I haven't read the book, only reviews, but based on those, it seemed to me that that McAfee's thesis relies on carefully selected numbers. For example, using "number of cattle" as a proxy for resource input into the dairy industry, even though everyone else just measures actual resource input.
It leaves me wondering if the book (which I admittedly won't be reading; my existing reading list is already long enough) ends up functioning a bit like that satirical paper on smoking and long distance running from 10 years back. In it, the authors "convincingly" argued that marathon runners should all be smokers. The overall structure was a sort of numeric Gish Gallop: the paper focused on a constellation of individual measurements that are plausibly associated with both smoking and long distance running performance, while carefully steering clear of any measurement of smokers' actual athletic performance.
Hm, seems to contradict this: https://en.wikipedia.org/wiki/World_energy_consumption
Here I see energy consumption going straight up.