* If the real goal is to find overlooked, qualified people, we should expect to see more moneyball type thinking, and less lazy diversity-for-diversity's-sake thinking.
* The current trend of hyperfocusing on social causes of behavior and making biological ones taboo seems dangerous. I would argue that a comprehensive, accurate description of how people work makes you better at grounding your compassion in effective results. To that end, overlooking something as big as biology seems an easy way to enact a policy that's well-intentioned but harmful.
* Some people may appeal to biology in a sexist way and maybe history has many examples, but some people really just want to know what the truth is. Another way of making this point: https://twitter.com/clairlemon/status/846520296240693248
Google is supposed to be a data-driven organization.
If the engineers they've hired can't tell the difference between statistical distance and the utility value of a Gaussian distribution, I think they have an entirely different issue on their hands...
But there isn't. There is an argument that women are, on average, better than men at chess. But that argument does not apply to an individual person. Why would you use male vs female average IQ when judging who becomes a member of the chess team? Why not assess the candidates IQ? Or, maybe have them play chess and see how well they do?
There is a persistent theory that diversity programmes exist to promote underqualified minorities over qualified white people. That's not true. The aim of the programmes is to find people who have been disadvantaged by this exact kind of (often subconscious) "well when you average group X out over the entire country ignoring educational achievement and employment history, typically they're not so good" thinking and have lost out on past opportunities.
You're right, I was using some excessive shorthand in my original post and will amend it - the assumptions are not genetic, they are social. But the point still stands, why be making these assumptions at all?