> My other comments expand on the biggest issue which is the advertiser-optimized engagement algorithm. A site which does not cater to advertising (using an alternative source of income such as a user-driven subscription model) can instead maintain a time-based feed. This is a much better user experience. It doesn't favor negative engagement, it's less addicting, and it doesn't create artificial virality.
Thanks I probably should have read the other comments.
I haven't registered for a new Bluesky account in a while now, but when I joined there was no algorithmic feed. The "Discover" feed was added later. The default feed for me at least is still the timeline. I do occasionally glance at the "Discover" feed, but I ended up writing my own feed algorithms for my own uses that I use more often.
Though I will say, even on my purely time-based feed, ragebait always wins. People like ragebait. Today <Jane Johnson> is a nazi. Tomorrow <insert bad customer service story> so corporations are evil. While my examples are colored by the folks I follow, I'll bet this makes up the majority of content online.
I've been trying to work on a feed that removes this kind of ragebait but it's proving real hard especially when the post is sarcastic. I'd love to connect with folks who have worked on algorithmic feeds in the past and know how to crack this problem. Right now Muted Words tend to do the trick when there's a new Villain (not always a person, sometimes an entity) of the Day on my feed, but I'd love to do this automatically.
Sure but we don't need to help it along. People are gonna people, we shouldn't create machines that encourage the worst tendencies.
If Bluesky goes the advertising route (likely) they'll find just like everyone else that stoking fires is the best way to ensure engagement, which ensures eyeballs, which benefits advertisers.
My problem with ads is not the ads themselves, it's that the ads become the customer, and the users become the product being sold to the advertisers. This creates a series of perverse economic incentives that significantly shift the way the site is designed and developed.
Last I heard Bluesky was contemplating a paid app with enhanced features and also some rumors about hosted domains. They haven't ruled out ads. I'm not sure where that landed and how far along that process they are. The most I know about the situation is this interview [1] here on TechCrunch.
There are increasing numbers of new ways to generate feeds that suit your needs. I use the list functions to segregate people I follow into categories. There are a couple of built-in feeds that help discover new people.
But, the best thing is exemplified by graze.social which gives you a huge amount of ways to filter the firehose feed into whatever you want. Leave out the rage bait? It’s a filter sequence away.
Me? I do not want to avoid the villain a day problem. I want to see and block them myself, to know what’s going on.
Yeah I'm aware of feed generation and as I said I wrote my own feeds to do some of this filtering. I really should use Lists more often, you bring up a good point.
I'm a huge fan of how easy it is to have editorial control of your feed on Bluesky and the concept of a standard Bluesky identity.
Thanks I probably should have read the other comments.
I haven't registered for a new Bluesky account in a while now, but when I joined there was no algorithmic feed. The "Discover" feed was added later. The default feed for me at least is still the timeline. I do occasionally glance at the "Discover" feed, but I ended up writing my own feed algorithms for my own uses that I use more often.
Though I will say, even on my purely time-based feed, ragebait always wins. People like ragebait. Today <Jane Johnson> is a nazi. Tomorrow <insert bad customer service story> so corporations are evil. While my examples are colored by the folks I follow, I'll bet this makes up the majority of content online.
I've been trying to work on a feed that removes this kind of ragebait but it's proving real hard especially when the post is sarcastic. I'd love to connect with folks who have worked on algorithmic feeds in the past and know how to crack this problem. Right now Muted Words tend to do the trick when there's a new Villain (not always a person, sometimes an entity) of the Day on my feed, but I'd love to do this automatically.