I think it's a bit of a popular straw man that a 50/50 split is desired. It's more along the lines of, "There are no obvious artificial disadvantages for people

I think it's a bit of a popular straw man that a 50/50 split is desired. It's more along the lines of, "There are no obvious artificial disadvantages for people in these groups, and if that happens then the resulting distribution will almost certainly change." If you look at the published #s right now, they're... very striking. It makes the hypothesis that it's just the preference of women and people of various minorities quite an aggressive hypothesis, in a bayesian sense.