Yes, that could be the case. I'm not a mathematician but I gave my fair share of struggle towards trying to understand the mathematics behind physics that I was studying. You don't need to know what a smooth manifold is to solve problems in classical mechanics, just as you don't need to know what a tensor product of vector spaces is to learn what tensors are in the context of neural networks. The former is always harder because the standard of rigor is higher, and that's because of different goals. So it could be that a person tries to look up what a tensor is, finds the mathematician's definition first, thinks "what the hell is this", then finds a simpler explanation in applied contexts and goes away with the idea that mathematicians unnecessarily complicate things, because the reason why the mathematical definition has to be so complicated escapes them.