Yes, exactly this, obviously in the 56th hour next week if you want to you can 'do something'; the point must surely be (to mean anything at all) that if you attempt to sustain that, the 'first 55' suffer more than the 56th rewards.

As another commentor says, it simply can't be that marginal productivity drops to zero. At least, if you're trying to produce, there's some external motivation, then it'd take something really serious (starvation, massive sleep deprivation, etc.) to make it actually zero; more than just 'a long week'.

It's like if you're super motivated and work all-night on something exciting: if you were rational for a second, you'd realise you'd probably accomplish more on it with a few hours' break to sleep. But that doesn't mean the alternative is doing nothing in the last x hours.

> As another commentor says, it simply can't be that marginal productivity drops to zero.

It can even fall below zero. Imagine the totally overworked surgeon killing his patient because of fatigue and total exhaustion. His net productivity has fallen, reducing the outcome of the last 12 hours.

No, that's 'just' a bad outcome of his productivity.

His contribution to 'product' is providing the service of surgery; he has done that.

But yes, overworked and tired surgeons are more likely to kill patients; killing patients bad; overworking surgeons bad. (It's just not a 'productivity' issue.)