I think the big question the author fails to address is: Do these deluded engineers end up regretting their decisions in the long-term?
I think most people in the startup community understand that failure-- and the misery that comes with it-- comes with the territory, and the key is whether you can learn with each failure. And I would argue that there's no better to place to learn than an environment where your income depends directly on whether you're genuinely solving a problem for people (the people who "run the lottery" are your target customers, not the VCs, by the way). In my case, I've failed at over a dozen projects (some prior to my time at Google) while finding a few moderately successful ones along the way, and I'm continuing to fail, learn, and grow.
And I like to think that this better understanding of harsh realities of how the world actually works gained from doing a startup-- whether you succeed or fail, whether you end up at a small company or a large company-- is a return on your investment that continues to serve you for the rest of your career.