Live data from Hacker News

Twitter's Recommendation Algorithm

blog.twitter.com

321–330 of 1001 posts

Re: Twitter's Recommendation Algorithm

#321
post #247
post #182

Earlier quoted context omitted.

Spoiler - there isn’t.

Yeah they track author_is_elon, author_is_democrat, and author_is_republican but they don't appear to be used for favoritism anywhere in this code.

Why do they exist then? No code references it, but that's Scala/JVM so many things depend on runtime initialization, so maybe some other systems do? wich ones?

Is is it there to help fight impersonations? should be solved with Twitter Blue already?

There was reports of people receiving notifications about Musk tweets despite not following him, so..

Re: Twitter's Recommendation Algorithm

#322
post #227
post #165

Earlier quoted context omitted.

Mastodon is conveniently also AGPL...

On reason I use Mastodon is that there is just chronological timeline. Quick scroll and you are done. Bad for advertising platform - good for user.

I'm not sure why Mastodon would be interested in Twitters non-chronological timeline. It seems to be pretty antithetical to its goals.

Re: Twitter's Recommendation Algorithm

#323
post #236

I wonder what the "author_is_elon", "author_is_power_user", "author_is_democrat", and "author_is_republican" labels are for [1]. [1]: https://github.com/twitter/the-algorithm/blob/main/home-mixe...

Elon is addressing this in the Twitter Space right now. "It definitely shouldn't be dividing people into Republican and Democrats; that makes no sense[...] you've identified something we should be getting rid of right away."

but how can we be sure it isn't doing that?! first, we would need to figure out a way to identify who-

Re: Twitter's Recommendation Algorithm

#324

Earlier quoted context omitted.

Most of the major news outlets are not doing so either. The Elon stans are crowing that this will be the long-overdue end of legacy media, but it strikes me that the new 'blue check twitter' might end up becoming even more of a social bubble than what it replaced. There are so many low quality accounts sporting a checkmark now that users who value substance will soon be incentivized to just block anyone they find ann…

yeah, they, uh seem to have realized that's gonna be a problem. https://variety.com/2023/digital/news/twitter-business-month... and of course all this means is that the organizations most likely to be able to afford it won't have to

Good find! It'll be funny to see if the incumbents respond with 'don't do me any favors.' Also to see whether Musk's frens sulk aboit him selling out to the elites or so - their gratitude has an extremely short half-life.

Re: Twitter's Recommendation Algorithm

#325
post #236

I wonder what the "author_is_elon", "author_is_power_user", "author_is_democrat", and "author_is_republican" labels are for [1]. [1]: https://github.com/twitter/the-algorithm/blob/main/home-mixe...

Elon is addressing this in the Twitter Space right now. "It definitely shouldn't be dividing people into Republican and Democrats; that makes no sense[...] you've identified something we should be getting rid of right away."

As if Elon has a clue what that feature is or is not being used for.

Re: Twitter's Recommendation Algorithm

#326
post #302

Earlier quoted context omitted.

About 40% of US voters are not registered with any political party, so at least they will avoid whatever this triggers.

So you imagine these tags are set by looking account names up on state voter registration lists?

Well, if not, then the tags have such a large error bar as to be meaningless.

Re: Twitter's Recommendation Algorithm

#327
post #105

The twitter algorithm sucks balls and heavily overweights who's paid for a checkmark. The default feed view has grown increasingly useless over the past ~6 months.

I don't think any changes to bias towards bluechecks have been made yet.

Re: Twitter's Recommendation Algorithm

#328

Earlier quoted context omitted.

I'd love to see the exact date author_is_elon was added. Too bad they didn't publish the commit history

IIRC it was very recent, there was a Twitter engineer that was fired after explaining to Elon that the algorithm was not biased against him: https://www.salon.com/2023/02/10/petulant-elon-musk-fired-tw... Almost certainly after that event Elon had them explicitly bump his tweets in their reach.

This speculation seems mostly informed by a negative opinion of Elon. Is there any real indication he gave this instruction?

Re: Twitter's Recommendation Algorithm

#329

Earlier quoted context omitted.

Here is a screenshot in case this changes later: https://i.imgur.com/F8GSeyH.png And, no, this wasn't in a merge-request, it was in the "main" branch of HomeTweetTypePredicates.scala.

What's all the "DDG"? Is this data from DuckDuckGo?

I doubt it, since the isElon thing also has the abbreviation

Re: Twitter's Recommendation Algorithm

#330

Earlier quoted context omitted.

Well yes, but I am pretty sure, elon envisoned warm welcome by the OS community and help for free and now that is off to a bad start.

I doubt he thought he'd get any help for free, given I don't think you can run this code there isn't a meaningful way to contribute. It's just about transparency or PR, take your pick.

"given I don't think you can run this code there isn't a meaningful way to contribute."

Hm, there is lots of code released, I would think that some of it, can be run and might be forked and useful in other context, but mainly it is a PR move, sure.

Post reply on HN