The US Air Force wants a supercomputer to help contractors navigate its bureaucracy. As Camron Gorguinpour, the USAF’s “director for transformational innovation” told Air Force Times earlier this month, “there are thousands of pages of policies, laws, and regulations that affect Air Force acquisitions,” and contractors need help understanding what they all mean. In theory, if the USAF can “teach the system how to understand context,” it will be able to “answer questions accurately and become a resource that personnel can access.”
On second reading, I found the whole thing reminiscent of Project Cybersyn, the brainchild of Salvador Allende’s Marxist government for centrally managing the entire Chilean economy with a pair of IBM 360 mainframes. All the inputs and outputs of every enterprise in the country were theoretically to be modeled with a ginormous optimization program—the New York Times once described that as trying to “find the right software for socialism.” Alex Tabarrok of George Mason University more cogently described it as an expensive joke—in retrospect, those two 360s had less processing power than an iPhone, and there’s no app for running an economy. There’s also no app for running a bureaucracy either, but there are quotidian ways to make it worse.
As Exhibit One, consider this new business of IR&D sponsorship. Alongside his effort to get suppliers to “innovate,” Under Secretary Frank Kendall has proposed a new regulation. Reimbursable research and development expenses are just fine, he thinks, but they should only be so independent. Thus, all new IR&D projects will need some sponsor in the Defense Department to sign off that the effort is useful. Whether that’s a flag officer or maintenance technician has been left unspecified for now, but Kendall insists that getting approval should be no trouble. But if it’s no trouble, then why require sign-off at all? If there’s no wicket to clear, then what value is added by the review?
Quite in contrast, I foresee two problems, The first is from the bureaucracy this requirement adds. Creating another box to check might be only one box right now, but as Gorguinpour would confirm, those boxes have been piling up for decades. At this point, by the estimate of the Professional Services Council, some 30¢ on every federal contracting dollar is spent on compliance costs. Creating any new regulation without first deleting a bunch of others sends a terrible message to the bureaucrats—go ahead, add another requirement, no one ever holds you accountable for anything anyway.
The second problem is that the need to secure a sponsor for a new idea really means only new ideas we say want. So, finding a sponsor will be easy? Great—which office in the Air Force will get working on a novel idea for a counterinsurgent ground attack airplane? Of late, the USAF, and the US Army too, have been pretty good at discouraging dissent from the official line. Hopefully contractors will be able to find relative mavericks who buck that pressure. These people systematically think outside the box, but some also tend to color outside the lines, and get in trouble as a result. At some point, there will be a list of who’s allowed to sign those IR&D forms, and don’t expect that the John Boyds of the world will be on it.
As Dan Gouré of the Lexington Institute noted last week, if the USAF needs “a cognitive thinking machine [to] analyze the vast quantities of data, track cost and benefits, and easily navigate the library of U.S. government codes and regulations,” then something is deeply wrong with the regulations. So here’s another idea, Camron. So before innovation becomes the new transformation, try doing something actually transforming. Just cut those thousands of pages down to dozens.