I should probably first say that this is unlikely to be implemented, there isn’t much support for it on the iNat team. There are bookmark services you can use across platforms that could fulfill the bookmarking function. I do like the idea of being able to search observations you’re following, however.
I wanted to share my perspective on Faving and iNat, since I’ve been looking through faved observations almost every day for nearly four years now (for Observation of the Day). Right now about 288k verifiable observations have been faved, which is about 1.2% of all observations. So, a tiny fraction of all observations have been faved, and I bet a decent number of these were faved by users who were “bookmarking” them for later, in addition to students who like faving their own observations.
In my experience, a lot of observations with “pretty” photos are faved, and that’s great, but I think even more observations are faved because they show a cool/interesting interaction, or a new species to iNat, and a good number of them are of what most people would consider “ugly” organisms. Oftentimes the photos aren’t anything special, technically, it’s what observed that was special. That’s just the kind of stuff our community likes, which is awesome. And that even extends to our social media audience - the more popular Observations of the Day tend to be the “weird” organisms that aren’t conventionally thought of as pretty or cute. Things like lions, bears, and other generically popular organisms aren’t the biggest hit with the iNat crowd.
Finally, iNat doesn’t have an algorithm that automatically displays faved observations, and I think that is what really fuels the addictive and competitive aspects of platforms like Instagram (which I have personally stopped using) and Twitter. So I don’t see a major risk by keeping the fave function.