I recently had the privilege of attending a Silicon Valley conference attended by leaders across the national security “innovation ecosystem.” The term reflects today’s veritable freshet of interest in defense innovation, from self-styled “virtuous insurgents” and defense “hackers” to individual agency innovation offices and entirely new outfits with on-the-nose names such as the Defense Innovation Unit and the Defense Innovation Board. All this may suggest that the national security apparatus is at last confronting the need for long-overdue changes to how we do business.

For two days, I listened to senior people from the military services, large defense agencies, and major components of the intelligence community as they described various “mission acceleration” efforts—that is, finding shortcuts that allow us to do what we’ve been doing a bit faster, a bit cheaper, a bit better.

This is a problem.

Innovation—from the Latin innovare—literally means to “make new.” But defense and other national security leaders often confuse it with automation or modernization. Automating an existing process doesn’t change the process itself. Nor does it change the game to incrementally improve the range, speed, or—forgive me—the “lethality” of existing weapons. Such efforts are like a homeowner fixing a broken window, painting a dilapidated wall, or adding a bathroom without considering the decaying foundations of the house itself.