In defense of the interviewer, if somebody told me they wrote an OS for the Raspberry Pi I would assume they wrote it mostly in C, not mostly in Bash.
Someone who picks up that info and turns it into interview material should be well aware about the difference between making a distribution and writing an OS!
The idea here is that to achieve something quite useful like a distro ("made an OS") for the Raspberry Pi, the important part is to figure out a problem that needs solving and a tool to solve it. A person who made something cool might be actually quite dumb, and that's alright. A good answer is something that foremost solves a problem. A bad answer is a fancy technology that doesn't do it better.
Here I sucked at my Go solution and I deserved to fail. But here it's just a lame fail because my initial solution worked and the rule changed to accommodate the fact that it was not the technology he expected.