My problem with this line of reasoning is the inferential leap. The other day my wife was surprised that I could box jump higher than her, even though I rarely work out, while she had been training for a stair climbing competition. Being a man has its advantages. But then we're going to make the jump from that to saying "and that explains why men just prefer/are better at programming and thus make up 90% of the industry!"
My second problem is that men set the metrics in ways that benefit them. Let's take the stereotypes as an assumption. What is a good programming team? One full of cowboys and rockstars, or one full of people who can concentrate on detail work for long periods at a time and communicate with each other? Studies of children show that girls are better at the latter (which is why they get higher grades).