I would take issue with the assumption that code that changes the world and makes money needs to be written in a system-level language.
One of the beautiful things about the way technology is going is that we're able to take advantage of a range of technologies together, and use the most appropriate one at each layer.
Any world-changing app is likely to end up using code written in a system-level language, but that doesn't mean the app has to be written in that system-level language.
Take ML: with something like TensorFlow, I gather a data scientist is likely to write most of their code in Python... not a system-level language. The core of TensorFlow no doubt needs system-level work, but someone could write world-changing ML without using that language themselves.
Of course, you could view TensorFlow itself as a "serious, money-making, world-changing" task - and I'd agree, but I wouldn't want anyone to think that those words can't also be used to describe applications. Intrastructure and apps enable each other - let's not encourage a view that one is inherently better than the other.
I should note that I don't work in any system-level language myself. I've heard mostly-good things about go, but haven't used it. I've written C and a tiny bit of C++ in the past, but nothing significant. I don't think that's impeded either my career or my possibility of accomplishing things. But I'm still really grateful to people who do write infrastructure, and it's nice to see that the ecosystem of system-level languages seems like it's doing well.