I've been an Emacs user for over 20 years. I like it, and find it productive. But it's stuck in a rut, in a very similar manner to Perl 5. If that isn't addressed, it will fade into the background. I'd argue that's already happened. None of my coworkers in any of the jobs I've worked in have used it, and that includes academia and various industries. It seems to be restricted to the uber-nerds who find out about it and go through the pain of learning how to drive it effectively. I used to use it for everything, from GNUS to org-mode, but my usage is declining.
The keybindings need to be fixed. CUA has been the norm for what, 35 years at this point. C-y is paste. C-z minimises the window (WTF!) C-_ is undo. It's not just nonstandard and counterintutive, it's unnecessarily nonstandard and counterintuitive. They aren't more efficient, they aren't more effective. It's a barrier to adoption which doesn't need to be there. It has a certain familiarity for us long-time users, but that's about it. Every other application uses something based on CUA, so it's not like we couldn't switch over.
While I certainly wouldn't want a "GNOME 3" type of makeover, where all the actual userbase are cast aside to be replaced by a hypothetical potential userbase which doesn't actually exist, I do think it's well worth a look at present-day alternatives and to make sure it doesn't fall behind. Because I think it actually has fallen behind, and it needs to get its act together to retain its relevance.
Recently I got a personal JetBrains subscription and started using CLion. While I still miss many little small features, there are quite a number of big features which Emacs doesn't attempt to do, and so the small annoyances are worth it for those large benefits. Emacs needs to be able to do better than this to not only keep up, but to do better than its competitors.