Comment by lobocinza
1 year ago
> This means there was yet another update since I looked. The previous note at least made the user suspect that there was an attempt to point fingers at some other country. Now it merely corrects the location in China. It is good to link to a higher quality video, but I don’t rule out that the note will with time drift to suit the agenda of the government.
I didn't calculate any statistics but exploring the data I saw are way more "anti-China" notes than "pro-China" notes.
> He killed the feature that labeled accounts associated with governments though.
Yeah because Western government funded medias cried rivers when they were correctly labeled as such.
> If they applied the same algo to weighing tweets and replies, they could’ve gotten the same results but without making people trust blindly. But of course this defeats the point of paying for Elon’s blue checkmarks.
Doing so wouldn't make sense as the algorithm needs prior data from tweets to calculate ratings. What are your expectations? That Twitter hide (soft ban) tweets/accounts that an algorithm labels as misinformation because it was massively flagged? That happened before, it was easily abused, it was censorship.
> I think community notes are not his invention so I don’t blame him for them, but they are very poorly implemented and are strictly worse than tweets themselves.
You could give only one example where community notes were abused to spread misinformation and with time the correct note prevailed.
> exploring the data I saw are way more "anti-China" notes than "pro-China" notes.
Maybe that is the problem, it is seen as pro/anti X instead of facts/lies.
> You could give only one example where community notes were abused to spread misinformation and
If you ask this, you miss the point. How exactly do you expect me to tell a true note from a false one? The medium is the problem here.
> with time the correct note prevailed.
As long as it prevails before the heat death of the universe that’s OK, right?
> Doing so wouldn't make sense as the algorithm needs prior data from tweets to calculate ratings.
Twitter has prior data from tweets. I don’t get it.
The algorithm they use to present one community note can be used to capture feedback and sort tweets instead. Problem solved. People have better access to balanced views but are not being nannied by the platform or elonsplained what is truth.
> Maybe that is the problem, it is seen as pro/anti X instead of facts/lies.
What I meant is that there are more notes correcting pro-China lies than anti-China lies. So the removal of community notes would benefit pro-China propagandists.
> If you ask this, you miss the point. How exactly do you expect me to tell a true note from a false one? The medium is the problem here.
Use your head. Do your research. Trust your guts. I think you're expecting impossible things from technology.
> As long as it prevails before the heat death of the universe that’s OK, right?
No, it's not right. It should be quick but I have no way to tell how long it took for the better note to prevail. AFAIK it is quick enough.
> The algorithm they use to present one community note can be used to capture feedback and sort tweets instead. Problem solved. People have better access to balanced views but are not being nannied by the platform or elonsplained what is truth.
Nobody wants that. With community notes you can choose to ignore the context but you see the tweet. With your proposal people would just no see some tweets, they would lose agency.
> Use your head. Do your research.
Exactly. Just show me tweets well-sorted and let me decide. Don’t tell me “here is truth” (which is what “here is context” is intended to look like).
> right. It should be quick but I have no way to tell how long it took for the better note to prevail
You can’t. Because the implementation is botched.
> Nobody wants that.
Nobody wants community notes as is. Anyone who wants them only wants them because it’s a great way to disseminate disinformation.
> With community notes you can choose to ignore the context
Just like you could have chosen to ignore lies in the actual tweet without any community notes adding yet another layer of lies, only harder to ignore because it is now called “context” but really is just some guy’s opinion that won a popularity contest.
The context should just be a top reply, but then who would pay Elon 8 bucks to show up first?
> That Twitter hide (soft ban) tweets/accounts that an algorithm labels as misinformation because it was massively flagged? That happened before, it was easily abused, it was censorship.
You mean something like this? https://twitter.com/chrisswartout/status/1753634492894347759