One thing I am always surprised at when people discuss Twitter/Mastodon clients is that there are apparently still so many people who attempt to consume every post in their timeline. I pretty much gave up any pretense of treating my timeline like an inbox years ago.
(Not an argument for algorithmic timelines at all BTW. I just don’t care that much about rock solid timeline position restoration because I don’t even pretend to follow everything that happens on here.)
@buzz I learned to --love-- tolerate the algorithmic timeline at some point. My hope for mastodon is that I’ll be able to tweak the variables to some degree and end up with something that works a bit better. I don’t need much more optimization than the 2000s era Bayesian spam filters tbh.
@jimray My main problem with the Twitter algorithmic timeline is it felt like the focus on engagement as a signal meant that your feed was always dominated by the same, like, 10 journalists who post 500 times a day like deranged addicts, while meanwhile there is a vast long tail of normies whose occasional posts might as well not even exist.
@jimray It also basically meant that anyone who wasn’t habitually posting epic multi-message threads all the time basically disappeared.
@buzz @jimray and now the epic thread posters have come here and think they've got to keep that up to be relevant
@powerllama @jimray Over on Mastodon, it is *not* time for some game theory!
@jimray Perhaps a more thoughtful algorithm might habitually say “Hey, here’s someone you haven’t seen in awhile! Guess what—they still exist!”

@buzz yeah, some combination of “I always want to see posts from my wife and my good pal Buzz (whom I miss and hope I get to see again soon)” and “this is probably actually interesting to you in a way that isn’t creepy”

YouTube actually does a good job of this, I’m hearing TikTok is similar. I suspect it’s somewhat easier since those are both built around surfacing interesting content, not building on top of social graphs.

@jimray @buzz that was the exact case that first made me hate Twitter’s algo view: missing a tweet from my wife. (The “someone faved this” stuff that made faves into soft RTs was the other big source of ire.)
@marcprecipice @jimray Yeah, that stuff was a huge mistake. Made Twitter feel like a panopticon.
@jimray @buzz did you know that you can enable notifications per followee so you won’t miss those posts?
@buzz Weirdly, Facebook has this feature. Last time I checked it out, it wasn’t actually that compelling to me.
@jimray “Blessed are the meek, for they shall inherit the feed.”
@buzz @jimray a non-algorithmic feed + a very good algorithmic “highlights since you were last here” would be ideal for me, and I love the idea of infrequent posters you follow getting highlighted
@mcbramhill @jimray Haha, yeah, it's like...hey maybe we could try using algorithms to create something other than perpetually self-reinforcing power laws?
@buzz It's less about being a completionist and more about the annoyance of being in the middle of something and having it all fwoosh away, for me. If I'm going to put up with computers, I expect them to be able to perform tasks like keeping track of a position.
@masto True, that is annoying. That brief moment of seeing something, going “Oh, what’s this?” and then BOOM, bye!

@buzz I have honed my skim-reading skills over many years of information overload, and they are pretty good. (Proof is that I spotted this toot and I’m replying to it.)

You don’t care about TL position restoration or don’t understand the need — fine. Please let those who care decide whether such a feature is useful or not.

Simply taking the user back to “right now” is an extremely crude way of dealing with information overload. Surely developers can come up with better ways.

@buzz

Right.
It's too vast.
I read about the first 10 posts, check my responses/reblogs, then leave.

@buzz In my case, I’m a completer. If I don’t check avery single post in my timeline I feel I’m doing it wrong.
@buzz you follow 400+ people. You could probably benefit from lists.
@sanguish Probably. I built a bunch on Twitter years ago but just never really got into the habit of looking at them.