Paul Graham's essays contain much wisdom and truth, but even Homer nods, and let's be honest, the paragraph you quoted is bullshit.
Even if you care about nothing whatsoever except your job (which probably isn't - and certainly shouldn't be - the case) there is an optimal level of effort that delivers peak sustainable output, and for tasks with a significant intellectual component, the available data says that optimal level is about thirty or forty hours a week. Beyond that, sure, you'll get more done in the first week, but chronic fatigue will quickly build up, reducing your effectiveness to the point where you're actually getting less done than you would have in a thirty-hour week.
Worse, chronic fatigue is like being drunk: it impairs your judgment to the point where you can't tell how impaired your judgment is. You feel like a hero, when your actual performance is more akin to someone coming in to work drunk every day.
That doesn't mean you shouldn't do a startup, if that's what you think is your best option. It does mean if you're going to, you should do it in the understanding that it is not actually a get-rich-quick scheme, and it doesn't suspend the normal laws of human biology and psychology.
Even if you care about nothing whatsoever except your job (which probably isn't - and certainly shouldn't be - the case) there is an optimal level of effort that delivers peak sustainable output, and for tasks with a significant intellectual component, the available data says that optimal level is about thirty or forty hours a week. Beyond that, sure, you'll get more done in the first week, but chronic fatigue will quickly build up, reducing your effectiveness to the point where you're actually getting less done than you would have in a thirty-hour week.
Let's see some of that research. "A Grand Gender Convergence: Its Last Chapter" Goldin, NBER working paper suggests that there are superlinear returns to hours worked for differentiated labour. Where workers are basically substitutable pay has an almos 1:1 relationship with hours worked. Where people can have unique knowledge and relationships and combinations of same working more hours makes your total output greater and makes you less substitutable, which makes your bargaining position better, if you're an employee.
Even if you care about nothing whatsoever except your job (which probably isn't - and certainly shouldn't be - the case) there is an optimal level of effort that delivers peak sustainable output, and for tasks with a significant intellectual component, the available data says that optimal level is about thirty or forty hours a week. Beyond that, sure, you'll get more done in the first week, but chronic fatigue will quickly build up, reducing your effectiveness to the point where you're actually getting less done than you would have in a thirty-hour week.
Worse, chronic fatigue is like being drunk: it impairs your judgment to the point where you can't tell how impaired your judgment is. You feel like a hero, when your actual performance is more akin to someone coming in to work drunk every day.
That doesn't mean you shouldn't do a startup, if that's what you think is your best option. It does mean if you're going to, you should do it in the understanding that it is not actually a get-rich-quick scheme, and it doesn't suspend the normal laws of human biology and psychology.