The problem with this reasoning is that you assume you have a perfect view of how disadvantaged every single subgroup is. Also, given that this a zero-sum game

The problem with this reasoning is that you assume you have a perfect view of how disadvantaged every single subgroup is. Also, given that this a zero-sum game (the company only has a fixed number of hours to interview a single game), you are necessarily making someone else worse off when you give advantage to a sub-group of candidates. Also consider that many candidates can belong to a privileged group and a disadvantaged group at the same time. Of course none such nuances are being considered. How could they, when all you have on the person is their 1 page work resume? You know literally nothing about them, except a few projects they claim to have completed in the past. Now you're not actually hiring for skills, but playing disadvantage roulette with your hiring pool. Ok, maybe not, and you're still screening for skills, but at least call a spade a spade.