Well, it's hard to argue with her basic claim, which is that it's not at the college or at the job-recruiting level that we need to focus our attention if we want to be more gender-balanced; it's on the little kids. Speaking as a college educator, I can confirm that the social biases are already set, if perhaps not in stone.
I guess you could claim that gender-balancing the field is undesirable, which is strange, or that it is impossible, which is an interesting claim but I'd need more evidence. But otherwise you'll need to look at the little girls and work from there.
But that's exactly my point; we've already done a lot of work on that front. You can see the results everywhere else. Are we really sure that after decades of work on this matter, work that has obviously come to fruition in field after field after field, that there isn't something other than "social bias" at play? At what point does the field get to stop self-flagellating?
I guess part of what I'm saying is that at this point I reject the idea that the burden of proof is somehow on the field to prove that it is open to women or that girls are somehow being excluded in a special way, after these decades of work. I'd say the burden of proof at this point is on those to demonstrate that somehow computer science has specially somehow failed to solicit woman in a way that the other fields didn't, in an environment in which this stuff was becoming cliche even by the time the women entering college today were being born.
Twenty years from now, are we still going to be having this exact same conversation and making the exact same accusations of entire fields, if it does indeed turn out that women of their own free will don't want to go into computers on average as a career?
I guess you could claim that gender-balancing the field is undesirable, which is strange, or that it is impossible, which is an interesting claim but I'd need more evidence. But otherwise you'll need to look at the little girls and work from there.