It's a weird jump of logic to say that because the biggest Ruby project (Rails) is overkill for a particular architectural flavor (JS + REST backend) that this

It's a weird jump of logic to say that because the biggest Ruby project (Rails) is overkill for a particular architectural flavor (JS + REST backend) that this will lead to Ruby's decline. Even assuming the market moves wholesale to pure REST backend (which is quite unlikely frankly, because not everything on the web is a CRUD-heavy app, documents still dominate), and Rails does not fit the bill (which is also arguable since moving to pure REST API only makes views/helpers irrelevant, which is at best 20% of Rails), then you would see other frameworks like Sinatra/Padrino rise to prominence. The only way Ruby becomes irrelevant is if a new paradigm becomes dominant that just isn't easy to implement in Ruby. Threaded/evented/asynchronous frameworks do fit the bill, but my sense is that enough people prefer Ruby so much to a language like Javascript (in the case of Node), that if something like web sockets become indispensable as an 80% use-case, then the Ruby community will figure out how to make it easier to use them. I'm still using MRI which has some difficult problems to solve, but it seems JRuby is pretty much ready to pick up the slack if it really becomes necessary.