I think that there seem to be two schools of thought in the USA on what defines performance and competence. If you follow the money you’ll see exactly the same thing happening in China happening here.
You've got wall street on the east coast and investment banks then you have the west coast with VC funding smaller businesses. As the size of the businesses get bigger and they are no longer created to be sold in six years, the downside of failure becomes much much higher. It's easy to assess performance in a small company and it's easy to kill projects that don't work in a small company. When you're Ford and your Firestone tires kill some folks life is a bit different, funding projects is a bit different, cash flow necessary to stay in business when disaster strikes is different, the types of employees you need to stay in business are different. What does this have to do with education and credentials? In these large companies it's much harder to measure performance and you get back to the old world view of who knows and trusts who. You're looking for a minimum level of reliability and competence since the downside of a screw up is much larger and the consequences are very different. I think these small companies are funded by people in large companies eventually buying them and those people agreeing to make the purchase I suspect went to Harvard and Stanford as did many of the VC partners who organized the funding. At the beginning and end of the project, the money passed through the ibanking system.
Silicon valley creates tons of companies which are bought and produce nothing but cash for the VC and deliver no or very little incremental profit to the company which buys it. It can no doubt be produced faster with a small company driving the project, but if you step back and look at the quality of the investment, you have an entirely different way to measure performance.
That means that the VC founders got money from east coast investors and the executive authorizing the purchase came from the exact same social circle. Performance metrics at the project level Paul Graham refers to are clearly met by any qualified professional from any school, but that employee amounts to an assembly line worker. If you measure performance in profit
I would suspect that the people who funded the startup attended top universities as those who bought it. Access to this kind of sure thing bet is limited to a group of people who are trusted by the previous generation who were doing the exact same thing.
The USA is good at moving money around into investments that often have no value to the investor. The founders and VC are out of the investment and the executives that authorized it are out of it long before anyone knows if it delivers real value. To call this performance based rewards is a pretty narrow view of performance. I think Paul Graham is brilliant, but he lives in a west coast piece of the puzzle which exits the process long before the end result of the pieces he’s created is known.