I do question if a breaking change for a custom language is the right way to go.
Built in Lua support in Neovim (with LuaJIT) has done wonders for the plugin ecosystem.
There is a lot of activity, and a lot of pluings with great quality have quickly replaced older ones.
That's no surprise. LuaJIT is fast, much easier to write, has a has a sizeable ecosystem and a good LSP server.
Even an overhauled Vimscript won't be anything more than a odd niche language with worse performance.
I would have loved for Vim to also adopt Lua and the Lua API from Neovim to prevent a complete ecosystem split.
I just don't see how implementing a new language makes sense in this context, except for fun.