3 ms·

"Followers" as a metric exists because Twitter demands a centralized feed that's constantly unstable so you go back to it often. Why can't it be like RSS, where

"Followers" as a metric exists because Twitter demands a centralized feed that's constantly unstable so you go back to it often. Why can't it be like RSS, where everybody has a feed, and it tracks what posts I've read and which ones I haven't, and I can catch up with all the unread posts? Probably because that way, I could check it once a week, instead of every 30 minutes, and Twitter needs those DAUs to survive. Web3 isn't worth the digital paper I'm writing this screed on, but if there's anything that a decentralized social network should be able to do, it's get us out of hyper-aggressive VC-subsidized optimization for eyeballs and growth, and put us back into humane territory. The mis-step of adding back "status" into the network is just showing us nobody even recognizes the problems here, or more likely, intentionally want to make a worse version of what already exists. also, tbh, the "follow" system is a confusing web of permissions. Primarily it means "I want to see their posts in my feed", but also sometimes "I want the ability to see someone's private feed" and "I want to be able to DM someone", because inexplicably, there's only one relationship available in Twitter... the follower.

comex 4y ago
> Why can't it be like RSS, where everybody has a feed, and it tracks what posts I've read and which ones I haven't, and I can catch up with all the unread posts? Probably because that way, I could check it once a week, instead of every 30 minutes, and Twitter needs those DAUs to survive. I use Tweetbot, which approximates this by saving my place in the timeline (though it doesn't let me mark tweets as read/unread/needs-followup on a more granular level, which would be a useful feature). But ironically, I suspect I might be better off using the standard unstable feed, precisely because I only check Twitter anywhere from once a week to once a day. Every day, my timeline gets hundreds of tweets. The actual volume of text is not that high, compared to the long blog posts I'm accustomed to reading at a fast pace. But the tweets are about many different topics, and many of them are tips of icebergs, having links or being part of longer threads. So the timeline feels overwhelming, and I almost never read it all. Given that, perhaps I'd be better off without the pretense. Ditch the saved place; prioritize recent tweets, so that they're still live discussions in case I feel like replying; but rather than solely prioritizing recency, use a filtering algorithm to identify which tweets are most worth reading. But for now I'm stubbornly sticking to Tweetbot – partly out of habit, partly because I don't want to use an algorithm unless it's something open source that I can customize.

comex 4y ago

> Why can't it be like RSS, where everybody has a feed, and it tracks what posts I've read and which ones I haven't, and I can catch up with all the unread posts? Probably because that way, I could check it once a week, instead of every 30 minutes, and Twitter needs those DAUs to survive. I use Tweetbot, which approximates this by saving my place in the timeline (though it doesn't let me mark tweets as read/unread/needs-followup on a more granular level, which would be a useful feature). But ironically, I suspect I might be better off using the standard unstable feed, precisely because I only check Twitter anywhere from once a week to once a day. Every day, my timeline gets hundreds of tweets. The actual volume of text is not that high, compared to the long blog posts I'm accustomed to reading at a fast pace. But the tweets are about many different topics, and many of them are tips of icebergs, having links or being part of longer threads. So the timeline feels overwhelming, and I almost never read it all. Given that, perhaps I'd be better off without the pretense. Ditch the saved place; prioritize recent tweets, so that they're still live discussions in case I feel like replying; but rather than solely prioritizing recency, use a filtering algorithm to identify which tweets are most worth reading. But for now I'm stubbornly sticking to Tweetbot – partly out of habit, partly because I don't want to use an algorithm unless it's something open source that I can customize.