I was rank 8/432 and applied to: Harvard, Princeton, Yale, Columbia, Stanford, Pomona, Rice, and UW. Got rejected from everywhere except UW so that's where I ended up going. I think only one other person in the top 10 went to UW with me, and he had plans to be a CS major. Funnily enough over 50% of my graduating class just went to UW because it was: good enough, close to home, and relatively affordable.
You stand out by being a top performer, and will attract other top performers as long as you search them out. Also, in those scenarios, professors will give you more attention, because you are a better student and more enthusiastic than most of your peers.
Perhaps if you are a true prodigy, you might need a top-tier program to reach your full potential... at that point your peers are the professors and high-performing graduate students, but for most students, I think there are pros and cons for being in a top 5 vs a top 50 program.
Yes. That's why I used the top 50 CS programs as an equivalence class in my post:
>> the difference between "unranked LAC" and "top 50 CS" is difficult to overstate.
The US has 5,000 colleges and universities. Not 500. 5,000.
You are absolutely not going to find more than one top performer every half decade or two at a small non-selective LAC or the branch campus of a university system. If ever. I spoke with on faculty member at a branch campus who said that he's never had a single student who is as good as the average undergrad he taught at <top 5 program>. He's been teaching for 20 years. Those types of institutions comprise the vast majority of US colleges and universities.
I think the "rankings are just noise" attitude is mostly held by people who don't even think about the existence of 90% of US colleges and universities. If you consider Stevens Institute of Technology a "backup" as opposed to a "reach", then I guess the attitude has merit. But if you're one of the 50% of college students who get rejected from Stevens -- or the even larger percentage who don't even apply because they know they can't get in -- then the world looks different.
(Also, he didn't state this as a negative or a positive. Just as a fact. "Different institutions serve different clientelle". You don't have to be a hotel snob to say that the Holiday Inn you manage isn't as nice as the Ritz, or a elitist that the youth swim team you coach has nothing on the US olympic program... some people -- particularly educators -- aren't obsessed with being "the best".)
This is entirely compatible with the idea that top tier schools are overselling the quality of their students.
For this reason, rankings (unfortunately) have more value that OP suggests. The "best" students will cluster around the "best" institutions. Does the average student differ muchst between #1 and #5? Probably not much. But the difference between 10 and 100 does, by a lot, and the difference between "unranked LAC" and "top 50 CS" is difficult to overstate.
If you're measuring is "how good is my peer group", rankings are often a bit of a self-fulfilling prophecy. Not so much at fine granularity, but more so if you measure in terms of "20-50 position overlap equivalence classes". This is particularly true in fields like CS, where there are huge differences in curriculum between the top schools and the not-so-top schools.