It doesn't seem like market efficiency was the central thesis of the essay. He was saying that countries which don't allow their citizens to get rich are worse countries to live in. That's empirically true. Based on that observation, he argues that income variation is actually a sign of health. None of that relies on efficient markets.

Most "highly valued" startups usually disappear in thin air once founders get their payouts as if they had never existed (ahm... Viaweb?).

Viaweb was designed to be sold. Pg didn't want to work on online stores the rest of his life. But it went on to become a profitable part of Yahoo. That seems like a win-win for both Yahoo and the Viaweb team.

now you can argue founders are 100X better in skills to exploit market and make believe everyone that their goods are far more worth than it actually is. Should they be paid 100X for these skills? If you think of benefit to the human kind as a whole, perhaps not.

If someone paid the founders a certain amount of money for those skills, then those skills were worth that amount of money.

What does it mean to "benefit humankind as a whole"? That presupposes humankind has a purpose. Who gets to decide how much benefit something has to humankind? It seems like the answer is "the market."