviewing easier if there were some code quality metrics being gathered automatically, like the amount of errors/warnings being added or flash and RAM used.
Manually checking everything is wasted effort.

CisienRole icon, MeshCore Legends — 10:47 AM
Automation can at least flag quality issues

[jbrazio]

wtf this PR is the one that looks AI-suggested.

[Blake-Latchford]

wtf this PR is the one that looks AI-suggested.

I mean, that's a little rude.

[Blake-Latchford]

I guess my core complaint was that it doesn't look like there is any feedback on community PRs. I guess I should consider "your work is garbage" as an improvement, since it is indeed feedback. Something more actionable would be appreciated though.

[jbrazio]

I mean, that's a little rude.

Sorry.. this PR makes no sense.

"Have you reviewed at least 5 other community member PRs"
"Has your PR been reviewed by at least two community members"

.. in what world does this make sense to be a requirement ?
Instead of "AI slop" we would have "human slop".. reading endless comments just for the sake of meeting the requirements.

Just look around a PR gets comments/feedback if it needs them.

[Blake-Latchford]

Just look around a PR gets comments/feedback if it needs them.

I did, though if it would be helpful I could go gather specific counts. The vast majority of PRs produced by people that are not part of the core dev team have received no feedback. This is doubly true for issues. If you look at the changes merged into the dev branch over the last few months there have been only a very small number of non-core team members that successfully got changes of any kind merged.

To quote recrof:

I'm bit worn out from all the reviews tbh. LLMs made a nightmare getting thru them all. lots of non-coders are submitting PRs with questionable quality or monster PRs. I only have so much energy which I channel elsewhere. not talking for whole team tho. ... motivation gets to the bin if you need to swim thru garbage every day

https://discord.com/channels/1495203904898728149/1495410606985969765/1548630343337115680

The intention behind the change was to try to help solve that problem. I think there is a lot of good work that has been produced by the community, and it doesn't seem like there is any indication that work well ever make it upstream. If the root cause of the lack of feedback is that there is too much slop coming in the pipeline. If that is indeed the problem, then it seems like one possible solution would be to have the community act as a reviewer for baseline quality, which might help filter the input to the dev team.

Reasonable people could argue that the MeshCore project doesn't need or want to solve this problem. Or that it isn't a priority. I'm pretty surprised to hear "a PR gets comments/feedback if it needs them".