Curious if this will ever become a concept to help put a stop to spam/bots.
While karma might help spam/bots in some ways, I feel like it would also lead to karma farming, which I’m personally happy to not have here. Maybe they could instead allow communities to set requirements for minimum time subscribed or minimum interaction (voting, commenting, etc.) before people could post? I’d prefer that be set per-community, though, and not a site-wide mandate.
What about a tier system, something like “excellent / great / good / fair / neutral / cautious / warning / negative”? Hide the actual number from the user, so there’s no ‘mine is bigger than yours’ contest.
That might be better, but I’d still worry about people rep farming (for lack of a better term). Any time you give people a score, title, or other personal metric, you run the risk of people posting to influence that metric rather than to post for the sake of contributing content.
It’s possible the good such a system could do would outweigh the bad, but it will definitely always have elements of both.
No thank you. We already have enough reddit reposts, let’s not give people further incentive to spam the same shit we’ve all seen a thousand times.
No thanks. Happy to have left karma farms. Upvote wh*res. Hive minds. Karma gates and shadow bans. Keep this place open to diversity of opinions.
Agree. At least I can have an opinion here without killing my account entirely.
Karma was useless for that.
Account age helped some.
But, mostly, active moderation is what made spam controlled. You can’t totally do away with bots because the people running them will adapt to anything automated. But a small team of humans (or a single human on a small enough forum) can keep things to the point where any bot posts are taken down before long.
Mind you, reddit automod could help a shit ton. Once you figure out how to set it up, you can fine tune things. Hell, just the ability to make flairs mandatory reduces bots because it isn’t worth the time and effort for the bot runners to set up to auto flair posts. Some will, but not many (and never the usual scam bots, they target subs on reddit with bad moderation to begin with).
You could set up filters to target any spam a given sub would develop by keyword, urls, etc.
So far, lemmy doesn’t need that level of automation, and karma really wouldn’t help things as lemmy exists now. Karma just makes the barrier of entry a tiny bit higher because it’s easy to game.
If lemmy ever does hit hundreds of thousands of active users per hour, automation would be necessary, but the numbers aren’t high enough yet to make it worth the dev focus. And when it does happen, there will probably be user generated solutions anyway.
Right now, we’ve just gotta be willing to report the bots and spammers.
Hope Lemmy stays that way
Lemmy does already have that available on the backend… it’s a UI decision to not display it.
I don’t feel like it ever worked exactly how reddit envisioned. 100 helpful comments in small communities might end up with less karma than one well-placed joke in a major community.
I don’t feel like it’s super helpful as far as anti-spam, personally. Looking over someone’s most recent posts tells me enough, and communities that have karma limits are just annoying. People looking to spam can easily get around that (start a community, 10 accounts, vote each other up, or just make a few reposts of popular content) while it discourages legitimate new users.
Lotta good it did on Reddit. Eh? Karma provides motivation for botting.
Congratulations, you have a reputation of 1,427 as observed on kbin.social!
Kbin / mbin do expose reputation (karma) even for federated users. e.g.
https://kbin.social/u/@GreyTechnician@lemm.eeOkay, how on earth do I have positive rep?
Simple, this isn’t Reddit.
More complicated, Kbin doesn’t propagate downvotes from the Federation as a whole. It only counts the ones made on Kbin. Upvotes do propagate though, and reputation is determined by the ratio you get of upvotes to downvotes. So technically your Kbin reputation is inflated and not entirely accurate.
Only it is more complicated than that too …kbin has boosts as well as upvotes, and boosts count double, so reputation is:
Boosts x2 + upvotes - downvotes
and all of that is as observed by that instance, so much of your history could well be on communities the kbin instance doesn’t know and didn’t see.
Lemmy already has a system in place to only allow in approved members, why would it need karma?
Lemmy is federated. You can set up your own instance, create 100 spam bots, and go spam other instances. No account approval required.
but then those instances can simply defederate.
Yes, all 1000 instances can individually have their admins simply defederate.
Then the spammers make a new instance in 10 mins and start again.
The problem that karma attempts to solve is that you stop spammers before they can spam, not after.
I’m an instance admin, spam is a massive challenge at the moment and lemmy is still small. We need better tools before it grows too much.
Karma would only bring more spam and bots, so I surely hope not.
It didn’t work for Reddit, it won’t work for Lemmy. All it does is incentivize bots to spam AI-generated comments and posts before they launch their campaign of whatever malicious links they intend to spam. It delays the goal at the expense of an influx of a bunch more garbage posts. Might as well ban them right when they post the malicious links or whatever.
With how federation works, they really need Karma so they can set karma requirements
Like, if a troll gets banned from lemmy.world, there’s literally 100+ instances they can go and make an account on to post on lemmy.world again. Takes just a few minutes to get back at it.
There needs to be either a federation wide IP black list, or some kind of bot that notifies instance admins when a flagged IP register somewhere else, maybe even auto bans the new accounts for instances that have them flagged, but they can still post to the rest.
It’s not about stopping 100% of trolls, it’s about making it so inconvenient to get around that trolls get bored with it and troll somewhere else.
IP bans are a bad idea except if you want to like ban a country.
That could still result in purchased accounts due to time instead of karma. Idk if there would be an easy way to gate keep spammers without them getting around it.
@GreyTechnician I don’t think Lemmy is really interested in that sort of thing or at least not the Karma stuff as it really get’s toxic quick between users. If you like the idea of Karma there’s Kbin / Mbin instances that have reputation that is the same thing.
When it comes to bots though there’s not much you can do other than tell the admin of the instance to block them out. If they are running on another instance such as I’m on the Kbin.run instance then the admin can straight block that whole instance if it’s just full of bots. So really it can only be sorted with moderation.
I like that on kbin karma is not a highly visible number on profiles… on desktop at least, you have to look at the little stat bar and scroll over to see the number.
Yeah it doesn’t become the key part of the platforn but it’s a cool stat to see.
I think the only solution is, people reporting spam/bots accounts, and human mod review it.
No, objection.
Lemmy should not have the mech to censor, and should not have algorithm ranking/censoring, …
By having an upvote / downvote system it already has a censoring and ranking system.
If you look at most threads people downvote dissenting opinions regardless of the quality of their arguments, some instances become echo chambers that way.
You can easily control the narrative by upvoting someone that argues against you, most people will see your comment as
1 upvote 1 downvote
And the person arguing with you as
2 upvotes
Therefore they will be biased to downvote you further and upvote the other person as their social calibration defaults to thinking you are in the wrong as someone else said so already, it’s more economical that way as you won’t really have to do that assessment.
It’s a very interesting phenomenon and I’d like to do some formal testing with it across multiple social media with upvote/downvote functions, but I believe it does censor opinions that do not adhere to the usual feelings of the majority.
Just to point out that identities of upvoters / downvoters are visible on the fediverse. Though granted, you could use a proxy account for the voting
