Also want an algo that doesn’t simply sum up all engagements. A post with 100 likes and 1 RT is usually a personal announcement, a joke, or a cute cat picture. But sometimes I’d rather see a post with 3 RTs and 0 likes (eg a good question someone raises).
this; but also automated with LLMs. there's so much potential (and also technical means) for highly user-aligned/controllable recommendation/classification systems that go well beyond "bad on purpose to make you click + reply" maximizers
as a personal example i could imagine having tabs with categories "meditation practice discussion poasts", "art posts for [fandom names here]", "AI doom-n-foom discussion", "posts from people i follow who rarely post"; highly personal classification rather than an undifferentiated Stream Of Content
could be a great way to reduce the “your feed is eternally horrible tragedy posts interspersed with shitposts and pictures of food” dilemma of modern social media
Hashtags filled this function largely - what their eventual usage became. You could follow hashtag, set up a separate tab. Is there a future for hashtags here? I’m neutral on that.
Comments
I've wished for community-curated metadata in other apps too (Reddit; Amazon).
Conscientious users are surprisingly powerful
Also want an algo that doesn’t simply sum up all engagements. A post with 100 likes and 1 RT is usually a personal announcement, a joke, or a cute cat picture. But sometimes I’d rather see a post with 3 RTs and 0 likes (eg a good question someone raises).
I love media tagging, so this would be really cool
But yes, the functionality I would like.
And lists or something similar
Or maybe smart hashtags