Actually the GIL is not necessarily an issue for NumPy since C extensions can break the GIL. I don't see why this needs to be a winner take all competition betw

Actually the GIL is not necessarily an issue for NumPy since C extensions can break the GIL. I don't see why this needs to be a winner take all competition between languages for statistics. It's worthwhile to have different approaches and just because in the author's opinion R is better than Python and Julia doesn't mean that it is always the better option no matter what.