The Air Force wants a new supercomputer. Not as part of the Defense Innovation Initiative which seeks to more rapidly develop revolutionary military capabilities. Nor to process the data from ISR platforms operating against ISIS in Syria and Iraq. Nor to counter the massive, sustained and relatively successful attacks on Pentagon networks by Russia and other adversaries. Not even to figure out ways of penetrating China’s anti-access defenses and defeating its area denial capabilities. No, the Air Force has an even more impenetrable adversary in mind.
Apparently, the Air Force finds defense acquisition so challenging that it is starting a program to develop a supercomputer to take on the task of dealing with the system’s impenetrable regulations, endless reporting requirements and overly-complex and cumbersome contracting processes. According to Military Times, the Air Force doesn’t want just any old computer but “a cognitive thinking machine, that can analyze vast quantities of data, track cost and benefits, and easily navigate the library of U.S. government codes and regulations.” The Air Force recently awarded a contract to develop this supercomputer with a target delivery date of 2018.
The Air Force’s success in the conflicts of the past several decades have been largely the result of its use of computers to achieve orders of magnitude improvements in navigation, weapons delivery, air tasking orders and countering hostile air defenses. It should come as no surprise that it wants to apply this experience to the problem of guiding private businesses and government officials through the labyrinth that is the acquisition system.
That the acquisition system has become increasingly complex, cumbersome and costly is not news to those who have to deal with it. Recent efforts at acquisition reform, which included directives to do away with unnecessary and counterproductive regulations and processes, have had no demonstrable impact on cost, schedule or performance. The only glimmers of light in the otherwise impenetrable darkness that is the acquisition system comes from entities such as the Rapid Equipping Force, the former Joint IED Defeat Organization — now the Joint Improvised-Threat Defeat Agency (JIDA) — and the Counter Terrorism Technical Support Office that have special authorities, unique funding mechanisms and different attitudes.
The Air Force’s idea of using a supercomputer to attack the acquisition system reminds me of a suggestion made a few years ago by Arnold Punaro, the then-chairman of the Pentagon’s own Defense Business Board. As reported in Breaking Defense, Punaro declaimed that if the Pentagon wanted a system that delivered on time and doesn’t cost too much, then it should take decades of regulations, totaling thousands of pages, and “put a match to it.” Personally, I think Punaro’s solution would be more effective than that proposed by the Air Force.
Does anyone else appreciate the irony that the Air Force is attempting to go through the existing acquisition system to develop a supercomputer to figure out how to navigate the acquisition system? Care to bet whether or not the new thinking machine will be delivered on time and on cost? Or that it will work as advertised? I hope the Air Force is collaborating closely with DOT&E to develop the appropriate test plan for its miracle machine.
The supercomputer program has an inherent flaw. Its goal is to build a “natural language query system … to provide users insights into defense contracting statutes, regulations, practices, and policies.” This assumes that the acquisition system was built on the basis of rational thought which can be boiled down to a set of rules that can, in turn, be imbedded in search algorithms. In truth, the acquisition system defies the rules of logic, common sense and good business practices. Sometimes it even attempts to defy the laws of physics.
The Air Force is putting too much faith in computers. Unlike Watson, the IBM computer that defeated the best human competitors on the television quiz show Jeopardy, my bet is that Congress and the acquisition bureaucracy will be able to stump the Air Force’s new supercomputer.
Find Archived Articles: