I am happy to report that I agree with Scott Aaronson – and obviously with Steve Pinker – that the universities should focus on the learning and scholarly work while sports and similar things should be treated as cherries on a pie.
Also, I agree with them that the standardized tests are – if you allow me to use a quote we invented along with Winston Churchill – that the standardized tests are the worst method to "rate" applicants except for all other methods that have been tried. ;-)
I've served on a Harvard undergraduate student admission committee, Harvard graduate student admission committee (repeatedly), and a Harvard postdoc admission committee, and I know something about the things that decide and whether they look OK. Just to be sure, Harvard is insanely selective, reducing MIT, Yale, as well as Princeton to schools where pretty much everyone can be admitted.
It's great if someone has a rock band or plays a sport game or flies to Africa to feed the children or something like that. I am usually touched or impressed when I hear or read such things. But at the end, it is not what universities should look at when they pick students (or other members of the community).
Also, standardized tests produced boring numbers and don't see the multi-dimensional qualities of a human being, her X-factor, invisible creativity, and blah blah blah. No doubt about it, it's true. But the admission committee ultimately assigns a student even less than one real number – it attaches one bit of information (accepted/rejected) so it's not surprising if a real number is used as an intermediate step to calculate the final bit.
At the end, the multi-dimensional perspective on the applicant is what encourages favoritism, nepotism, corruption, and similar things. The greater role is played by undefined criteria that are only seen by individual human members of committees, the less kosher and "demonstrably fair" the admission process is. There's more room for the bad things which – if we ignore their moral problems – show up as noise reducing the efficiency of the admission process.
Last year, there was a summit at the local science center asking how to stimulate the best high school students to work on their physics/mathematics skills and I defended the usual mathematics/physics olympiads against high school teachers who found them obsolete – and the logic was fully analogous. Vague, less quantitative ways to compare "top mathematicians at schools" etc. are less objective, less fair, and for those reasons, less motivating (I mean motivating them to do the right things), too.
I do think that even though MIT isn't quite Harvard in most respects, we may see that there exists evidence that Harvard has made a mistake when it did not accept Scott Aaronson to a much more lowly position than what Aaronson occupies at MIT today. It is not waterproof evidence, of course, but it is evidence, nevertheless. One shouldn't overlook that the odds of being accepted to the physics graduate school at Harvard are something like 1:15.
Peter Thiel is probably right that college education is overrated in general, especially when it comes to a young person's plan to become a great manager, inventor, entrepreneur, and so on. However, I still think that there are activities in which the universities are unchallenged leaders. If you will try to find people who investigate and follow research on string theory and who are not affiliated with a university, you will have a hard time to find the three people and even if you succeed, you will ask who is the third (and the second, for that matter). The number of string theorists in the Academia across the world is of order 1,000. You can't really expect a serious research of fundamental physics in the environments that look nothing like the Academia.
Aside from the string theory research, there are not too many other, very different yet indisputably great examples of highly valuable activities that universities are best at (or even enjoy a monopoly status, more or less). But there are some and I find it obvious that the universities aren't really needed for (or unchallenged at) the production of best athletes, among many other things, which is why the people at the universities shouldn't be led to focus on such things and students shouldn't enjoy major advantages because of these extra skills.
Scott and/or some of his readers argue that some of the tests were introduced to encourage the rural youth to get to the universities – which I find sensible because the scholarly environment is often isolated and degenerating (although these adjectives are often positive, however). Others, including the sport requirements, were meant to suppress the Jews whose average IQ is 10 points higher which makes them overrepresented at the universities. Scott believes that the Jews are as good at sports as everyone else these days. I am not sure but it's clear that a suppression could be achieved if someone demanded the students to be experienced in farming and agriculture – search for these two words in the famous 2005 speech by the 30th cousin of Scott Aaronson, namely Larry Summers.
Of course, I am suggesting such a thing as joke – as an example of the type of a contrived multidimensional selection that the universities should surely avoid.
Universities may train the students to become great at many activities and it's just fine. They may feel and surrender to financial incentives to produce future billionaires or powerful politicians etc. However, the more "common" the character of the training is, the less clear comparative advantage the universities (I especially mean the top universities) may boast. Preparing someone who seems excellent for the environment itself – i.e. for the scholarly work – is something that should always remain at the top of the good universities' agenda.