Tim Oren's Due Diligence

The Other Side of Milbots

Further to yesterday's post about the Grand Challenge, the invaluable StrategyPage takes a look at procurement of frontline combat robots:

The U.S. Department of Defense has awarded a $140 million contract for more battlefield robots, as well as parts, repairs and training. This is just for one model of MTRS (Man Transportable Robotic System), the iRobot Packbot. Actually, there are several models of Packbot, but the most popular job for these remote controlled vehicles (that look like a miniature tank, with an extendable arm where the turret should be) is checking out possible roadside bombs, or booby traps, or the insides of caves or buildings believed to hold hostile gunmen. For over a decade, there have been larger (over a hundred pounds) remote control vehicles like this for bomb squads. But better designs, and smaller electronics, have made the "man-portable" (under 40 pounds) units possible. Two years ago, there were several different MTRS systems in Afghanistan and Iraq, but the iRobot equipment has emerged as the most popular, and most effective. This means troops no longer have to get close to possible bombs, and risk getting blown up when nearby terrorists detonate the device. The MTRS has become very popular with the troops, and saved many lives. The Packbots proved very user friendly and reliable, so now over a thousand of the $115,000 MTRS units are on order.

Comparing these devices with Grand Challenge style autonomous ground vehicles is really an apples and oranges mismatch. Very few of the 'combat' bots have any weaponry installed - they are recon and EOD specialists. And their range of autonomous action is at most a limited amount of self-steering and obstacle avoidance.

For the investors our there, iRobot - the maker of the described Packbot - is in S-1 registration.

Let The Games Begin: Grand Challenge Underway

The second running of DARPA's Grand Challenge autonomous vehicle contest is underway, and the entries seem to be doing better than the first time (PDF): This time a nearly half the first day's entrants made it through the shake-down course, while the first time around only a quarter of them were able to make it out of the starting gate. For those of you down in SoCal, the preliminary trials are going on in Fontana through noon, October 5, and it's open to the public:

Spectators are welcome to witness the semifinal activities, which are conducted at the
California Speedway in Fontana from 7:00 a.m. to 6:00 pm. daily. Admission is free. Food and drink
concessions are available; however, DARPA encourages spectators to bring plenty of water, a hat, and
sun block.

Real 'Robotwars' - for free - what's not to like? There's then going to be a for-real 150 mile race in the Nevada desert on the 8th, for the top twenty qualifiers, with a $2m prize for making it to the finish line. Don't expect too much: Last time the 'winner' made it seven miles, which actually exceeded the DARPA program managers' expectations. This time the teams have had a lot more lead time to prepare and test, and have picked up some major sponsorship, but the DARPA guys have also made the race course a lot nastier than the first time, with obstacles designed to defeat systems relying on single sensors (such as GPS).

To the extent this makes it into the MSM, we're likely to see a focus on the 'war bots' potential, but that's really not the game in this case. Infantry combat bots seem to be evolving in the direction of being cheap, portable, and largely controlled by the troops. DARPA is instead likely aiming this part of the autonomous vehicle program at another military bugbear: logistics. The military has long fancied a mechanical 'mule' for logistical support, and has in fact spec'ed one out, but it has heretofore lacked any convincing path to getting control software to drive it. If the GC project eventually works out, there will be fewer troops exposed to IEDs and ambushes while hauling supplies, a larger fraction of manpower pulling triggers, and maybe - eventually - some mechanical assist for special forces schlepping ammo and batteries over hill and dale. Not a Bolo, but darn useful.

Thinking about Credential Risk

"High West" has a very nice post on an identifier indemnity index that puts some structure on the peak of my hierarchy of privacy unease, which I'd labelled "Direct Financial Loss". The framework usefully stretches that undifferentiated phrase out into two dimensions of value and impact (though they don't feel completely orthogonal).

My one quibble, likely resulting from the orientation of the original post towards (federated) identity providers, is that it's more than identity that's involved here - most of the entries in the relevant chart are really credential, not identifiers. It's often the binding of the credential to the identity, not the identity itself, that comes under attack.

Operation Katrina: Soldier's Relief Fund

This is a day when the whole blogosphere is making their endorsements of Katrina relief efforts. I'm late, but better than never:

I heartiful suggest Soldiers' Angels Operation Katrina - Soldiers Relief Fund. There's a lot of active duty and reserve military who are forward deployed, and call Louisiana or southern Alabama or Mississippi home. About now some of them are finding out that the families and homes that they thought safe behind them, aren't. This fund is to stand these families back up again. They and their loved ones have been putting it all on the line for us. Isn't this a good time to pay back? Or if you want to look at it another way, we know exactly what streets these families' men and women have been walking, and what they have been doing there. Isn't that the kind of Americans we want to support?

I've been over there and hit the Paypal button. You go too, eh?

Eric the Half VC

(OK, my name's not Eric, but I wanted a gratuitous Monty Python reference.)

Been a bit quiet around here lately. There's a reason. I've taken a half time leave of absence from Pacifica Fund, to become interim CEO of a seed stage startup. We're still in stealth mode, so all I'll say for now is it's in the area of database security, and more will emerge over time.

Not only is this a serious commitment, but for now it's heavily engaging whatever set of neurons is responsible for blog posts. Those also seem to be the ones required for pitch decks and business plans. I'm not hauling down the flag here, but activity will be rather sparse until we're through this phase. In the long run, I expect the contrast of having one-half seat on each side of the table, so to speak, will make for some interesting blogfodder when there's more time for reflection.