The Arpanet was a DoD funded research project. So was CPOF - it's the main C2 tool used in the Army for something like 15 of the last 20 years. So sometimes the research dollars produce results. But the DoD also invests in stupid things, as you cite above. Its location in China wasn't necessarily dumb from a national security perspective as long as nothing classified was transmitted, i.e. the data flow was all one way. The research itself probably was stupid.