And the bullshit about “can’t verify your domain” upon my second attempt to subscribe to a newsletter is just nonsense. I’m subscribing to the person I’m subscribing to, not to you. I will not be cajoled and threatened into installing your shit app that is now more like X than something for long-form writing (which used to be what you claimed you specialized in).
I desire with all my heart that one day, there will be a way more elegant way to follow people. I am not against algorithms or any tech for moral reasons, but so far, I just have never experienced them working in a way I want them to work–even as a consumer. If I were satisfied with algorithmically served recommendations, I would be much much happier. But the reality (at least for now) is that these algorithms are based on mere proxy signals: likes, shares, subscribes, etc. They aren’t based on the actual thing. What the fuck do I care if 500,000 people pressed Like on something? Likes are cheap. When am I going to be able to get recommendations based on the actual thing?
Anyway, for now, so how do I find what I want to consume? Word of mouth + coincidence.
Ironically, word of mouth includes Facebook groups. Facebook itself is heavily algorithm-reliant, but the groups within Facebook? Especially private groups? Unless the majority of the members in a group are bots, it’s simply way likelier that I will enjoy what is being recommended there more than some random algorithmic recommendation.
Yes, random! The present-day algorithmic recommendations are so random, they’re irrelevant. The algorithm might think that it’s the most statistically sane thing ever, but no, that is literally machine thinking. (Which is fine if machines are recommending things to each other, but not fine when their makers pretend that they’re ideal for human consumption.)
When I just ate a slice of pizza, I don’t fucking want another slice of pizza. I probably don’t even want coke-related recommendations, because I am not stupid enough to need such a recommendation. For God’s sake, if I am eating pizza and I crave coke, I will think of that myself, thank you. Anyway, probably, after eating pizza, I want something less cheesy and fatty. I don’t know, figs. Or cucumber. Who knows. And after I had that, I don’t want any more food recommendations!
There has to be actual random stuff in the recommendations, otherwise the supposedly statistically accurate recommendations will start feeling random. It’s like throwing pizza after pizza on a person who cannot find use in pizza anymore. That is so random.
It’s fascinating that apparently, tendencies like these cannot be dealt with algorithmically. Why is that, when these are so predictable? Is there no way to instill randomness, or did the makers of these algorithms in fact find that most people do want pizza after pizza after pizza after pizza?