Well, I think survivorship bias is separate (though, agreed, it is endemic in our industry) -- the logic wouldn't change in this argument even if YC had only funded Airbnb and Dropbox. It may be a case of misattribution by the founders concerned, but hey -- in the absence of RCTs, trusting them seems like the right default.
> The thrust of his argument is that YC is asking to be granted special privileges, in a circumstance in which they happen to directly benefit, in a way that only helps them unfairly cement their already privileged position.
I don't think Sam's argument is quite as self-serving, though I see how it could be interpreted this way. I think he'd agree that YC should not be a gatekeeper. (As in PG's original proposal.) But if the argument against the original proposal is that we somehow need more data before rolling it out everywhere, running a smaller experiment with YC would seem to be better than nothing at all.
I completely agree with you that a steady-state world of privileged visa-granters is not what we should be trying to bring about.
Wouldn't objective analysis be the right default, rather than trusting those that were successful within an organization that has incentives that are, in the majority, contrary if not in outright competition, to your own?
YC isn't a charity, as I'm sure you're aware, and success is not guaranteed. They're in the business of throwing spaghetti at the wall to see what sticks.
Nobody wants to be the spaghetti that doesn't stick, but the vast majority will be, and it's in YC's interests to keep convincing new people to jump in the pot.
So no, I don't trust them, or their motives, or think that it's to the benefit of anyone other than YC to grant them special immigration privileges. I certainly don't think outlier examples justify or prove YC's value to the American people, or to foreign citizens who might wish to pursue their interests here.
> The thrust of his argument is that YC is asking to be granted special privileges, in a circumstance in which they happen to directly benefit, in a way that only helps them unfairly cement their already privileged position.
I don't think Sam's argument is quite as self-serving, though I see how it could be interpreted this way. I think he'd agree that YC should not be a gatekeeper. (As in PG's original proposal.) But if the argument against the original proposal is that we somehow need more data before rolling it out everywhere, running a smaller experiment with YC would seem to be better than nothing at all.
I completely agree with you that a steady-state world of privileged visa-granters is not what we should be trying to bring about.