Live data from Hacker News

Twitter tests Reddit-style upvote and downvote buttons

techcrunch.com

121–130 of 400 posts

Re: Twitter tests Reddit-style upvote and downvote buttons

#121
post #54

This seems prone to trouble to me. Twitter is already very groupthink heavy, and I can only see this amplifying that problem. You see it in Reddit all the time - people don't actually read the comments, just upvote the upvoted ones and vice versa. I worry this will just make an already very narrow set of acceptable opinions even narrower.

I think up and down voting could actually help the discourse on twitter if it's implemented correctly. Currently every way to interact with a tweet drives its engagement score: likes, retweets and replies all result in more visibility for a tweet. So if you think a tweet is unproductive or wrong, you can express that in a reply, but even that is resulting in more visibility for the original tweet. This incentivizes b…

I too think negative signaling is valuable, but I doubt Twitter will ever correctly implement that. This is evident from the state of "report tweet" menu. Reporting can be thought as a strong negative signal that happens to be also ToS violation, but its implementation is far from working; not just political figures, but I've seen even direct death threats against minorities with numerous reports ignored multiple times. Twitter seems to have no incentive to fix reporting and I don't think downvotes will be different.

Re: Twitter tests Reddit-style upvote and downvote buttons

#122
Suppose you're someone who identifies as a member of an activist group. Your goal is to spread ideas. You can't control the spread of a shared tweet but you can control the comment thread by voting en masse. Every popular tweet is then accompanied by top-ranking comments that your group controls.

Re: Twitter tests Reddit-style upvote and downvote buttons

#123
post #107

Earlier quoted context omitted.

> For example, if the #1 person believes something is being said that is really damaging in some way, they could put at stake the entirety of their karma and downvote it, which would pretty much obliterate the comment. We are building a P2P network with subjective moderation [0], so while we don’t know what works yet (we’re experimenting) I can at least tell you why we’ve found this is suboptimal. The issue is that t…

You've probably thought of this, but if you can associate votes to identities then a classic way to mitigate this behaviour is quadratic weighting, where e.g. your nth vote costs n^2, so a censor or propagandist needs to pay 1e4 counteract the single votes of 100 casual observers.

Yeah, quadratic voting is one of the potential solutions to this, but ultimately that removes the ‘fraud load’ by mostly moving it to another place, to number of accounts. That is arguably worse because it incentivises the malicious users to create as many accounts as there are votes so as to not be subject to quadratic voting (1 vote = 1 energy point equivalency only holding at one vote), and it’s substantially harder to deal with user account fraud than vote fraud.

The goal is to minimise the fraud load and make whatever fraud load you have applied against the strongest parts of your system. It’s like designing a castle except with incentives - it has to fail gracefully and still be functionally successful even in failure.

Re: Twitter tests Reddit-style upvote and downvote buttons

#124
post #90

Earlier quoted context omitted.

I'd say that it's perhaps more tied to moderation. Reddit communities vary dramatically in quality of discussion and one thing that seems to separate the highest quality from the lowest is heavy moderation. AskHistorians for example is extremely strict and has very high quality answers. Another example that's strict in a different way is CasualUK, where even vague references to politics or politicians is an immediate…

I think Reddit might be fundamentally flawed because of its voting system and its need for high moderation to make it workable. Mods influence so much power over what people are and aren't allowed to talk about. You use the example of CasualUK which is indeed a very good if slightly "police state" sub but it only exists because the other UK subs became unviable. The old defence has always been "if you don't like it g…

> I think Reddit might be fundamentally flawed because of its voting system and its need for high moderation to make it workable.

Perhaps. Or perhaps it's groups of people that are fundamentally flawed.

> The only positive change i can think of is to cap active participants in a sub to a specific small number.

Which would break a significant number of use cases of reddit.

Whatsapp groups have no voting, typically pretty limited numbers of participants but can be hot flaming piles of garbage and a huge source of misinformation spreading around the world.

Re: Twitter tests Reddit-style upvote and downvote buttons

#126
post #85

Earlier quoted context omitted.

Excellent, more bubbling.

That's my worry too, but I don't know if you can already make it worse on twitter. Reddit has the random button and easy access to two "feeds" (chronological and "personalised"), so I'd like more of that.

That random button is just the top 1000 most popular subreddits by the way. Better than nothing, but still.

Re: Twitter tests Reddit-style upvote and downvote buttons

#127

I've wondered why Twitter seems far more toxic than Reddit (certain now-banned subreddits excluded) and I believe the downvote is a powerful tool used to achieve this, along with Reddit's community-style moderation, even though that can border on toxic itself at times. I wrote about it here a little while back https://unlikekinds.com/article/trolls-and-consequences-what...

Twitter exposes the profile of the user a lot more and people have their own feeds. Reddit in comparison albeit with profiles seems anonymous all the same.

Re: Twitter tests Reddit-style upvote and downvote buttons

#129

upvote/downvote system is flawed allows to manipulate, promotes groupthink and kind of harassing people it maybe worked over decade ago or so, but it's definitely not state of art nowadays. edit. or maybe it just doesn't work for e.g news/discussions (especially about politics and other "hot" topics) meanwhile for meme sites it should be fine

> but it's definitely not state of art nowadays. What would the better version be? I can only think of chronological order (the most active/noisy wins) or any kind of engagement metric like number of retweets/hearts/followers, that it's basically a worse version of arrows since it only measures engagement without caring if it's positive or negative.

>What would the better version be?

in my opinion?

Forums with mature community and competent moderators.

Forums have way better model for discussions - better than fb, reddit, hn, twitter.

Discussions on forums can last weeks / months, meanwhile on reddit/hn when thread drops from front page and then 20 hours pass, then comment section under thread is dead.

But it doesn't scale

Re: Twitter tests Reddit-style upvote and downvote buttons

#130

I've wondered why Twitter seems far more toxic than Reddit (certain now-banned subreddits excluded) and I believe the downvote is a powerful tool used to achieve this, along with Reddit's community-style moderation, even though that can border on toxic itself at times. I wrote about it here a little while back https://unlikekinds.com/article/trolls-and-consequences-what...

I'm wondering about the same thing. To me, Reddit feels more like it emphasizes communities. While Twitter is more about the cult of personality, and not too centered around community. Also, on Twitter it's much easier to get to another social bubble by accident (just see some random tweet in the feed, and here it goes). On Reddit, you are much more likely to stay within your chosen subreddits (unless you look at the most popular, news, etc.)
Post reply on HN