Airbnb, Kickstarter, Twitch, GoodReads, Bloomberg, there is quite a list of companies who still run a significant amount of their site through Rails. Over time I imagine that particular bottlenecks and performance critical pieces have been pulled out into other languages and services as necessary, but I certainly wouldn't feel bad about it if I was developing or maintaing a Rails site today. It's significantly better than PHP, and a hell of a lot of sites are still run on that. The reality is that if you really need to scale, your primary bottleneck will probably be the datastore, not the frontend.
I'd say it's more use Rails for all of your standard CRUD operations and then use Go as a module in Rails using Quartz/FFI if you have any algorithms that need to be high performance. Of course you could always go down the microservice route and spin up a Python/Go service for your more intensive data processing modules.
Twitter using Rails to grow to the point where Rails became the problem was actually Rails's selling point. If you ever grow to the point where Rails is your problem, then you have money to solve that problem. But if Twitter hasn't used Rails, they might not event get to the point where their product is out the door.
It's absolutely correct to say Twitter blamed their issues on Rails. It's less likely to be correct that they wouldn't need a similar major change if they had started with anything else and grew from an MVP to one of the busiest sites in the world. I personally find it doubtful that they couldn't have done it with Rails.