The problem is, you don't know which 50 it is. I mean, we're talking about 22 year olds here, they probably don't even know. You end up having to "advertise" to all of them, which then leads to you interviewing some of them who are 90% sure that they're going on to a PhD, but just want to do the interview [for experience|to make their parents happy|to see if it might change their minds].
Full disclosure: I work for Fog Creek, and went to Stanford.
Well, that certainly sucks, and I e.g. in 1997 worked for a new grad who had interned at Microsoft and either had an offer there or had accepted it before his elder brother convinced him to work full time at the brother's startup, for which he'd programmed their prototype. (The other guy they really wanted was in SV and still happy their, until his company went poof much later).
Anyway, what you're staying is that it's difficult to qualify your sales leads at MIT and that results in a lot of critical wasted effort. So much it's not worth trying (hard)?
Exactly; it just doesn't make sense for us to go to them and recruit. Of course, that doesn't mean we're going to ding them for having gone to MIT or Stanford if they come to us. We're a small company, so we need to focus our recruiting efforts where they'll be most effective.
I had a similar experience - I interned at a government organisation that only recruited at Oxford/Cambridge because they wanted "only the best".
But it was such a crap place to work they their actual hires were from 3rd rate institutions. They would have been better focussing their recruitment on them. But that would mean admitting there was a problem - so much better to convince yourself that the Oxbridge grads all went to do PhDs instead.
Full disclosure: I work for Fog Creek, and went to Stanford.