Loren Amelang writes:

There is one catch in this example (and a lot of similar ones in this
thread) though... If you start with one liter of 200 psi gas, and let
it expand through an air motor into a receiving tank until you have
two liters of 100 psi gas, you can extract some work, but you don't
end up with those two liters in the receiving tank.

A valid point.

However the application to the original problem is that you can regulate
the pressure down via a regulator (essentially a feedback-controlled
variable restriction) or you could conceivably accomplish the same
regulation down via an air motor. The latter case produces work output
that somehow must be accounted for in the former, namely by waste heat.