November 22, 2002
I’ve actually worked on some of this stuff. It was for the National Reconnaissance Office, doing an unclassified (!) study of future information technology for the Intelligence community; the first phase was on looking 25 years out, and the statement of work explicitly said we could assume any technological breakthroughs we wanted to, except FTL and time travel!Darius Bacon makes some similar points, including an explanation of “truth maintenance,” here.
I have no brief for John Poindexter, and I’m thoroughly disgusted by the casual disregard this administration has for privacy, but there actually are a lot of technological ways to improve intelligence without Big Brotherish implications—mostly by improving how we convert data (of which we-the-U.S. government already have overwhelming amounts; the last thing we need is to collect more) to information, and (even more important) information to knowledge (meaning something you can actually do something with). Alas, there are an awful lot of people in the intelligence community who are absolutely (and IMHO idiotically) convinced that the One True Way of the future is the One Giant Database.
[…] [“Story telling, change detection, and truth maintenance”] have entirely legitimate technical meanings, and they’re real problems that lots of people have been working on since long before 9/11 or the Second Bush administration. “Story telling” is the process of conveying key information in a connected and memorable fashion, and is incredibly hard to automate; “change detection” is the problem of distinguishishing real changes from random noise and systematic errors in complex data (most often in images). I can’t say I’d run across “truth maintenance” but by slight extrapolation from some of the problems I looked at, I suspect it’s referring to trying to keep a single common picture of a situation (“truth”) updated when many people are independently both updating it and looking at it — the classic example being trying to keep all the military commanders in a theater looking at the same map, so the guy who got his information from headquarters at 9 a.m. doesn’t send his troops into the path of the airstrike that got scheduled at 2 p.m….
On the subject of the sheer loopiness of the IAO’s web site, Stefan Jones observes:
I wonder if the IAO has been taken less seriously that it could because its presentation—the stunningly brash name, the over-the-top logo, the jargon-laden web page—is so unbelievably campy.[01:06 PM]
People scared of terrorism might find this comforting (given that it’s a Republican administration…if the IAO were established under Clinton there’d be freepers in the streets of D.C. throwing rocks and burning cars).
Skeptics, on the other hand, might merely be amused. It’s like something out of a fundamentalist techno-thriller, in which it would be the agency in charge of bar code tattoos and kidnapping the hero’s girlfriend.