To me this reads more like "Even though humans can be modeled as distributed computing systems, we have to remember that they are humans because ethics" which is not my point. My point is that the human element is what makes people productive; it's not some tangent that we have to account for. IMO, the boundaries outlined here don't define the upper bound of meaningful work (the author's thesis) but rather the upper bound for mundane work if it isn't made to be meaningful. In the author's model, we'd say that superlinear productivity is possible _for individuals in isolation_ which definitely breaks the analysis since it's all about trying to hit that linear improvement mark. This doesn't strictly mean that their argument is "wrong" because they're reasoning about hours worked as their unit of measure but it does mean that it's not meaningful because a person's individual outputs can vary so greatly for a fixed number of hours worked.