Recently I praised the Pentagon's new offset strategy and repeated my support for a strong military R&D program. Today I see a War on the Rocks piece noting $53 billion in failed or cancelled programs that were part of the previous offset strategy, called the Revolution in Military Affairs [RMA].
To me the failures weren't in the concepts or the early research, but in the later phases of the design and procurement processes -- when unrealistic cost estimates soar and change orders complicate the program, slow it down, and add to costs.
I still agree with the view of the DARPA director when I worked on R&D matters for the Senate Armed Services Committee, who said, "For DARPA to succeed, we need the right to fail." He meant that DARPA would work on many cutting edge ideas that failure was inevitable and even helpful in showing what wouldn't work.
We still need to fix the acquisition process, but shouldn't worry so much about failed experiments.