Interesting observation, though it becomes fairly readily explained on several points.
The militaries of WWII were fighting an active, unrestricted, war. The US military of today is active, yes, but highly resstrained. aRather than occupy and subdue, its primary role is to project power -- think "gunboat diplomacy" for an earlier example, rather than exercise it. The nuclear arsenal as a prime example.
Where the US does exercise force, it is typically highly focused and targeted. It is not unreasonable for foreign journalists to repprt on ongoing air raids from exposed locations within the cities directly targeted (see Bagdad, in both gulf wars, in ways that reporting from a firebombed Tokyo or Dresden would not have been. Murrow's rooftop reporting of the London Blitz was notable specifically for its daring.
Several authors have noted (some, such as Robert J. Gordon apparently unconciously) of the near le=inear relationship between energy input per worker and productivity, particularly in the case of factory manufacturing of the 1st half ofvthe 20th century. Whether measured in mechanical horsepower or kilowatts, the relation holds. The US military brings a tremendous amount of energy (itself largely responsible for the pollution mention in TFA) per soldier, sailor, airman, and marine. That includes embodied energy in vehicles, arms, and structures, as well as motive and destri]uctive power, again with the nuclear arsenal as its apex.
To wrap my examples: the military is also a highly-tuned logistics system. Much of the task of a military is package delivery -- parking hot lead, dU, HE, or other actives, where they will have a desired effect. Getting the parts and support where and when needed is complex, and the DoD is a masterwork of central planning efficiency. Result: fewer workers -- combatants -- required.
As a parallel to the role of mechanisation and automation in general employment, there are multiple lessons to draw.