The original also failed to break down paper writing. For a top-tier conference, you're going to spend several weeks worth of full days working on a paper, and you're probably going to do it twice per paper because, as they say, even Simon Peyton-Jones has papers rejected occasionally. This time is on top of work you already did to get your system working! (well, unless the paper is the result, as in some areas of theory)
You really, really have to want to live it. I enjoy it, and as a PhD student I don't mind spending all my waking hours that aren't devoted to my wife or a little bit of exercise doing research. But just to be competitive, that's sort of the baseline. I'm not sure, though, whether it's truly a demand of the job or just the default because academic research attracts people who work like that (I also worked like that when I was a developer at MSFT).