sj_zero on Nostr: tl;dr: You're mostly right with some clarification about the potential impacts Here ...

tl;dr: You're mostly right with some clarification about the potential impacts


Here it's not just a matter of us seeing what we want and not what we don't want, but other people choose to see what they want and not what they don't want as well.

I think it's pretty important in both ways. It means that if you engage it's more satisfying, but less addictive.

Without an algorithm specifically going "We think you'll like this and we think you won't like this", you're going to see everything from the people you follow, which can have 4 effects: First, you can end up seeing posts nobody could have predicted you'd like that you like. Second, it means you see a lot of noise from people that you neither like nor dislike, which I think may actually have a slightly positive effect because it lets the dopamine drop off and sometimes you go "ah, there's nothing interesting on the TL right now, I'll go do something else". Third, it means that sometimes you see posts you absolutely categorically hate and they'll get displayed from the people posting it whether you like it or not which means you might unfollow certain people who occasionally say things that are good but occasionally say things you absolutely cannot allow.

Then there's the effect on your own posts. You get interaction not because an algorithm likes what you wrote and sends it to a bunch of randos, but because people who follow you liked your post and chose to interact with them among the different posts they see (or alternatively because they see your post on the instance timeline which is made up of local users and remote users that local users are following). Then people following you might share your post and through networking someone might see it and decide to either interact with the post or to follow you.

For mastodon (which isn't what everyone uses), I'm aware you can follow hashtags too, meaning that you get the portion of your instance timeline that tags a certain thing.

For completeness you also would need to recognize that each instance has its own defederation policy so there's potentially a block of the fediverse you'll never see at all.

Overall then, it's somewhat more granular what you see and don't see, but it's also broader that you'll see stuff that isn't so stimulating to you, and you'll see the stuff you like from people you like but also stuff you're ambivalent to and stuff you hate from people you like (and the people they follow and repost)

Contrasting the flat chronological view of mastodon and its alternatives, is the algorithm of big tech sites.

The algorithm often seeks out stuff it thinks you'll love and is often justified as cutting through the chaff to get to all the stuff you like. Even people on the fediverse often express disappointment that there's posts they might like but can't immediately see because there's a lot of posts they don't particularly care about, and express that an algorithm could help with that.

The algorithm doesn't just show you stuff you love, however. It also actively seeks out stuff that is tuned so you hate it just enough to leave a dislike or a comment saying "I hate this you're an asshole" it but it isn't so bad you close the app. Whereas organic growth certainly can surround you with people who believe exactly what you believe, I feel like being specifically shown stuff you hate on this level is almost worse, because whereas if you're surrounded by people you agree with then the fact people who disagree with you doesn't affect your life, an algorithm bringing people who hate each other together on one platform means you can't just quietly coexist because the platform itself will be constantly shoving you in their face.