Emacs shows how to craft a very diverse and flexible system still without UX issues :D

Beside that Nostr is "an everything network" so it's normal that clients should be " everything apps", and any modern everything app beside Emacs and Smalltalk from the ancient PARC, is actually limited by the underlying UI/UX limited and limiting model.

Nostr could be:

 - a blog platform, so those who want to blog well... Need a web skeleton like WriteFreely or Hugo to build their own personal Nostrsite

 - a generic discussion platform, so something like Lemmy/Reddit is equally desired 

 - a distributed address book, or rather a social network for humans in the style of Facebook and classic phone books

 - a direct textual messaging platform, long form as emails, posts as chat messages

 - a media sharing platform, for video, audio and images, so a gallery, player, playlist manager is a nice addition to have, supporting both private usage (like Jellyfin, Immich etc) and public sharing

 - an economic platform, the web3 who succeeded, not just limited to few-sat zaps but also substantial jobs with the "blog/website" mentioned early who are also job boards like UpWork, MTurk etc

 - for chats, nowadays people expect also VoIP call, so a TURN/STUN etc servers + the needed UI is in the wishlist of many...

 - for easier management a single integrated Nostr app-service including bitcoin core and lightning node to be fully sovereign on our own homeservers is also in the wishlist of many

 - since many have no knowledge to deploy homeservers BUT have some friends who have one with Nostr a web client with all the above stuff for them as well is on the dreams of many 

IMVHO the current Nostr problem is that there are a gazillion of projects started as experiments with no clear vision where to go, often abandoned, so most stay away because they do not know what to do and while interested have not enough docs to understand what to do for themselves. A "reference implementation" with the above stuff, maybe "modular enough" to allow third party features on the same codebase could be a gamechanger, like Matrix vs XMPP where Matrix succeed over XMPP even if technically not much nice just because they offer "a complete package" others with little knowledge of the project could get and play with.

This is not mobile of course, but could be as app on phone + webapp/service on homeserver. Try for a lower entry barrier for casual users it's a no go because they simply cost too much for a FLOSS project who start to have filled up blossom servers. A geek easy game to start could spread enough to push more people in at a peace slow enough not to clogs the public servers.