Admissions at Harvard or Stanford doesn't use a random number generator to determine admissions after a 'meets minimum requirements' filter. It would be as if a student got accepted to Harvard, Stanford, MIT, 5 state schools and 10 community colleges and then used a coin flip to determine where to go.
That’s another way to solve the >10x applicants issue.
With this scheme, if a school inflates their grades it will have a larger adjustment down. It's entirely based on past performance of people from the school, so there's no standardized tests to worry about. Theoretically, one could have an adjustment upwards, but I've never heard of it happening (I participated in department meetings where they discussed admissions in general). The adjustments are usually somewhere between 20 and mid to high 20s percent. Some are higher (not many), but a few are lower but not by much.
Reputation is also very fluid over time, Yale arguably took a big if likely short term hit from George W. Bush. Of course with 5 presidents Yale clearly has some serious connections.
I have even seen things swing both ways as schools are looked at as a great talent source. This lets less capable people in on reputation. Resulting in a backlash where people think X is overrated.
PS: Don't forget the supply of Genius is smaller than the number of slots at top schools and many of them take other paths. So, you got a pool that's above average, but not that far above average.
Working in concert and amplifying each other over time, these factors are responsible, ultimately, for creating the benefits that we all perceive as intrinsic to an Ivy League education -- namely, engaging and networking with lots of smart and motivated peers and mentors.
In the private sector there is a bias to cut out overqualified candidates ex: MS or several years experience for entry level work. The assumption is someone may need a job now in a downturn, but that would move on quickly or have higher salary expectations.
Asking about it someone also said "People also flat out lie which makes a significant qualification / role mismatch suspicious."
And finally, I suspect there may be some jealousy involved.
Because everyone else is farming the same universities, making these candidates overpriced. Google and other tech companies have considerably widened their net, when they could no longer rely on Stanford et al alone.
As for the 4-yr university degree, it's always been more about discipline than anything else. A completed degree is as you say - a signal for discipline - which is probably why more and more employers are caring less about the exact type of degree, and more about whether it was "quantitative".
This is highly dependent on the field. (see mechanical engineering), but you are also correct in that is signals this person has the ability, dedication and perseverance to learn complex tasks and items for an extended period of time (in terms of IT/tech).
Back then, there was absolutely no substitute for the in-house training I got from going to classes and doing my homework and taking the test etc etc...
Today, I am far from convinced that people cannot simple learn the informational part of it all through (yes albeit intense) online courses, videos, and training.
Do we want people building bridges and building without proving they understand their static and dynamic physics and structural limitations of materials and such??
Of course not...but these things can be solved via various technical means that do not necessarily require sitting in real-world, outdated classrooms.
Of course, networking has always been a big part of the benefit being accepted to an elite university, so while I accept the learning part of it will be easily disrupted, I am not really convinced (yet) that the signalling and the networking can be replaced so easily.
[edits]