Lots of ink has been spilled saying that the internet (and in particular, twitter) was better back in the good old days of reverse chronological feeds. When Twitter eventually capitulated to this common user request, there was less ink spilled on the results - it seems like the people who really wanted this feature used it happily, most folks never switched to it, and some switched back after realizing it wasn't the panacea they thought it might be.
While reverse-chronological gets a ton of love, it's worth thinking about this as a single instance of a larger question: how could users tune the parameters of the algorithms that decide what they see?
It seems like such a rudimentary idea, but below the surface there are tons of different mechanisms and formulations of this, many of which have been implemented under disparate names across different platforms. The reality of algorithmic recommendation is fractally complex, and not something I'm particularly well equipped to write about. Exposing a user to every lever that every engineer who has ever worked on a recommendatino system would be needless. However, a system that allows users more fine-grained control over the content they see seems worthwhile, allowing people to express with their feet their preferences for the inevitable tradeoffs that these algorithms embed within them.
I expect we will see this emerge in the fetaverse organically, but one could imagine a highly motivated platform using this kind of customization as their differentiator, or a regulation passing that compels this kind of customization.