"For instance, while Scala boils away a lot of repetitious code, it's still there under the hood. It feels like an architectural band-aid"

I don't understand this argument. Even so-called elegant languages like Scheme still have 'repetitious' MOV, JNE, etc instructions under the hood. Unless you're running on a lisp machine, everything you write executes very repetitiously in machine language under the hood.

Why make a distinction between one abstraction and another? All of them are architectural band-aids, to use your parlance.