Rails needs websockets, just like years ago when it needed a better way to handle organizing and deploying assets. The modern application demands it, and if you've ever had the pleasure of trying to build a websocket application in Ruby you'd understand why it's actually really exciting that we're going to get the Rails community to build this for all of us.
If you want to be disenchanted about some aspect of Rails then pick the fact that the need for diversity at RailsConf this year has yielded some of the worst presentations I've ever seen (like, literally, the presenter spoke poor/broken english and the Docker oriented talks were about 6 months behind).
Rails moves the way it does and absorbs the ideas it does because they become popular. It's much easier to add a feature that everyone wants, instead of inventing a feature that makes a solved problem that much easier/faster/what-have-you. DHH seems to have to explain this every few years, but go back in time when Rails adopted sprockets and you'll see posts not unlike the one you just wrote.
Realistically, Ruby and Rails jobs are still among the highest paying in the industry. Startups and other more established companies still choose Rails as their platform, and not just for bullshit brochure sites. Certainly there remains a "best tool for the job" mentality but, right now, the alternatives are to: build it yourself, hop on a less "mature" framework -- Pheonix? I don't even know what you pick when you adopt Elixir as your platform language and need to build an API or front end apps but Pheonix won the google -- or choose from one of the popular technologies like Express.js, Laravel(?), or Django. That part of building technology (and business) didn't change, and there are still hundreds of thousands of Ruby and Rails engineers out there.