In general, you'll hire people who, * are smart/productive enough to automate the drudgery, * are smart enough to get by, but whom the drudgery provides some la

In general, you'll hire people who, * are smart/productive enough to automate the drudgery, * are smart enough to get by, but whom the drudgery provides some lasting challenge, or * are screwing around and will likely screw around more as you keep them on for more hours. So, even assuming that simple work will allow you to get a significant long-term productivity boost working people 8 hours instead of 12, in the long run, you're going to be working with a set of people who'll likely be challenged by the work, so fatigue will cause problems if they're challenged at 8 hours/day, or you could be getting smarter work out of them if they challenged at 12 hours/day. Of course, building the tools to automate drudgery does take some time, and considering Ars Digita apparently operated during the first .com boom, time-to-market was critical, so some of this logic doesn't apply to the original example.