Conventional wisdom, and many studies, hold that 40 hours of work per week are the optimum before exhaustion starts dragging your productivity down too much to be worth it. I read elsewhere that the optimum is even lower for creative work, namely 35 hours per week, though the sources I found don't all seem to agree.
In contrast, many tech companies in silicon valley demand (or 'encourage', which is the same thing in practice) much higher work times. 70 or 80 hours per week are sometimes treated as normal.
How can this be?
Are these companies simply wrong and are actually hurting themselves by overextending their human resources? Or does the 40-hour week have exceptions?
How high is the variance in how much time people can work? If only outliers are hired by such companies, that would explain the discrepancy. Another possibility is that this 40 hour limit simply does not apply if you are really into your work and 'in the flow'. However, as far as I understand it, the problem is a question of concentration, not motivation, so that doesn't make sense.
There are many articles on the internet arguing for both sides, but I find it hard to find ones that actually address these questions instead of just parroting the same generalized responses every time: Proponents of the 40 hour week cite studies that do not consider special cases, only averages (at least as far as I could find). Proponents of the 80 hour week claim that low work weeks are only for wage slaves without motivation, which reeks of bias and completely ignores that one's own subjective estimate of one's performance is not necessarily representative of one's actual performance.
Do you know of any studies that address these issues?
Yeah, leaving the industry is extremely common, but in my opinion not outcome-optimal for the employers who are driving their employees to extremes (or, more commonly, not encouraging normalcy). There are indeed young recruits who are willing to come in and spend huge amounts of time and effort on game projects, but there is huge variance in work quality in this group, such that the flipside of fast, positive work from a talented, unstable, young programmer is the risk of losing time on poor work from a weak, unstable, young programmer... with little way to know the difference. A senior engineer at 40 hours is substantially better than a junior engineer at 80 hours, and so companies probably should invest more to keep talent on future projects, which is of course hard to do when crunch rolls around on the current project.
I used to hear this saying around the industry: "When you're on the outside of games, it seems like nobody is looking to hire good talent. When you're on the inside, it seems like there's nobody talented applying."