I agree. I think the problem of choosing an instance in federated social media is quite difficult. From the outside, it’s hard to tell if an instance is generally good or not (e.g. whether they’re anti-bigotry, well-moderated, have a good atmosphere, etc.) because you don’t have much information except for what the instance owners tell you. And even once you create an account on an instance it could still be hard to judge it well, because maybe the instance admins are assholes but you can’t tell because they ban and defederate from anyone who points out things they do wrong and because anyone who remains on the instance either is ignorant of or tolerates the admins’ bad behavior. (I don’t really know if this happens on Lemmy specifically, but there is a Mastodon instance that is listed on joinmastodon.org that I used to think was chill, but I later found out that one of the admins has a habit of harassing people. How are you supposed to find this stuff out without ruining your time and mental health constantly following the latest federation disputes?) (And choosing a server from a list also has the disadvantage of centralizing the Fediverse further.)

So the most trustworthy way to pick a server I can think of is this: Have a friend or someone you already trust who uses the Fediverse, and use their judgment to help you pick a server. Now, this has the pretty big limitation of requiring you to already know someone on the Fediverse, which I imagine most new people on the Fediverse (including myself when I joined the Fediverse) don’t. So a new joiner to the Fediverse has to do the hard work of weeding out the really bad instances and the instances that tolerate the really bad instances, and then they have to hope they didn’t land on one of the subtly bad instances. That seems unpleasant and I imagine most people don’t have the time and willingness and just give up.

A possible way to alleviate the problem might be to maintain carefully curated instance lists along with honest notes about the general experience of being in each instance. But that would take a lot of work to create and maintain, and it runs the problem of also having to ensure the trustworthiness of the person who makes such a list. For that reason, I don’t think this solution that I’m proposing is likely to happen.

I do feel that flatness and context collapse are major problems for Lemmy and the threadiverse. In a chat group or a smaller forum, you can have a general idea of who will see and respond to what you post, and because everyone knows everyone else (or knows someone who knows everyone else) there is a social context that can be built on, with the community reaching consensus on certain social norms. But on Lemmy, even in small, niche communities, anyone can just drive by and post a comment after seeing something in their “All” feed, and unless it’s bad enough to warrant moderator action then there may not be many social consequences outside of that one comment feed. In this sense, I feel that the design of Lemmy encourages certain patterns of behavior that feed into some of the problems of Lemmy’s social environment. For example, the existence and prominence of the “All” feed makes it easy for people to express disdain at others’ expressed preferences when it’s not in their place to do so, because the “All” feed treats all posts essentially the same and lessens distinctions between the social expectations of different communities, for example if some communities expect members to have put in some effort to understand the topic of the community.