I'm pretty sure that we actually agree about this, generally; my point is that having a sense of urgency that comes about from recognizing that the company is unprofitable due to salary expenses has, in my limited experience, done wonders to get people off of needless "optimizations" and onto making stuff people want.
I do insane workweeks because I'm a Japanese salaryman and people would talk if I didn't. What the heck is your excuse. ;)
Seriously, every time I hear stories about it, I get the feeling people are trying to be Real Manly Men doing their Rough Startup Labor (say them in the Monster Truck voice), finding things to spend 80 hours a week on. Which, since you can't sustain productivity for 80 hours a week, ends up getting frittered away on negative productivity (introducing more bugs than you fix) or just spinning your wheels doing things that don't matter (backend).
(The "you" here is a collective "you".) You may think you personally sustain productivity for 80 hours a week. I think you are too much a Real Manly Man to have ever collected data to substantiate that. How did I do? (We track it at work, which is finally starting to impact the culture here. And we have a bunch of Real Manly Men looking at the chart showing the brick freaking wall in the units-of-engineering-labor created versus time graph and saying "Well, sure, it might LOOK like we're not being productive... but that is just because some people need to man up and work harder.")
I used to work in banking and VC. 80 hours was easy. More work in banking, more travel (which sucks just as much) in VC. I don't get gaming developers complaining about crunch, but I'm always thankful on Labor Day that labor gave us the five-day, 40 hour workweek as a standard for when (if?) I get married and have a family.
People enjoy working on the stuff they're good at. A lot of times success comes from working on the stuff that you're not good at. Rebuilding a backend sure can "feel" productive, but it often isn't the right decision for that moment. That's why I prefer to think of it as working "harder" - it's a useful shorthand for "don't neglect the stuff that sucks to work on". A need for revenue, based on a little higher cash burn, can often make the correct resource allocation decision more obvious.