Why would you do that? Why would you have a follow function and break it? And what else do you show people? I mean if your user follows someone, then it is what they want to see, why wouldn't you show them what they want to see?
Why would you do that? Why would you have a follow function and break it? And what else do you show people? I mean if your user follows someone, then it is what they want to see, why wouldn't you show them what they want to see?
Ideally the user is happy to discover new content that both complements their existing interests and occasionally introduces them to something new and exciting, while the company is happy to have an algorithmic lever that helps with ad placement. This theoretical win-win ideal may not get realized very often in practice...
"Ok, this user follows N. But N's latest post is already 15 hours old and it had poor engagement. Probably that means it's low quality and it's not worth including at all."
It makes sense, if you think you know better, what the user wants to see. Or, if you somehow make more money by doing so.
In either case, personally I like to decide, what I see myself, but I might actually be a minority (soon). Many people are apparently fine with intransparent algorithms making the decisions for them.
This is so insane to me. Some weeks ago I had to create a twitter account because I wanted to contact a developer for an app I'm using and the only contact information he had was his twitter account. I created a dumb account with my Google account and I was surprised that the first tweet I see is from Elon Musk. How small your ego has to be to request being featured first for every new account that is created? I remember I saw his face in the account creation process and I thought to myself "I would never follow this idiot" and still this guy is the first thing I see.
I wont be opening that cesspool anytime soon.
Sadly providing an algorithm that provides this is not what makes money so with most solutions you're stuck with an algorithm that just tries to show you what will keep you from clicking away instead.
The stupid thing about this is that Twitter has a great measurement system for evaluating the tweets. They have likes, retweets and quote tweets. It would not be that hard to make some inferences based on those metrics mixed with some randomness that would create a nicely curated timeline for most. You could even allow users to up the randomness or give some more explicit preferences.
I would say twitter stays about 1.5 months ahead of every other site I know of on AI news in particular. I've usually had time to demo new tools for quite a while before hearing about them on HN or elsewhere.
The first appears to be randomised or from The Algorithm, but as I far as I can tell, it’s mostly random.
The second also seems randomised but from people you follow.
That’s not true.
The following tab is only showing you who do you follow with posts in chronological order