←back to thread

190 points amichail | 1 comments | | HN request time: 0.201s | source
Show context
grishka ◴[] No.42195272[source]
> The use of algorithms to filter information has become the norm because chronologically presenting information from followers creates a confusing morass for the average user to process.

Can't disagree more. Call me old-fashioned but I hate any algorithms at all meddling with what I see. If I follow someone, I want to see their posts, all of them, without exceptions. If I don't follow someone, I only want to see their posts if they were knowingly reposted by someone who I do follow. If I want some posts filtered from my feed, I'll set up word filters myself, thank you very much.

It's a recurring theme in the modern IT industry that "the average user" can't be trusted to take their own responsibility. It's sometimes taken as an indisputable truth, even. Why does this keep happening? What can I do to put an end to this?

replies(15): >>42195452 #>>42195484 #>>42195500 #>>42195534 #>>42195550 #>>42195552 #>>42195629 #>>42195648 #>>42195715 #>>42195785 #>>42195811 #>>42195814 #>>42196019 #>>42198305 #>>42198449 #
1. 015a ◴[] No.42196019[source]
The reason social media apps use more complex global discovery algorithms (over a chronological feed) is because chronological feeds always run out of content. That's literally the only reason. At some point, some team at some gigacorporation invented the "hours spent with us" KPI, and tasked their hundred reports to increase it. It turns out, it doesn't matter how many people complain, if the "hours spent with us" KPI keeps going up.

"But users prefer algorithmic feeds": There's no evidence of this. The KPI is measuring an increase in hours spent with the app; it is not scientific-method A/B testing a preference between two options. Even if an app could do this, what does "preference" mean? You could measure how many users pick one experience versus another, but I've never found an app that, if it offers both experiences, durably and reliably saves your choice for a chronological feed between re-launches. Also: Maybe I want both experiences, at will. Hours spent in one experience versus the other? This is not communicating a preference; if I choose spending an hour driving during my commute to one job, versus ten minutes walking to another, have I revealed a preference for a longer driving commute? Obviously not.

You can ask users directly: And users may actually reveal their preference that social media never existed at all because your company isn't actually delivering value to the world [1]. Oops. Uh, don't run that survey again, bury it, make sure shareholders don't find out.

All social media is trash, and should not be consumed by anyone who has even an ounce of self-respect. Honestly: HackerNews is in that bucket, but at least its not as bad as most platforms.

[1] https://fortune.com/well/article/nearly-half-of-gen-zers-wis...