Carr's notion of "war" is outdated. The short history of internet hacking has shown that a) national borders are close to meaningless b) a defined start and end of hostilities is difficult, if not impossible to ascertain and c) the attackers and defenders need not belong to traditional defense establishments.
What the US government should do (and they may require this already, I don't know) is create a standard that would ensure security is built into any system developed for government use. Of course, relying on existing operating systems with known vulnerabilities means that the foundation of US government security is shaky.
The whole "Wolfgang-Pauli-Karl-Jung-I-Ching-We're-too-grounded-in-a-causal-mindset" seems like a pretty specious argument for why we would lose a war. I have a feeling that Carr read a book by Nicolescu and wanted to show people how smart he is.
Math is like love -- a simple idea but it can get complicated. -- R. Drabek