There is clearly a problem that most of the politics and news communities on Lemmy are unpleasant places to take part in discussion. People yell at each other. The tone of disagreements is that of saying what your opinion is, and insulting the other person if they don't agree with your opinion, or a bunch of people giving quick one-off statements like "well I think it's this way" or "no you're wrong" which adds nothing. I've heard more than one person say that they simply don't participate in politics or news communities because of it.
Well, behold:
I have made some technology which attempts to take a much heavier handed approach to moderation, by just detecting assholes or people who aren't really contributing to the conversation, in other communities, and just banning them pre emptively en masse. In its current form, it bans about half of hexbear and lemmygrad, and almost all of the users on lemmy.world who post a nonstop stream of obnoxiously partisan content. You know the ones.
In practice it's basically a whitelist for posting that's easy to get on: Just don't be a dick.
I'd like to try the experiment of having a ~~political~~ community with this software running the banlist, and see how it works in practice, and maybe expand it to a news community that runs the same way. There's nothing partisan about the filtering. You can have whatever opinion you want. You just can't be unproductive or an asshole about the way you say your opinion. And the bans aren't permanent, they are transient based on the user's recent past behavior.
(Edit: I think making a general news community might fit better with slrpnk than politics. In thinking about it and talking with people, I think electoral politics just doesn't belong in the slrpnk feed, but maybe general news specifically with the political bickering that comes along with it being muted, would be a positive for the instance at the same time as I get to test out my little software project.)
I don't want to explain in too much detail how the tech works, because I think some segment of assholes will want to evade the tech to come into the community and be assholes again. But I'd also want to set up a meta community where anyone who did get banned can ask questions or make complaints about it. (As long as that offering doesn't turn into too much of a shit show that is.)
Is slrpnk a place where a little experiment like this could find a good home? What does everyone think of the idea?
I have concerns about your vision of an ideal community, and I'm cynical of how far technical means can go in achieving that vision, but those concerns are overwhelmed by my support for experimentation. I agree with the prevailing opinion that moderation on Lemmy is hamstrung by a lack of adequate tools. Your project, even if it fails to achieve your vision, could serve as a stepping stone to some future success.
My primary concern is that you may be filtering people into whitelists and blacklists by feeding their comment history with a prompt into a Large Language Model like ChatGPT. If that's the case, it is a deal-breaker. You cannot submit content via an LLM API and also avoid having that text absorbed by the model as training data. Since you would be submitting the comments of other people, this violates the principles of respect and consent. Many people exited corporate social media for Lemmy to protest this hoovering of their data by 'AI' companies; while some have gone as far as to add an anti-AI clause as a comment footer, it should be assumed that every Lemmy commenter does not consent to their intellectual labor being exploited for the profit of tech capitalists unless they explicitly state otherwise. If SLRPNK endorsed a moderation tool that abused other Lemmy users in this way, we would quickly become a pariah instance.
When it comes to software, I'm a fan of transparency. I hope at some point you're willing to share your code, though I acknowledge your reasons for keeping it obscure. I would advise you to be open at least about the mechanism your filter uses while hiding your parameters if you can, so that you can alleviate any concerns that your code is feeding Lemmy comments to an LLM.
Perfectly reasonable. It's not feeding any users' comments into any LLM public API like OpenAI that might use them for training the model in the future. As a matter of face it's not communicating with any API or web service, just self contained on the machine that runs it.
As far as transparency, I completely get it. I would hope that the offer to point to specific reasons for any user that wants to ask questions about why they can't post will help to alleviate that, but it won't make it completely go away. Especially because as I said, I'm expecting that it will get its decisions wrong some small percentage of the time. I just know there's an arms race between moderation tooling and people trying to get around the moderation tooling, and I don't want to give the bad actors any legs up in that competition even if there are very valid reasons for it in terms of giving people reasons to trust that the system is honest.
Other things that have occurred to me in the meantime: