> Just calling it an array might be underselling it, at this point I like the NDArray terminology used by numpy. I think it gets the point across more clearly

> Just calling it an array might be underselling it, at this point I like the NDArray terminology used by numpy. I think it gets the point across more clearly than “tensor”, and conceptually they’re pretty much equivalent