Problem here is also that your instance may not know about all communities from the instances you’re connected to. This could probably also be improved.
Yes, that’s what I mean by not trivial, a centralized system can do analysis like this a lot easier. But even on your own instance, they could find the N users with the most overlapping subscriptions and check which communities they follow to give you recommendations.
I guess the ‘simple’ way of doing this would be adding tags to communities like ‘art’ ‘hobbies’ ‘sport’ ‘football’ etc. This might then let the app suggest others based on the tags you are subscribed to.
It would probably still require some AI/analytics to work out the links based on user activity in different communities/tags but I think it would make it easier to group interests and promote smaller communities.
It could also improve Lemmy visibility in Masterdon if the tags are used as hashtags or something. (Would require more work)
Kbin lists “related magazines” which are similar communities in the fediverse. Not sure how it works but I think it may be based on hashtags like this.
Recommendation algorithms are a big reason for the enshitification of other social media. You don’t need to be connected to everything everywhere all at once. Enjoy your handful of small communities.
I don’t want random posts to appear in my feed from communities I haven’t subscribed to, but I want to have a feature that shows me suggestions for communities when I ask for it. That’s a big difference. Right now it’s (too) hard to find these communities.
Hashtags could possibly help with this. When making a post, a user can add hashtags which categorize the content. One can then search hashtags, or subscribe to them to find new communities. Probably not as passive as you’d be looking for, though.
I would like to see recommendations for communities based on my communities. It’s not trivial to solve, but discoverability isn’t great right now.
Problem here is also that your instance may not know about all communities from the instances you’re connected to. This could probably also be improved.
Yes, that’s what I mean by not trivial, a centralized system can do analysis like this a lot easier. But even on your own instance, they could find the N users with the most overlapping subscriptions and check which communities they follow to give you recommendations.
I guess the ‘simple’ way of doing this would be adding tags to communities like ‘art’ ‘hobbies’ ‘sport’ ‘football’ etc. This might then let the app suggest others based on the tags you are subscribed to.
It would probably still require some AI/analytics to work out the links based on user activity in different communities/tags but I think it would make it easier to group interests and promote smaller communities.
It could also improve Lemmy visibility in Masterdon if the tags are used as hashtags or something. (Would require more work)
Kbin lists “related magazines” which are similar communities in the fediverse. Not sure how it works but I think it may be based on hashtags like this.
Exactly this. I like how substack works and have no idea if it’s doable or not.
Recommendation algorithms are a big reason for the enshitification of other social media. You don’t need to be connected to everything everywhere all at once. Enjoy your handful of small communities.
I don’t want random posts to appear in my feed from communities I haven’t subscribed to, but I want to have a feature that shows me suggestions for communities when I ask for it. That’s a big difference. Right now it’s (too) hard to find these communities.
Recommendation algorithms are fine as long as they’ce user-centric and opt-in.
Hashtags could possibly help with this. When making a post, a user can add hashtags which categorize the content. One can then search hashtags, or subscribe to them to find new communities. Probably not as passive as you’d be looking for, though.