News for hackers. Stuff that matters.
Re:The Twitter Algorithm by Joeri on Saturday April 01, 2023 @02:25AM (#35397700)
I did find the article more enlightening than the source code. I always suspected the quality of a tweet did not factor in to its promotion, only its engagement, and now they have confirmed this to be the case. Now I understand better why twitter seems to be filled with people angrily retweeting what I consider to be low quality clickbait tweets.

As long as they don’t try to tackle tweet quality at all separately from engagement twitter will remain unappealing to me.

Re:The Twitter Algorithm by jrochkind1 on Saturday April 01, 2023 @08:57AM (#35399965)
What does "quality of a tweet" mean, how might you measure it?

We pretty much knew this is how all social media works, becuase a) engagement is what they want, why wouldn't they be optimizing for it, and b) how else might you measure 'quality'? Back when this started, I have no trouble believing some well-intentioned engineers thought that engagement was a good proxy for quality. A bunch of users give it the "like", isn't that a collective assessment of quality? Who is to say what quality is, overruling the users in aggregate?

I agree it has the negative effects you mention; and I've read lots of people writing about this, it's of course not a new observation.

But I agree it's good to have an explanation of what's going on, even when parts of are what we basically knew was happening on all social media networks. confirmed is better than "basically knew", for understanding how these things that effect our experiences (and our society) work.

Re:The Twitter Algorithm by dbbk on Saturday April 01, 2023 @09:26AM (#35400148)
"Quality of a tweet" is impossible for a computer to understand. That's like asking to quantify the value of a work of art just from the image.