3 ms·

I'd love to see a site like Reddit or HN experiment with using relationship graphs to research this. Start with some fixed points of "this person is a credible

by Pxtl 5y ago

I'd love to see a site like Reddit or HN experiment with using relationship graphs to research this. Start with some fixed points of "this person is a credible source of information" vs "this person is a neo-nazi conspiracy theorist" and examine the networks of people who support them and oppose them, and use that to establish credibility-scoring system-wide and weigh people's votes based on that. I mean, it would be super-biased but at this point I'm all about systems that are biased in favor of scientific consensus and human rights. Yes, you could reverse the "credibility" of those paragon users and make a system where the people surrounding the neo-nazi conspiracy theorists are the ones who hold all the power but we already get sites like that without any moderation or voting at all so that would be redundant anyways.

anigbrowl 5y ago
I'm very into that (because I spend a great deal of time studying neo-nazi conspiracy theorists and similar people). I think platform operators to some extent have to eschew moral judgements and let things play out, eg if a variety of distinct and persistent clusters emerge naturally, and 98% of people in a community make nazis' lives miserable for being in the 'nazi cluster' then it doesn't need any additional help from the platform operator. Of course the nazis will whine 'so much for the tolerant left' but honestly, who cares. The key is that if enough people in a community tag individuals participants in a particular negative way, those individuals can either accept it (and then be habitually ignored) or change their ways. This isn't always a moralistic dichotomy. For example I post a lot of environmental stories to HN, which are not-interesting to some readers who prefer to focus on pure technology. I'm OK with them ignoring such posts or even downvoting them/flagging if I make too many in a row. Likewise there are many technical topics of wide interest on HN which I find deathly boring, but which I am happy to ignore. Researchers at Stanford have come up with some results on clustering and conflict in online spaces, which you might find interesting: https://snap.stanford.edu/conflict/ https://snap.stanford.edu/conflict/
- raxxorrax 5y ago
  > because I spend a great deal of time studying neo-nazi conspiracy theorists and similar people Really sorry to hear that. I hope it is a hobby and not your job. I remember when Nazis where a joke on the internet. That they are not quite that anymore is not completely their own making as there are some people obsessed with them and even start to mirror them. Or others that want to build a profile about standing against the obvious. Suddenly there is a significant group that denies civil liberties like free speech or freedom of association. Free speech in interesting here, because it seem to trigger those that fear that Nazis are overtaking the internet. To me this is very indicative of reactionary behavior towards a vastly overestimated threat.
mrighele 5y ago
What about clustering them by upvote/downvotes patterns (I.e. find the bots) and then “shadowban” their votes? (I.e. make them count as one or in any case a fraction of the total) Maybe they would be hard to distinguish from regular user clustered because of group thinking , but even in that case I wouldn’t consider it a completely bad thing… :-)
pessimizer 5y ago
Make it even simpler. Mark yourself as the only source of karma. Then when you upvote people, they also become sources of karma; and when you downvote them, they become sinks (their upvotes might actually subtract karma.) This effect should attenuate with distance from you (i.e. your upvote is +1, the person you upvoted's upvote is +0.9.)
- AnimalMuppet 5y ago
  That sounds like a recipe for creating filter bubbles.
  - hutzlibu 5y ago
    It is. But diverse bubbles. I think this is still better, than just one bubble, even if the intention is a "pro-science" bubble. Because there are not many people claiming to be opposed science. Allmost everyone claims to be "pro-science", even flat earthers. So it still comes down to what the specific mod thinks.

by Pxtl 5y ago

anigbrowl 5y ago

I'm very into that (because I spend a great deal of time studying neo-nazi conspiracy theorists and similar people). I think platform operators to some extent have to eschew moral judgements and let things play out, eg if a variety of distinct and persistent clusters emerge naturally, and 98% of people in a community make nazis' lives miserable for being in the 'nazi cluster' then it doesn't need any additional help from the platform operator. Of course the nazis will whine 'so much for the tolerant left' but honestly, who cares. The key is that if enough people in a community tag individuals participants in a particular negative way, those individuals can either accept it (and then be habitually ignored) or change their ways. This isn't always a moralistic dichotomy. For example I post a lot of environmental stories to HN, which are not-interesting to some readers who prefer to focus on pure technology. I'm OK with them ignoring such posts or even downvoting them/flagging if I make too many in a row. Likewise there are many technical topics of wide interest on HN which I find deathly boring, but which I am happy to ignore. Researchers at Stanford have come up with some results on clustering and conflict in online spaces, which you might find interesting: https://snap.stanford.edu/conflict/ https://snap.stanford.edu/conflict/

raxxorrax 5y ago
> because I spend a great deal of time studying neo-nazi conspiracy theorists and similar people Really sorry to hear that. I hope it is a hobby and not your job. I remember when Nazis where a joke on the internet. That they are not quite that anymore is not completely their own making as there are some people obsessed with them and even start to mirror them. Or others that want to build a profile about standing against the obvious. Suddenly there is a significant group that denies civil liberties like free speech or freedom of association. Free speech in interesting here, because it seem to trigger those that fear that Nazis are overtaking the internet. To me this is very indicative of reactionary behavior towards a vastly overestimated threat.

mrighele 5y ago

What about clustering them by upvote/downvotes patterns (I.e. find the bots) and then “shadowban” their votes? (I.e. make them count as one or in any case a fraction of the total) Maybe they would be hard to distinguish from regular user clustered because of group thinking , but even in that case I wouldn’t consider it a completely bad thing… :-)

pessimizer 5y ago

Make it even simpler. Mark yourself as the only source of karma. Then when you upvote people, they also become sources of karma; and when you downvote them, they become sinks (their upvotes might actually subtract karma.) This effect should attenuate with distance from you (i.e. your upvote is +1, the person you upvoted's upvote is +0.9.)

AnimalMuppet 5y ago
That sounds like a recipe for creating filter bubbles.
- hutzlibu 5y ago
  It is. But diverse bubbles. I think this is still better, than just one bubble, even if the intention is a "pro-science" bubble. Because there are not many people claiming to be opposed science. Allmost everyone claims to be "pro-science", even flat earthers. So it still comes down to what the specific mod thinks.