Seems to me that Haskell's ability to infix named functions is a much nicer solution over-all
I wonder if a whole set of infix matrix operations could be added to reduce the need to load numpy for simple tasks.
A @* B
A @. B
A @+ B