But I do a lot of REST API (and WebSocket) work and all of the workload that happens inside my Node.js program is extremely lightweight.
If I need to perform some heavy computation, I will offload it to a separate child process - Node.js forces me to put that code in a separate file/module but I actually like this because it encourages separation of concerns. It feels very natural so I don't really need any other special constructs.
I looked into goroutines a while ago; it looks cool, but I probably wouldn't use them much because I don't like the idea of having code from the same source file splitting off into multiple processes/threads; it makes is harder to read and reason about the code (this is a bit like what happens with multi-threaded code when you have mutexes all over the place).
To me, this feature has the same utility value as the ability to define multiple classes per source file - Ok, that's cool, but is it a good idea to do that?