Hacker News Reader

Comment by rottencupcakes | Hacker News Reader

Twitter has surprisingly become much better since Elon purchased it. The community tags are great and non-partisan, they fairly report on all inaccuracy.

Even advertisers aren’t immune, I’ve seen community tags on paid ads calling them out for being drop shippers.

View on HN

I have had the opposite experience with X, I find it hard to escape any cheap rage bait on my feed regardless of how I would tag or hide pages. It’s interesting you have this experience, though.

Pointing out yet again that Community Notes long precedes Musk's purchase of Twitter. It got going in early 2021; I was one of the first batch to sign up for it, although I think it had been in developments for a year or so before that. January 6 rattled a lot of people at Twitter so I suspect that influenced the timing of the rollout.

Perhaps the confusion is that CN was originally called 'Birdwatch' and people don't realize they're the same thing.

Notably, Community Notes (formerly Birdwatch) was released in Jan 2021. Elon just flipped the feature flag.

Worth noting Birdwatch launched alongside the ability to report tweets for misinformation, a feature Elon disabled. For that reason it’s difficult to conceive of Community Notes today as a simple continuation of the Birdwatch effort.

I think he was right to disable the ability for Community Notes to actually report and remove tweets. First, it's a vector for abuse. Second, misinformation is a loaded term and having the context along side the content is more important for an informed public.

Of course anything illegal or against ToS should be removed and only be the purview of Twitter's moderation team (or what's left of it).

I guess you haven't read any news for a while.

Advertisers are fleeing Twitter and nobody knows how much longer Twitter will exist, because, well, they don't want their ads to be displayed next to Nazi content, in any circumstances.

If you find Twitter better, good for you for swimming in that kind of content.

In a 2 day span I saw George Soros and Tucker Carlson get hit by Community Notes.

Gave me some hope that Elon is serious.

Musk has criticised Community Notes a few times for calling him out.

More often he has expressed how well it works and how great it is.

He praised it many times, but just a few weeks ago was complaining when he posted BS and community notes essentially called him out. In any case, I think it's a good feature and hope it doesn't go away.

> get hit by Community Notes

How does that work? can anyone post a "community note", or by what process is that decided?

https://help.twitter.com/en/using-x/community-notes

Thanks! But I didn't find a clear description there of how this works. What elevates an ordinary comment to a "community note" and what determines whether a "community note" stays up as a sort of verdict on the original post? This seems like a hard thing to get right at scale. If you rely on voting (i.e. likes or whatever), you'll just end up with a parallel comment system, no more authoritative than the original one. So there must be some other process deciding which proposed notes get this special status.

(Perhaps I should add: yes, I'm being lazy; no, I'm not being critical - I just want someone to explain this to me so I don't have to work to find out.)

Be a verified (by phone number, no need to pay) twitter user. Apply to be a Community notes contributor. Wait. If accepted, you start with 0 points. Twitter will start to show you proposed (as opposed to published) community notes. You can vote yes / no. If a note you voted yes (or no) will get enough other yes (no) votes to get published (rejected), you will get 1 point. If you voted "wrong", you will lose a point. After you have a certain number of points (I think 5), you can also write proposed notes yourself, not just vote. If your note gets rejected, you lose 5 points.

But twitter is not only counting the yes / no votes. They need people who have voted differently on some previous notes, to agree in their votes on this note, before their algorithm makes the decision to publish or reject the note.

A lot of proposed notes will never get enough votes to be decidedly published or rejected, so they will linger in the "proposed" stage forever.

Thank you! That's clear and interesting.

Some related HN discussions:

https://news.ycombinator.com/item?id=25906672

https://news.ycombinator.com/item?id=25908439

https://news.ycombinator.com/item?id=33478845

Thanks - it sounds from https://blog.twitter.com/en_us/topics/product/2021/introduci... (via your first link) that they have a rating system whereby community members get to vote on which notes they think are true (a.k.a. which ones they like), and some sort of reputation system to compensate for the weaknesses of relying solely on votes. Did they publish the details of these systems?

Edit: this, from https://communitynotes.twitter.com/guide/en/about/introducti... (via your third link) is interesting:

Community Notes doesn't work by majority rules. To identify notes that are helpful to a wide range of people, notes require agreement between contributors who have sometimes disagreed in their past ratings. This helps prevent one-sided ratings.

Then they link to https://communitynotes.twitter.com/guide/en/contributing/div..., which expands on that. It seems they're basically trying to control for ideological perspective, i.e. to identify signal that doesn't just boil down to "I like / agree with this". I've often wondered if something like that could work.

>Did they publish the details of these systems?

I didn't look into the details myself, but there is a 24-page paper linked from this page:

>Details of our methodology and past findings can be found in our paper.

https://communitynotes.twitter.com/guide/en/under-the-hood/g...

https://github.com/twitter/communitynotes/blob/main/birdwatc...

Reply on news.ycombinator.com