I think Neovim itself has been stabilizing quite a bit in terms of upgrade breakage.

A common reason for breakage is/was:

  - Neovim changes some API (deprecating, ...).
  - User upgrades Neovim and theres some incompability, OR user upgrade plugin and that plugin assumes a much newer version of Neovim. (I've often seen Neovim plugins "mandate" either the latest stable Neovim or even HEAD).

But:

  1. Neovim has been including some popular plugins (or at least their functionality domain) in the past few years. This obviates the need for plugin-for-$THING, and reduces breakage.
  2. ISTM that the pace of (new) plugin development in the Neovim-sphere has slowed down.

The latest example of #1 is vim.pack, which is a plugin manager similar to vim-plug, mini.deps (vim.pack is based on mini.deps), lazy et cetera.

I can remember removing vim-commentary (from tpope) a while ago because Neovim included something like it in the main distribution. Granted, that specific plugin never broke because it uses the stable viml API.

I think a big difference is that in the end emacs often makes a call and adopts one of the very popular packages to the core - eglot, modus-themes, use-package - there are certainly more, and more will come. It may not make everyone happy, but is sets the baseline - e.g., I am using eglot as package manager, but I wrap it into use-package commands for compatibility reasons.

No such thing exist in neovim (or at least in times when I was using it), so that churn never ends. Also I find, that neovim ecosystem is concentrated on one (very productive) developer in an unhealthy manner - folke often takes time off and half the packages one uses stands still.

But in the end, while I like neovim, I also find that emacs ecosystem has better ideas - which-key, embark do not stop to amuse me (I will not comment on whether it is a good thing for a text editor). I also do not like lua and actively dislike the experience of debugging and configuring neovim with it (maybe less of an issue with LLM these days).

In my experience, running in a terminal absolutely adds a bunch of rendering/performance issues and all kind of surprising failures with hotkeys.

Neovim suffers from the Javascript kids mode of development. Constant change, constant churn, the mirage of stability always behind the corner, you always require third-party packages for functionality that should be core, completely erasing the Lindy effect of vim proper.

It’s a bit sad Neovim has stolen the thunder from the original work of Moolenaar & co. My guess is that neovim will splinter itself down the line further again once lua stops being attractive, while vim & Emacs will keep chugging along for another half century.

LSPs keep getting reimplemented, package managers keep getting reimplemented. It's a bit like the react version of text editors.

I used it more than I use emacs, but I agree with the assessment of doom emacs vs neovim.