I don't recall him writing that it was a foolish thing to do. He acknowledges the risk of failure but doesn't ever seem to come to terms with the consequences of failure. Again, with the caveat that losing 5 years of your 20's isn't the worst thing in the world.
Until you're in your 30's and start thinking about what else you could have done with the time, I guess.
But you can't look at each bite at the apple independently, either.
A failed startup might very well be a better experience than three years at a great company.
But two consecutive failed startups is probably not better than six years at a great company, if you take the company job as seriously as startup founders supposedly do. After a couple years at a great company, an A-player is going to get much more interesting work.
I'm also not trying to argue against starting companies. Starting a company is a great idea. I just think you should be making money from day one, get yourself to a place where you can live comfortably with your startup, and not bank your time on a tiny chance of getting acquired. Time is precious, and 20-year-olds suck at accounting for it.
Only if you're talking about the quality of the experience, and even that's entirely subjective.
If the goal is financial independence there's no comparison. Your shot at making FU money working for someone else is essentially zero. As a founder your shot at achieving financial independence is considerable, and making FU money is at least a possibility.
Your odds of making "FU money" are so remote that it is irrational to bank your career on it. Remember, during all the time that you're chasing the "FU money" (that you again are never going to see), you are also not contributing significantly to your retirement.
There are plenty of reasons to start companies, but money isn't a great one.
I'm not sure what you mean about it having been disproven that starting a startup is a bad way to make money. Presumably the same sort of statistical evidence that would prove that randomly selected people can't do anything hard. But what matters to someone asking this question is not whether random people could do it, but whether they could, and if the right sort of person is asking the question, the answer becomes very different. If the young Bill Gates is asking the question, the odds of making FU money are close to 100%.
(Microsoft was already doing well before they got lucky with the DOS deal.)