Idk if zero tolerance is the way to go, but conceptually at least allowing things with “significant” AI usage is kind of horrifying to me. The entire rest of the fucking world is completely uncritically embracing it so i see no reason why places like these have to allow it as well, i think it should be obvious that we want some communities that reject AI usage so we have control groups and safe havens for the future…

No, I think that there should not be a anti-AI rule. People can downvote those post if they are anti AI.

I think they should be allowed if they have actual effort and maintaining put into them, but that’s usually the problem. Most LLM projects get abandoned very fast.

This is a major difference with a project like Linux. Even if they start using slop code there, you know that they’re not going to suddenly abandon it.

Maybe we need to think out of the box, and instead of an LLM tag, we need a tag like how long the repo has existed which can be automatically checked by a bot.

Require an AI tag for anything written in part or in whole by AI.

Remove them

Just make people disclose it in the title

Minimum age requirement and activity requirement. If it was built all in one go it’s more likely to be abandoned and full of problems and just pollute the landscape. If it’s been under development with multiple commits over a time-span of months or a year or more it’s more likely to be higher quality, have human in the loop somewhere, continue development, patches, etc. Linux itself for example would be exempted because it has a significant age, significant recurring activity in development, multiple human developers, etc.

I do think exceptions on a case by case basis for clever simple programs that address a not previously addressed need might be fine as well. I really think there needs to be a little discretion for mod action.

What a lot of people hate and I do as well are these “look daddy I can ‘code’ too” baby’s first/second/etc vibecoded programs that people proudly toss out there without any real thought beyond pride they told an LLM to make something and it did it. Those rarely add much value, often are redundant with one another (why reinvent the wheel again and again?), and would tend to be lower quality and end up just being so much spam that makes it harder to understand what’s worthwhile.

remove

Remove.

Require a tag but if someone wants to share something that solved a problem for them as an open source project they should be able to.

Remove them

If it’s not open source, remove them.

If they’re open source then the people posting them are in the right place.

A policy like what’s being suggested in the op will only be walked back in a few months.

E: also disclosure is a stupid requirement. All it does is give people something to fight over and there’s no industry accepted process for evaluating ai code as opposed to human written code, so disclosure doesnt even serve a useful purpose. It’s literally just giving people something to argue about when the hypothetical post is ostensibly about the software or what it accomplishes.

There needs to be an actual framework for what you’re excluding and why. The majority of developers use LLMs to some degree, an outright ban is unrealistic and motivated by emotion not logic.

I’d suggest limiting submissions of brand new projects with no history and nothing to indicate they’re going to be maintained going forward. There also needs to be some proof the maintainer actually understands how their own project works but that’s far more difficult to prove.

If you’ve been working on a project for 6 months and plan to continue doing so then I’m happy to learn about it regardless of whether AI was used in the process or not.

All these “remove” votes will only accomplish developers not disclosing their use of AI 🙄 It’s very silly.

I vote to remove myself but I’d be fine with just having to state AI involvement