> The social media mogul said Tuesday that Facebook and Instagram will shift to a community notes model
I'm curious, what are HN's opinions around community notes on X/Twitter? I find them to be pretty good (and sometimes quite funny) but I wonder what others think.
Ultimately: It can take some time for community notes to correct misinfo posts; by the time community notes gets to it, the firehose has ten more misinfo posts community notes needs to fix, the people who saw the original post rarely see the community notes correction, and given the algo has a massive temporal bias the original post was probably on the way out of new viewers feeds anyway.
Its a good system that should exist. It could probably be made better by finding any way to help it act faster (e.g. the moment a post is even in the process of community notes voting, attach a temporary banner like "The Community Notes people has suspicions about this post, it might be misinfo, we're doing our research").
I am not sure about that. I see a lot of posts that get CN proposed for bogus reasons that never see the light of day. This is usually done by interest groups who view CNs as another avenue of their strategies. These interest group CNs almost never see the light of day so I would strongly advocate for not putting warnings just because a rando attempted to put a CN on a post.
Imagine every post by Donal Trump getting a CN from a Democratic activist group or vice versa, every post by Kamala getting a CN by a Republican activist group. Every single post will have a warning on them if you get what you are proposing.
The best thing about community notes IMHO is that you get notified if something you liked previously has been community noted. Gives you a chance to reassess.
> I'm curious, what are HN's opinions around community notes on X/Twitter?
I notice that there are influence groups who are trying to play them, but so far I haven't seen them be successful. I see tons of proposed notes from certain groups but rarely do I see them actually be shown on X, so I guess it is robust in a way.
The main place it doesn't work is on popular polarizing accounts. For example, generally the followers of Elon Musk like him so if you try to correct him there, a bunch of his followers will just say the note is wrong even if it isn't, they vote out of loyalty to Elon. So polarizing large popular accounts generally cannot be community noted. Same applies to most political leaders.
So community notes is for correcting plebeians, not for the patricians.
Yikes. That almost seems worse than doing nothing. The presence of community notes for plebeian matters will make it seem like politically popular figures are never in the wrong.
Yes, it's useless on really polarizing topics like Israel/Palestine, Russia/Ukraine, immigration, and Elon Musk.
But it's really good when it's nerds debunking other nerds. It could maybe even be something that future peer-review systems can take inspiration from.
I'm not really a twitter user (I just get posts from friends occasionally), but I think community notes is actually an awesome feature for a social network.
Community notes don't work when there are small number of users for a particular topic.
For example, Indian Twitter/X user were posting with old/unrelated/fake photo/video saying Hindus are being killed in Bangladesh (after change of fascist government that was under controll of the indian government), hindus homes were burnt. None of them are true, but there are not enough Bangladeshi Twitter/X users to community note them.
I'm really giving Bluesky a shot and I want it to succeed, but it really needs to be something more than a stronghold for the left. Maybe I've been put into some bubble already, but the vast majority of posts there are political. Some of the people there as rabid as their right-wing counterparts are on X.
I find it really hard to find posts that are not people raging/venting about Trump, Musk, some other right wing billionaires or left-wing cause. X is, ironically, more diverse on this front.
If the goal is to prevent, or limit the spread of "misinformation" community notes is a failure. I see community notes merely as additional context on what is now a wide open machine for spreading any information, regardless of truth.
I think they're potentially good, but are used nowhere near enough. So many posts on X simply do not have community notes attached, even though they contain misinformation and lies.
We do need to be able to discuss divisive topics on social media, and any single organization determining truth faces an impossible challenge. We shouldn't rely on the goodwill of Meta or the government as the basis for our trust on social media.
Community notes-style systems are... similar. Instead of a single group of people, we rely on a single algorithm/system to determine truth. At a minimum, such a system NEEDS to be open source. But even if it is, I seriously doubt any one algorithm will ever give us the low-effort high-accuracy truth heuristic we desperately need.
We need better models of trust/truth that support a rich ecosystem of personalized approximations of whether to trust or believe something.
Honestly, I think the real solution is far, far simpler: just have social media operate more 'privately'. My Bluesky feed shows me content from people I follow, that's it. I don't get extreme opinions forced upon me and, if I did, I would simply unfollow that person. It doesn't need any censorship in the same way email doesn't.
> “We’re going to work with President Trump to push back on governments around the world that are going after American companies and pushing to censor more,”
Nothing says American Freedom than interfering in how other governments conduct policy so you can push your own platform full of fake users and perpetual lies.
What an embarrassment. Maybe it's time said governments start just banning or outlawing these platforms if they're just going to flagrantly admit to doing what they can to bypass the law. Fuck em. As much as I hate tiktok the whole 'rules for thee but not for me' shtick is old.
>Maybe it's time said governments start just banning or outlawing these platforms
They should and it's long overdue.
