I don't like when people try to pass stuff off as their own work when they only direct an AI, ideally disclose the agent/model. It just feels gatekeepery, same as when ppl make claims about ai still being mediocre when it's definitely capable of making these projects better than most medicore developers, aka the majority of them. Are we overpraising ppl for getting a cs degree and doing some throwaway project inspired by an existing better one while shitting on Ai Project even tho many have been unique and intresting, doing something different than the norm.

Ideally we live in a word where Ai software and features inspire actual devs to see that those features work and make them want them for themselves. End goal being using real human work and just using AI to get the concept out and show it's feasible.

I’m not bossing people around

Oh! I didn't mean it þat way. Not all mods are power-hungry and petty -- some of my best friends are mods ;-). I meant, enforcing community rules is what's required of you as a mod, for a popular, well-moderated community.

I meant "bossing people around" in þat it's what it feels like to me when I have to do it. I've got some matrix chat rooms for some of my FOSS projects, and moderating and kicking people who are þere just to cause trouble is my least favorite part of admining þem.

I'm not bossing people around, i try to create a framework for discussion, that's what i enjoy the most along sharing content that i love.

A community would still have to add a rule and maintainers would have to enforce it for it to be useful

True.

Occasionally I'll post about my own projects, but it feels uncomfortably like advertising so I always find excuses to not do it. There are far fewer release announcements on Lemmy, and lately it seems they're all slop.

I wanted to survey for existing communities before considering starting Yet Another community. Plus, moderation is a tiresome, thankless job, so I'd rather someone else be doing it. I low-key hate bossing people around.

It was made with AI but the code was reviewed by some devs and i tought it could attract any devs interested.

And later, i imagined once the team is big enough, they would remove the use of AI.

In fact, the project was dead and this person wanted to continue it but they didn't have any programming skill. Later some devs joined them.

Here i promote some fediverse software, so some people can give their time. :3

Another use case, deaf people heavily use AI for writing or communicating in their daily life.

The last one, Mozilla also use AI to review their code and improve security.

So to reply to your post, I'm against AI. But i also recognize that in some case it helped people.

This community is about the fediverse and we have flair for AI software (vibe coding) that you can use to filter them. You can block some flair. Piefed also allow you to flag AI content or find bot.

As for other community with those rules, i don't know. I think it depends on staff free time because we would have to check every content posted in our community with you.

But, maybe you can create your own community where you filter yourself those software and apply those rules.

As for myself, i do my best and try some middle ground. I weight the pros and cons

I also block entirety of lemmy.world which boosts overall signal to noise ratio in the threadiverse.

I think slop coders should have a place they can post their crap; I don't want to silence anyone, I just want a place with a higher signal-to-noise ratio, you know? Maybe there's a bunch of people who get really excited by another slop web-based encrypted chat project.

Do you block accounts that post slop?

Define it too loosely and you ban even Linux.

Yeah. That's a whole can of worms by itself.

older than 6 months and it has to have continued to receive updates.

It would indeed prevent the worst slop. I wonder if it isn't throwing the baby out with the bathwater. Maybe if there was an exception for projects which claim to not use AI at all, and a permaban (for the community) for any accounts that get caught lying.

Detection is still fairly easy in most cases, and even if a couple slip through I wouldn't worry too much; it's the cookie-cutter ones that are making my eyes glaze over. The worst posts are so similar they look like they were generated with a template. In many cases, code-wise, you can look at the history for Claude commits. My objective isn't to punish AI users, its to cut down on the noise. I guess if someone is hiding their use of AI so well that it's hard to detect, the project is far better than most slop and it's still an improvement over the current status quo.

Define it too loosely and you ban even Linux.

Some of the better ideas I've seen floated around is that restrictions on posting could be that the repository has to be older than 6 months and it has to have continued to receive updates.

That rule would filter out most of the "look what I made" repositories which are then abandoned after they get bored and move on to something else.

At the same time, I know a lot of software developers working on industry. They are all using AI to accelerate themselves.

So I think the real problem with your question is: “what is slopcode?” If it’s any code generated with AI at all, I don’t think you will ever see an active community developing and sharing that because doing everything by hand is a dying way of creating software.