The problem for me, speaking from personal experience with a spouse, significant others, friends, colleagues, my own degrees, reading job posts here on HN, is that it is the case that there are those who filter on elite schools before even considering any actual qualifications otherwise. It happens too often. After seeing this repeatedly, I personally do not believe that the selection effects associated with these institutions, controlling for other factors, is mostly due to markers of attributes like "motivation and team orientation" or even intelligence or skill, at least relative to other institutions. My sense of reading about other things like Navy Seals is similar — that some of the selection criteria are performative and of very little real-world utility, or might even be detrimental to the functioning of the institution or its aims.

This isn't to say that there aren't exceedingly competent people going into these elite institutions, only that my personal experience is such that the "magic sauce" beyond all the life history, accomplishments, test scores, grades, and so forth, is often bias or distributed gatekeeping.

Looked at differently, let's say you have an institution that aims to be elite, but the information provided by your selection criteria hits a wall, and the number of actually qualified individuals by prediction exceeds your capacity. In that case you have to either (a) basically do a lottery, which is honest but weakens the rationale for your institution over others, or (b) create criteria that are essentially useless but have a false veneer of rigor.

I'm not sure I think diversity quotas and so forth are the way to go either, but I also believe we need to stop pretending that these selection criteria are perfect or even near perfect, and that there's no bias either. I feel as if these discussions always proceed the same, that questionable or even objectively harmful (to the institution) criteria are pointed out, and then there's some outcry that lowering them will decrease standards, even as alternatives are never tested.