This is a particularly silly opinion because Lemmy is an algorithmic social media platform. It’s just an algorithm that you happen to have access to documentation for. Almost certainly, any fediverse algorithm would have to work on the same principles as Lemmy (open and based on public interactions). Likes and upvotes are king. User similarity ranking is wildly inefficient on the fediverse due to its distributed nature and keyword systems are easily gamed (although some hybrid is possible).
Sure, Lemmy has algorithms, but it’s not the same as being algorithmic. We can continue to argue on the semantics of technical definition, but my point on the issue with toxic mess was clear a few messages above. Intent of the implementation really is the factor that defines the different between platform being algorithmic and using algorithms.
If a user has a chronological feed, and it reaches the top (or bottom) of the content, the user is likely to close the app/page and do something else, but an endless feed will keep the same user longer.
And using user input to sort content is not the same as prioritizing content based on past interactions to achieve a predefined goal.
E.g. Twitter API allowed to pull content in chronological order of only people you followed, but the official app used For You feed as default and even if you changed it manually to chronological in the settings it would reset on app restart. Contrast that with being able to set any sorting method as default on Lemmy or having no algorithm on Mastodon (yet the API allows anyone to create one if they wish) and the difference of intent is clear.
This is a particularly silly opinion because Lemmy is an algorithmic social media platform. It’s just an algorithm that you happen to have access to documentation for. Almost certainly, any fediverse algorithm would have to work on the same principles as Lemmy (open and based on public interactions). Likes and upvotes are king. User similarity ranking is wildly inefficient on the fediverse due to its distributed nature and keyword systems are easily gamed (although some hybrid is possible).
Sure, Lemmy has algorithms, but it’s not the same as being algorithmic. We can continue to argue on the semantics of technical definition, but my point on the issue with toxic mess was clear a few messages above. Intent of the implementation really is the factor that defines the different between platform being algorithmic and using algorithms.
If a user has a chronological feed, and it reaches the top (or bottom) of the content, the user is likely to close the app/page and do something else, but an endless feed will keep the same user longer.
And using user input to sort content is not the same as prioritizing content based on past interactions to achieve a predefined goal.
E.g. Twitter API allowed to pull content in chronological order of only people you followed, but the official app used
For Youfeed as default and even if you changed it manually to chronological in the settings it would reset on app restart. Contrast that with being able to set any sorting method as default on Lemmy or having no algorithm on Mastodon (yet the API allows anyone to create one if they wish) and the difference of intent is clear.