The sentence in question was: "You can't run a startup at crunch time with devs putting in 7.5 hour days." The "at crunch time" was meant to qualify the rest of it -- to suggest that standard hours might work some or even most of the time, but not when you get that big client on the hook who wants something custom next week. I think that's in line with what you're saying.

With that said, most of the successful startups I know involved long hours early on, and not just occasionally. It absolutely involves borrowing from future productivity and present quality of life. Sometimes taking that loan is worth it, and sometimes it's just a stupid, pointless death march and the urgency is manufactured. The latter might be more common but the former definitely exists.