Article is confusing "startup" with "small business", if your field and your methods are old enough to be tried-and-true then, yes, by all means, apprentice yourself and learn the ropes. After spending the 7 or so years working in a social-networking company, you too can create another stable small-business in social-networking that (for example) staggers users' twitter posts or sells slightly ridiculous shirts to facebook hipsters. If you balance your books and keep at it, you'll probably make a handsome chunk of money and maybe even make it really big one day.
If your field / technology / process is dangerously new and untested because you just invented it (e.g. you're mining titanium from the moon using gravity fields, you're desalinating sea water, you're selling a headset that streams movies into your right eye checks your social feed in your left eye and sucks your dick), there's going to be preciously little other people and books can teach you. You've entered the realm of vague opinions, uncertain grey zones, and questionable futures. In that case, you're only choice is to either not do it and wait for more tests to return, or do it blindly in hopes of learning on the job.
PG and his cronies recommend the "just do it" path for startups in the second category because if you succeed, you'll find yourself in a temporary market monopoly where you'll be entitled to make astronomical amounts of money. PG and his cronies then invest maybe 1% or so of their utility while you invest around 99% of your utility. If your business succeeds, PG and friends get a 1.5x boost to their utilities, while you get maybe a 2x boost to yours. If your business fails, PG and friends shrug it off and offer you some words of encouragement as you pick up the splattered mess of your life off the floor and write a book about how the world wasn't ready for the movie-streaming-eye-sucking machine.