I'm honestly wondering if this was a good idea, at least at this time. I'm sure you had good reasons to do it, and I certainly don't understand nearly half as much about the technical aspects as you do, but being able to find active rooms on a map view was really neat. As was having a choice of a variety of different clients that could all talk to each other. Spam was a problem, yes, but Nym showed that it was fairly easy to get rid of by simply using a relay that has rate limiting (such as Damus's). I understand that it would of course cause even more centralization if everyone was going to use that, but unless traffic was already causing severe resource strains, wouldn't it have been better to wait until there is a more significant userbase? A lot of people I talked to had trouble even understanding how it worked at this stage, how are they going to understand that there may essentially be several different versions of the same room on different relays, and depending on which client you use, you might see entirely different conversations? If it is not possible or desirable to expect other relays to rate limit, couldn't that be also done on the client side instead? For instance, any single account sending more than one message per second is likely spamming, as is anyone sending the same message from different accounts (such as the semicolon troll). Other alternative clients have tried different approaches — For instance, Nym has a customizable keyword filter, and bitmap.lat has a rudimentary pattern-based filter that worked pretty well that could potentially be ported to the app (you can see what it tests for in `src/lib/spam.test.ts` on their NostrHub). Basically I'm left wondering whether this move at this time killed all the momentum Bitchat had until now. My experience today was that most rooms on the app are essentially dead, and the ones that are working are those still using the global relays.