I think Y Combinator doesn't take its model far enough. What I'd like to see could be called "Human Resource Investment"- investing in people, not startups per se. This would be structured as a debt and/or equity deal. You find a talented person, and invest in them using money, technology, education, and social resources. Then you charge a high interest rate (for the reliable people), or make money off of a % of all their future income (for the high risk types) for the next 30 years or whatever (in addition to the principle).
The advantage to the investee that this has over credit card debt and other forms of debt is that you don't require anyone to make debt payments until they are making over a certain $ figure per year. This way nobody has to worry about going bankrupt, becoming poor, or social shame. They also get access to major institutional support, which is difficult for people to get (and social/institutional support is one of the few things people are generally missing in their lives, if they don't go to church). No longer do the poor, inexperienced, young- but energetic and talented- types need to scrap for everything.
The advantage to Human Resource Investment, Inc., is that even if there is a high rate of delinquency, just a few Zuckerbergs and Grahams and you make a fortune. You also can make money off of them, even if they fail a few times, since it's a long-term contract, rather than a project contract. This way other VC firms don't make money off of Y Combinator having educated someone with a failed venture, and then they do another, successful startup with Sequoia or whomever.
You also have access to a large pool of talented people with whom you have a good relationship, and you've already evaluated, who would make good potential hires for startups or whatever other companies you are affiliated with. If HRI were savvy enough, you could lock up most of the available talent the VC world relies on, before anyone is taking a percentage of the money they're going to eventually make.
Think of it as government for profit- you choose the people to invest in, helping them enormously (as government services are supposed to do) and you make a profit with "taxes". Given that governments have to invest in everyone, and are inefficient to boot, yet still roughly break even, I see no reason why a private company couldn't do the same thing, but focus on talented people, do a good job of it, and make a nice profit.
I've done the math before, and the conservative estimates I used worked out to a very handsome profit. But it gets complicated, and it depends very much on the assumptions that are made. If anyone wants an example, I'll be happy to provide it.