In middle/high school I put all the music I had on an off-brand mp3 player and would just set it to Shuffle All. I quickly realized the player’s shuffle fuction wasn’t purely random–it was weighted towards my favorite songs (aka the songs with the most plays).
Only I had never chosen those songs. They were just the random few to pop up the first time I shuffled everything, and they started playing more and more frequently as this horribly short-sighted algorithm fed itself bad data, until I was so annoyed at those few songs that I stopped listening entirely.
Anyway a few years later Facebook did the exact same thing with my friends list, siphoning me off from seeing most of my feed because OBVIOUSLY I interacted with them the most, therefore they must be my besties. But really they were just the only people showing up for me to interact with in the first place, until I was down to just a few people I never really talked to from high school, a college prof, and my racist uncle I kept calling out.
And shortly after that, YouTube followed suit, replacing “Subscriptions” with “Recommended” as the default category, and trying to find “things I liked” when it was really just whatever three channels I’d watched last, whatever unrelated viral vid it wanted to push that week, and weird perennials like Whose Line clips or lockpick reviews or YTPs that seem to hibernate for months at a time then return like locusts.
All this to say: the big mysterious algorithms that now run all the major platforms on the internet are never acting in your best interests. They’re just that junky mp3 player’s Shuffle All with a fresh coat of paint, and, to be clear, this is by design. They are VERY good at what they do, which is funneling users into nice predictable pockets of content that advertisers can exploit.
