51–60 of 138 posts
Recommender systems like Twitter are as much about data as they are about code. Without the dataset and the derived statistics and models that are used for ranking and recall, the code is not going to help much with transparency.
Heck, even create an alogrithm store where people could create and purchase different ways to view their feed.
It doesn't seem that difficult to add these multiple sort and filter options, but maybe it's more complex than I imagine.
Sounds like a bad idea. Bad actors would immediately exploit it and figure out how to most effectively flood everyone with spam, same as if Google's exact algorithms were public. Not everything has to be open.
On occasion, people make the same argument about all open source software, that making it public makes it insecure. Fortunately this has not turned out to be the case in reality. If there are flaws in the system they will probably be easier to find and fix with more eyes on it.
In content moderation vs spammers, you have spammers adapting to your tools in real time while posting content in the same way real users do, and the balance between "don't want to accidentally ban real users (who post stuff very similar to the bots)" vs "does it matter if someone gets past your defences a few times an hour" is massively different too.
Sounds like a bad idea. Bad actors would immediately exploit it and figure out how to most effectively flood everyone with spam, same as if Google's exact algorithms were public. Not everything has to be open.
Earlier quoted context omitted.
The discussion is about whether people can see what the algorithms are doing. Read only access to the code so they cannot manipulate people into their belief system or arbitrary shadow-ban rules. You’re right about human judgement but that’s not the topic. The central point, I repeat emphatically, is about transparency, not governance. Twitter can continue exactly the same way but just be transparent. The intense pus…
> Twitter can continue exactly the same way but just be transparent. No it can't. If the algorithm was transparent the only thing you'd see is spammers who have put tons of resources into figuring out the exactly optimal way to maximize engagement. Grassroots engagement would be impossible.
Also, Twitter's spam control has been objectively bad.
https://twitter.com/paulg/status/1487022342630957062?lang=en
People think that the entire platform has been hijacked by left-wing / progressives and the reason for lack of transparency is more insiduos than "spam". For example, being liable for what they told Congress.
The best feed is the chronological, non-algorithmic feed.
An unmoderated chronological feed would simply be a running list of every tweet from every one . A screen full of this feed would just be tweets made at the moment the data for the screen refresh was fetched. Is that what you or anyone wants?
It's not a global feed of every twitter in existence, it's a feed of accounts that you proactively choose to follow.
if(tweet.author.affiliation == "democrat"){
promoteTweet(tweet);
} else {
shadowBan(tweet.author);
}Make moderation guidelines and moderation audit trails public and open.
Not sure there is anything to gain here. Moderation is inherently subjective, especially at the scale of Twitter. The same people would just continue arguing about what should and should not be moderated and how.
I think it would be better if it was an informed argument.
The best feed is the chronological, non-algorithmic feed.
An unmoderated chronological feed would simply be a running list of every tweet from every one . A screen full of this feed would just be tweets made at the moment the data for the screen refresh was fetched. Is that what you or anyone wants?
"Open up the algorithm" is such a charmingly naive take on community moderation at scale that I can almost forgive the author for literally parroting Musk as the source of a lightweight take synthesized mostly from other people's thoughts in order to attract attention to his "Every" newsletter platform.