Yep, these are all excellent points about the tricky/gnarly parts of Mastodon!
1) is a good point, though in practice the question of which server one starts on isn't in my experience super crucial; the vast, vast majority of people who I follow -- and who follow me -- aren't on my server
2) Very true -- Mastodon is designed, I suspect, to limit the emergence of viral posts and a map of zeitgeist. Viewed one way that is a bug that can also be a feature; Twitter's public realm has arguably been driven faintly bonkers by a little too much emphasis on rubbernecking the daily/hourly/minute-ly zeitgeist, lol. But I agree it'll limit the growth of Mastodon ... people who really want to be on top of the subject of the moment won't find Mastodon satisfying ... it's more for conversation than cultural monitoring of trends
3) A good question! I don't know how the API works for one server talking to another. But my sense is there's no way for server X to seriously affect how server Y presents posts. Server X can determine how it itself broadcasts its posts to other servers. But it can't affect how any other server displays those posts ...
4) Yep, I follow a lot of folks! But it's not terribly unruly as yet, partly because Mastodon seems governed by the same 80/20 rule one sees in social interactions online: Only a minority of folks post much. (Itself a feature/bug aspect of how people behave online.) I can also set up lists, though, to make smaller groupings of folks to look at ...
4) Yep, this is a good point! Twitter already had quite a bit of that echo chamber effect happening, given the homophilic way that users picked who to follow. What might create more of an echo chamber on Mastodon is the lack of a quote-tweet: A common usage of quote tweeting on Twitter was for person in Group A to find a tweet from Group B that they view as horrible, then quote-tweet it to their Group A followers with a comment saying essentially "look at how ghastly folks in Group B are". It was, at least, a way for Group B and Group A to have exposure to one another -- and Mastodon will have less of that!
That said, the contention of Mastodon's designers (and early community, which settled upon the elimination of quote-tweeting) was that the downsides of this sort of Group A / Group B rage-quote-tweeted were more significant than the upsides. I don't have any particular data or wisdom to know if they're right ... though in general I respect the experiment.
It's also worth noting that research has fairly consistently found that homophilic echo-chamber self-sorting is more pronounced in F2F life than in many places online ... so even if Mastodon develops echo chambers worse than Twitter, it is likely to still be better than the physical/neighborhood silos we've sorted ourselves into IRL. Weak consolation, though?
