misskey
misskey copied to clipboard
misskey-dev
User-driven/configurable algorithmic discovery
Summary
For many reasons (I can elaborate if this is not an immediate "no"), I think that some form of algorithmic discovery is desirable, as it is simply impossible for a user to be online enough to sort through all possible content manually. However, algorithms built by large corporations which seek to maximize user engagement have very obvious downsides in many well-discussed ways. Still, I believe there is room for a good kind of algorithm which is opt-in and user-controlled. Since misskey has no monetary incentive to maximize user engagement, we're instead free to design an algorithm which centers the user, allowing said user to more effectively find the content they want to find.
It would be really great if this could be configured live in real-time in some fashion, i.e. making it into an intelligent search of some sort.
However, I know that in some places the mere idea of any algorithm being introduced is basically heresy. I searched here and the forum to see if I could find any discussion, but in english and my online-translator powered japanese searches I couldn't find anything about whether the creators of misskey have any stance on this.
This already is already in Misskey somewhat with the explore tab, where you can view popular posts and both local & remote users.
The ability to delegate recommendations to external services was implemented in v10.1.0 but has been removed as it is no longer maintained from anyone.
Oh, I didn't even know that was a thing. I might reimplement that in Calckey 👀
It was 4 years ago, so it is no wonder you do not know about it. And it seems that some of the logs are missing because the implementer deleted their account.
This already is already in Misskey somewhat with the explore tab, where you can view popular posts and both local & remote users.
Indeed! This is one of the reasons I thought this would be more well received by misskey than some other existing ActivityPub-based services.
Okay, sounds like there's at least room for exploration here.
No promises, but I'll try to draft up some sort of more detailed proposal and hopefully try to move forward with implementing something.