America, China, Russia and to a far lesser extent Japan and South Korea all have outsized political interference power upon other countries through the various social networks that operate out of them.
Any country with a desire to safeguard domestic politics against foreign interference and influence should be banning them without question and it should not be controversial.
Depends. My country doesn't moderate social networks in any way. Instead, it profiles people. Nationals who are not behaving get called, foreigners are turned in the border.
As expected, our posts are complying, you would think we live in Switzerland
Zuckerberg knows which way the winds are blowing in the US Capital and is ensuring he is aligned with them so to avoid political blowback on his company.
I suspect the changes to the fact checking / free speech will align with Trump's political whims. Thus fact checking will be gone on topics like vaccines, trans people, threats from immigrants, etc.
While the well documented political censorship at Meta affecting Palestine will remain because it does align with Trump's political whims...
It’s at times like this we realise how much better our situation would be if the internet remained decentralised and not consolidated into a few “social networks”. We messed up.
How would decentralization make a difference here? Fact checking only makes sense when a platform has people with multiple different beliefs about what is true. If people can self-select into their own bubble, then it doesn't do anything, does it?
Readit News
I'm curious, what are HN's opinions around community notes on X/Twitter? I find them to be pretty good (and sometimes quite funny) but I wonder what others think.
Its a good system that should exist. It could probably be made better by finding any way to help it act faster (e.g. the moment a post is even in the process of community notes voting, attach a temporary banner like "The Community Notes people has suspicions about this post, it might be misinfo, we're doing our research").
Imagine every post by Donal Trump getting a CN from a Democratic activist group or vice versa, every post by Kamala getting a CN by a Republican activist group. Every single post will have a warning on them if you get what you are proposing.
I notice that there are influence groups who are trying to play them, but so far I haven't seen them be successful. I see tons of proposed notes from certain groups but rarely do I see them actually be shown on X, so I guess it is robust in a way.
The main place it doesn't work is on popular polarizing accounts. For example, generally the followers of Elon Musk like him so if you try to correct him there, a bunch of his followers will just say the note is wrong even if it isn't, they vote out of loyalty to Elon. So polarizing large popular accounts generally cannot be community noted. Same applies to most political leaders.
So community notes is for correcting plebeians, not for the patricians.
But it's really good when it's nerds debunking other nerds. It could maybe even be something that future peer-review systems can take inspiration from.
For example, Indian Twitter/X user were posting with old/unrelated/fake photo/video saying Hindus are being killed in Bangladesh (after change of fascist government that was under controll of the indian government), hindus homes were burnt. None of them are true, but there are not enough Bangladeshi Twitter/X users to community note them.
Maybe the community notes are working as intended.
"Hindus in Bangladesh Face Attacks After Prime Minister’s Exit": https://www.nytimes.com/2024/08/07/world/asia/bangladesh-pol...
I find it really hard to find posts that are not people raging/venting about Trump, Musk, some other right wing billionaires or left-wing cause. X is, ironically, more diverse on this front.
We do need to be able to discuss divisive topics on social media, and any single organization determining truth faces an impossible challenge. We shouldn't rely on the goodwill of Meta or the government as the basis for our trust on social media.
Community notes-style systems are... similar. Instead of a single group of people, we rely on a single algorithm/system to determine truth. At a minimum, such a system NEEDS to be open source. But even if it is, I seriously doubt any one algorithm will ever give us the low-effort high-accuracy truth heuristic we desperately need.
We need better models of trust/truth that support a rich ecosystem of personalized approximations of whether to trust or believe something.
Personally I find that 10x better than someone just deciding I shouldn't be seeing this.
Nothing says American Freedom than interfering in how other governments conduct policy so you can push your own platform full of fake users and perpetual lies.
What an embarrassment. Maybe it's time said governments start just banning or outlawing these platforms if they're just going to flagrantly admit to doing what they can to bypass the law. Fuck em. As much as I hate tiktok the whole 'rules for thee but not for me' shtick is old.
They should and it's long overdue.
America, China, Russia and to a far lesser extent Japan and South Korea all have outsized political interference power upon other countries through the various social networks that operate out of them.
Any country with a desire to safeguard domestic politics against foreign interference and influence should be banning them without question and it should not be controversial.
As expected, our posts are complying, you would think we live in Switzerland
It must be utterly demoralising for anyone working at the US State department right now.
I suspect the changes to the fact checking / free speech will align with Trump's political whims. Thus fact checking will be gone on topics like vaccines, trans people, threats from immigrants, etc.
While the well documented political censorship at Meta affecting Palestine will remain because it does align with Trump's political whims...
- https://www.hrw.org/news/2023/12/20/meta-systemic-censorship...
- https://www.theguardian.com/technology/article/2024/may/29/m...
- https://theintercept.com/2024/10/21/instagram-israel-palesti...