It's amazing how much you can do when you start unbreaking things.
Edit: That was excessively brief. So many parts about the whole investment cycle were broken. Some are marginally less broken now. Having a negotiation between a fresh engineering graduate who does not know what "pre-money valuation" means and the principal of a firm with 9 figures under management, and expecting that negotiation to produce an equitable or efficient outcome, is broken. Contacting angels serially is broken. Using who has attended a dinner party with your buddies as the most efficient screening method available for predicting skill with conversion rate optimization is broken. Investment outcomes where the company succeeds and founders do not end up rich is broken. The work culture in Silicon Valley is insanely broken. etc, etc
YC has unbroken some of these. It turns out that an application process and a few weeks with a team scalably works as a decently efficient proxy for success. It turns out founders can be taught what pre-money valuation means and when they know this they get screwed less often. It turns out that spending half of one's time raising money is, as expected, just a deadweight loss.
Your comment made me look at YC as an extended extremely thorough onsite interview, and it really clicked. But then I remembered that YC companies pitch demo day to a bunch of relative strangers-to-the-teams, not the YC investors themselves. Or do Demo Day investors get quality time with the teams during the season?
Of course YC is overall a mix of interview and internship and summer camp and summer classes and more. The above is an attempt to think through one of the threads.
The angels get one extraordinarily significant bit about all presenters prior to the Demo Day presentation: they were in YC. That is a very significant bit relative to another similarly situated twenty-something which a gleam in their eye and prototype on their MacBook.
It used to be implicit that some angels were saying "If you're good enough to be in YC, you're good enough to get investment from me." That is now 100% explicit for at least one deal. The vetting is getting outsourced, streamlined, and productized.
This is a good point, but to your original point: "invest in me just because I impressed PG (plus more good reasons that may be true but the investor cannot see)" does not fall into your category of "unbreaking things" like the rest, especially at the start of YC. By now YC has certainly built institutional credibility, but even that can be broken if one day PG turns evil and sells out the brand like Pyrex or show or a grade-inflating college.
His interest is actually the same as for any kind of certification, and that is to increase the number of people getting certified for as long as the certification is seen as having value.
Thus it is a little scary that every Y-combinator class is bigger than the previous one.
However he is also figuring out how to scale the model and provide more assistance. For example current graduates benefit from finer tuned screening, have more kinds of assistance available, better assistance (pg knows more than he did when he started), a bigger network to draw on, and save more otherwise wasted fundraising time.
On balance the result is that classes seem to be getting better, not worse.
However I guarantee that a point will be reached where the Y-combinator model can't keep improving and growing at the same time. At that point his personal financial interest will still be to keep growing the program. It will be interesting to see what happens then.
Edit: That was excessively brief. So many parts about the whole investment cycle were broken. Some are marginally less broken now. Having a negotiation between a fresh engineering graduate who does not know what "pre-money valuation" means and the principal of a firm with 9 figures under management, and expecting that negotiation to produce an equitable or efficient outcome, is broken. Contacting angels serially is broken. Using who has attended a dinner party with your buddies as the most efficient screening method available for predicting skill with conversion rate optimization is broken. Investment outcomes where the company succeeds and founders do not end up rich is broken. The work culture in Silicon Valley is insanely broken. etc, etc
YC has unbroken some of these. It turns out that an application process and a few weeks with a team scalably works as a decently efficient proxy for success. It turns out founders can be taught what pre-money valuation means and when they know this they get screwed less often. It turns out that spending half of one's time raising money is, as expected, just a deadweight loss.