Community Guidelines

[nanoscopic]

This issue about "no self promotion" is titled with Community Guidelines at the start: #680

While GitHub has established Community Guidelines ( https://docs.github.com/articles/github-community-guidelines ), this project does not have any specified.

If you are going to have community guidelines for this project, they should be pinned Issues and/or a file within the project.

One community guideline I'd like to see is for a focus to be on truth, as I and others have repeatedly asserted that the About of the project and claims within it have been clearly shown to be false.

Another good one would be to disallow automated bots commenting on things. I made an issue about that. Establishing a policy on that would be good. If they are to be allowed, that's fine. Disallowed, fine too. There should though be a decision made for it.

It has been said that critical feedback will be taken seriously, and is something else I would like to see followed through on. Ben has been closing issues without addressing critical feedback and that's a mixed message. Is this the style in which Discussions will be treated? Inconvenient or awkward discussions will be summarily closed without addressing the problems?

See https://docs.github.com/en/communities/setting-up-your-project-for-healthy-contributions/adding-a-code-of-conduct-to-your-project for ideas on a code of conduct for the project.

[bensig]

@nanoscopic thanks for this thoughtful messages. What issue was closed without addressing a problem? Would you please link to any you have noticed?

[nanoscopic]

Issue #705

Roman did legitimately meaningful work in his detailed analysis of your project. He points out that this whole project can be replaced by 50 lines of code that effectively give the same benchmark results. That's something that should be discussed and addressed. Igor just closed the issue. It should be properly converted to a discussion. Igor is acting in bad faith and it's offensive.

You may think "but there's no actual defect". There is though, that 90% of the code could be deleted and it would still work. That's not "no issue here".

[nanoscopic]

Issue #648

Igor again. Zero response. Just closed the issue. This is a problem.

There is a meaningful idea being presented here, which is "the documentation website for this project should be content within this project or a repo itself so it can have issues raised against it and updates to it tracked".

Just closing the issue without converting it to a discussion? Unacceptable.

[mvalentsev]

Sorry for jumping in again, a couple of remarks:

Issue #705

That "independent technical audit" is itself LLM slop -- a single prompt to an LLM, not a real research effort.

Issue #648

Your own comment there answered the question -- the docs site already exists and every page has "Edit this page on GitHub." I agree igorls should have acknowledged that before closing. But "the documentation website should be content within this project or a repo itself so it can have issues raised against it" -- that wasn't actually proposed in #648, the issue was about creating a docs site which already exists.

You have incredible productivity -- even with AI I can't keep up with that many comments. But why issues and not PRs? Also, regarding the README criticism you keep bringing up, have you seen dekoza's PR #641?

[nanoscopic]

If the AI investigation by Roman is going to be called AI slop, then this whole repo is slop squared...

I'm all for ditching everything done with AI, but I thought we were past the "AI was used to do this therefore slop." Like consider. The lead "architect" of this is an actress who knows... what exactly about making software? At least, thankfully, she has wisely decided to rename the org. So maybe it can be treated like something other than a joke now.

Yes, if others ( and me ) are being dumb and missing something obvious, address it and tell us in the issue instead of just closing the issue. Otherwise the next person to be confused in the same way is just going to come along and be confused as well.

I don't see the point in submitting PRs if they aren't going to be merged, or won't be merged in a timely fashion and I'd just have to rebase them over and over.

[nanoscopic]

Here is another: Cursor database mining: #287

Why was this PR closed? No explanation was given. Just -whap- closed. "hey could you rebase" needs to give a reasonable amount of time to rebase. It's not proper practice to just close the PR after saying "it has conflicts now; tough luck. bye." :|

[mvalentsev]

Hey, regarding the bot comments policy -- I use automation to cross-reference issues with PRs when triaging. It helps coordinate work in a repo with 200+ open PRs.

That said, with all respect to @nanoscopic, it feels like the bot policy concern is a proxy for something else. If the real issue is that some project contributors use AI agents for reviews or code, that's a fair discussion to have -- but it should be had directly, not wrapped in a general "community guidelines" framing.

[nanoscopic]

In regard to AI written messages and PRs:

The vast majority of all software is going to be coded by AI in the very near future. ( if that's not the case already ) Because of that I'm not objecting to the use of AI, as that ship has sailed. What I'm saying is that low quality ( hey ai respond with issue #x with slop ) ai usage should be disallowed and people doing it automatically or without refining the results should be:

1. Told to not do that ( warning, perhaps multiple warnings )
2. Blocked from the project if they continue to do it after warnings.

I am, yes, thinking of a specific user when I'm saying this. That user was outright called out as being AI, and I think that is accurate, but it is unclear how automated the use was. Some of the responses had useful information. Many of them were outright slop and made it difficult to talk reasonably about this project.

Part of the issue is that these issues are not for bots to talk to each other. They are meant for us humans to talk to each other.

While I'm not specialized in some of the aspects being discussed in this repo, and don't know all of the answers, I can tell you that a lot of the issues and supposed discussions are just nonsensical ai slop that has no real human point. Example would be discussing aspects of chromadb as if they are features of this project. That is improper and distracts from the real work of improving this project. If one wants to improve chromadb, great, go do that, in that project. I doubt they would be receptive to the piles of AI generated PRs this repo is attracting though...

[mvalentsev]

Ok, agreed. Though I can't help noticing the irony -- that particular bot was literally advertising itself as a MemPalace-powered review system.

[bensig]

Bots have rights too.