It's a good case study in the downfalls of a waterfall type process and of committing too early to a system before it's true character is known.
I'm not sure that economies of scale can get you that far when you're talking something as complicated as a rocket. How many jet engines does Rolls Royce make every year? Rolls-Royce and GE make many hundreds (thousands?) of jet engines per year, but they still cost a crapload of money each.
You might be right, maybe fully-recoverable rockets aren't possible. On the other hand, being personally unqualified to judge, if I've got some random piscinymous internet commenter telling me that it's not possible on one hand, and Elon Musk telling me that it is possible on the other, I'm more inclined to believe the dude who owns a rocket company.
This was the logic behind the space shuttle. But unfortunately, they have to survive to high stresses that inspecting and repairing them between launches costs more than building a simpler disposable launch vehicles.
I honestly believe that as long as we will be using rockets, the best cost savings will not come from reuse, but from increasing the launch volumes to the point where launch vehicles stop being hand-built boutique items and start being mass-produced commodities.
This announcement gives me hope that this might happen during my lifetime.