If you doubt this, take a few hours to study what Transformational Technologies Portfolio holdings Orbital Sciences (ORB: NYSE) and SpaceDev (SPDV.OB: OTC BB) have done with budgets in the millions of dollars in the satellite and launch arenas, then compare it with NASA projects.

You’ll gag.

So I’m especially pleased to report that at least one Federal agency seems to be doing something right. Actually, it’s doing a lot of things right.

I’m talking about DARPA (Defense Advanced Research Projects Administration). It’s the Pentagon’s “gee whiz” arm. It’s charged with identifying and funding development of truly revolutionary technologies to empower tomorrow’s military.

At DARPATech, the agency’s annual tech trade show-cum-conference, they showcase a lot of this stuff. However, they don’t take themselves too seriously, which is perhaps a reason for their success. (When much of work is really play, people seem more productive.)

I saw military officers in full dress regalia doffing black eyeshades and pantomiming Men in Black (you know, the ultra high-tech uber-agency profiled in the Will Smith comedy of the same name).

It wasn’t a bad analogy. Not when you consider what these “uber-geeks” are REALLY working on…

One mathematician-in-uniform, Dr. Benjamin Mann, cheerfully announced that his program has discovered a new foundation for ALL OF MATHEMATICS. Specifically, he said that they’ve “proved” (that’s a very significant word amongst mathematicians) that all of math can be recast with data sets, rather than numbers, as its fundamental building blocks.

The good doctor cheerfully skipped over what this signifies in real-world applications. (He did express confidence that a whole new mathematics would arise sometime in the 21st century.) He has that privilege; in pure mathematics, it’s almost an indicator of superiority if one’s work is divorced from “reality.”

However, it’s not necessarily so. Consider Georg Friedrich Bernhard Riemann, the great 19th century mathematician (by the 20th century, most folks had decided that two names were sufficient). Riemann went to his grave proudly proclaiming that he had never done anything practical.

Guess what? His “non-Euclidean geometry” became central to understanding Albert Einstein’s curved space-time.

Since then, other “pure mathematical systems” have found surprising real-world applications.

So, when Dr. Mann tells us he’s going to recast all of mathematics on a new foundation, I’ll be the first to admit I don’t know what it means. Probably, neither does anyone else. But if there were a futures exchange for mathematical discoveries, I’d be aggressively long on this one.

DARPATech probably had more PhD’s per capita amongst the 2,500 participants than anywhere I’ve been outside MIT. Not all of them were looking ahead a century.

How about something practicable in the next decade or so? Dr. Ralph Chatham spoke of how new developments in software and user interfaces have ushered in the era of the “virtual co-pilot.”

It’s expensive and often logistically difficult to arrange for fighter pilots to train with actual wingmen and controllers. Now DARPA has developed “virtual,” artificially intelligent actors who are so good in flight simulators that they pass the Turing Test.

The Turing Test is attributed to the famous 20th century mathematician Alan Turing, who didn’t mind doing practical things. Essentially, what the test says is that if you can have a conversation with a computer, and are unable to distinguish it from a person, then it is truly intelligent.

Some have objected to the general nature of this statement. Who is conversing, and what is the subject? A computer might fool a layman on the subject of medicine, but not a doctor. Or it might be expert enough in medicine to fool even the doctor, yet know nothing of baseball (an ignorance which, however, many of us humans share).

Critics agree that a well-designed Turing Test would focus on a specific field of knowledge wherein the computer is “knowledgeable” and articulate enough to fool an expert practitioner. By this standard, various programs are now intelligent. (For instance, go play chess at a local game store and ask the program its opinions of positions on the board — many grandmasters do.)

Now, DARPA has taken this into a whole new realm. Artificially intelligent programs simulate wingmen and controllers in flight simulators so well that the trainee pilots cannot distinguish real people communicating with them through the console from the software doing the same.

Think about that. In principle, we are at the dawn of a time in which the whole field of air traffic control can be automated. Logically — though DARPA didn’t go here, perhaps for political reasons — the fully automated plane and battlefield vehicle should soon be feasible.

Closer to home, there’s no reason why this can’t enable automated driving of cars within the next 10 years. Wouldn’t you prefer to get into your car with a cup of coffee and read or enjoy broadcast media while your vehicle navigates the morning “traffic creep?”

Some of us who’ve sworn abstinence might even be willing to once again drive between the hours of 6:30 and 9:30am.

To your profitable future,

Jonathan Kolber
September 06, 2006

Your Tax Dollars, Actually at Work was originally featured in the Penny Sleuth.