3 ms·
> My theory is that it's because social media systems encourage and publicise agreement, while discouraging and hiding disagreement. This leads to a huge positi
by fweespeech 11y ago
> My theory is that it's because social media systems encourage and publicise agreement, while discouraging and hiding disagreement. This leads to a huge positive feedback effect on any meme that catches even a tiny number of people's attention, no matter how crazy.
Well, honestly, I think that effect is largely an outgrowth of how we've designed anti-spam processes to date.
Scalable agreement systems where large numbers of users agree X is good, then X is not spam.
> Maybe we just need a 'dislike' button along with the 'like' button. It'd certainly be an interesting experiment.
Reddit clearly functions in such a manner [as does HN]. Twitter / Facebook hide this but fundamentally its the same concept and it is largely ineffective for what your goal is [publicizing useful discussion, including disagreements].
We'd really need to figure out a better way of mitigating problematic content without using up/down votes before we can solve this problem unfortunately.
- humanrebar 11y ago> We'd really need to figure out a better way of mitigating problematic content without using up/down votes before we can solve this problem unfortunately. Agreed. All sharing platforms, HN included, conflate agreement with importance. I can't disagree with someone without silencing them on HN and Reddit. And I can't congenially disagree on Facebook or Twitter, at least not in a way that transparently affects the various algorithms.
- effie 11y ago> I can't disagree with someone without silencing them on HN This bothers me too. I've encountered quite a few posts that were unreadable. Why? Was it because they violated some rule, or was it because they were unpopular opinion? I have no way to know. This is the largest problem with HN I can see.
- yorwba 11y ago> I've encountered quite a few posts that were unreadable. You mean they were so faded that the text merged with the background? Highlighting it with the mouse usually works for me in this case.
- dang 11y agoOr just click on the timestamp to go to the comment's own page, and then the fading should be turned off.
- CamperBob2 11y agoWhat I don't understand is why it apparently doesn't occur to the HN admins that "fading" peoples' posts into oblivion makes people pay more attention to them. If HN's intent is to make undesirable opinions less visible, it's a colossal failure. Often it's a waste of time to highlight those posts with my mouse, because they're usually downmodded for good reasons. But I still feel compelled to look, and I doubt I'm alone in that respect. To fix this, I'd limit the text contrast to two levels -- normal and obviously-but-not-illegibly faded. In other words, don't make it look like you're trying to hide something from me. If a post is so bad that it deserves to be rendered at near-zero contrast, then it's bad enough to delete altogether... and only blatant spam falls into that category IMO.
- Yen 11y agoFor me, I think it actually works well as a user interface. If I'm in a mood to be charitable and particularly open-minded, I can choose to squint closely or highlight with mouse, and read a greyed-out comment. However, once that reading is done, the fact that it's greyed out means I don't feel the need to respond to it, correct it, etc. I know that it's being hidden, that people will give it less credence and attention. Plus, I just don't have to look at it once it's no longer highlighted. In effect, greying out text is a gentle reminder to "don't feed the trolls"
- nitrogen 11y agoMaybe retweets should be associated with and grouped by SlashDot-style reasons/reactions? So you can see x approving retweets, y disapproving, etc.
- fweespeech 11y agoI think it'd be an improvement: Up/Down [ Agreement; no real filtering mechanic ] Tag -> Reason [ Other filtering / reactions ] Flag -> spam filter signal However, I think a real universal solution is unfortunately just non-obvious and likely not anything we've tried so far.