Monday, August 31, 2009

Directionality of Hierarchal System Evolution

Too busy to write anything new on the Diagonal Economy at the moment.  I am, however, having an interesting email exchange with a friend who is currently deployed to Iraq and is tasked with shoring up their hopelessly over-extended intelligence databases.  We've been discussing how to "fix" the problem, with me (predictably) recommending adopting a flatter, decentralized information processing system that removes the analyst from the loop and uses some form of blog ecosystem to unify operator and analyst into a single position.  While I never really expected this to be an implementable solution, over the course of the conversation it became increasingly clear to me that hierarchal structures can't "go back"--they can't easily and selectively implement decentralized, p2p approaches to information processing because to do so effectively would be fundamentally antithetical to their constitutional form.  A short excerpt:

The more I think about the issue, the more it seems that there's just a fundamental directionality of hierarchal systems such as the military.  It's not possible to move away from the centralized analysis model to a decentralized model because that would lead to a landslide of decentralization, and the whole structure would break down.  The "blogging intel ecosystem" works best when it's done exactly the way the "enemy" conducts business (US cathedral v. "enemy's" bazar):  it's unclassified, anyone can participate, and often your very funding and operational capability depend on not only your participation but your success in that ecosystem.  If analysts were paid only by the number of hits and links their intelink blogs received (e.g. their google ranking, for lack of a better term), then suddenly you'd have an amazingly well populated and up-to-date system (side note:  the current system is so bad that most intelipedia pages are still largely the same as when they were cut and pasted from wikipedia in the first place!).  For example, if the US military stopped paying people, and instead paid for operational success in some kind of market-system, and if the US military abandoned all rank/hierarchal structure and let people organize in whatever way worked best to get their piece of the pie (payment for operational success), then there would be real value in open and decentralized reporting and analysis being conducted by the very people who are also using that information to operate.  Of course, anything structured like that would never have gotten itself into the royal mess we're now in...

More to the point, minus the mumbo-jumbo, and my conclusion is that this entire set of solutions that I'm hinting at is fundamentally unavailable to the military because of its structure.  For the concept to work as a solution, the military would need to abandon its structure to such an extent that it would no longer be in need of the solution.  And this structure is also the source of the original problem.  

In all honesty, do you think that this problem will ever get better?  Which is more likely to happen:  1) "they" add another 300,000 hours of predator video first to your analytical load without considering the consequences to your processing/exploitation/dissemination system, or 2) "they" conduct a top-down re-evaluation (complete with the budgetary authority to make real changes) of how their system functions and begin to collect data with the efficiency of the overall process in mind.  What you're grappling with is a symptom of a structural problem that will only continue to get worse, not better, until the structure is addressed.  While I have no doubt that you'll be able to improve the system to some degree, that will work in a way like Jeavons' paradox, and make the overall situation worse:  by improving system capacity by some amount, the immediate need to address the underlying structural problem will recede and you'll get 500,000 hours more predator video, not 300,000 hours, because now you can handle it.  Which, of course, will only get you back into the same jam you're currently in, but with more invested in a flawed structure and less elasticity of that structure to respond to future demands because you've picked the low-hanging fruit improvements already...

Begs the question:  to what extent is this unique to Nation-State military structure, or, as suggested by Tainter and others, is it impossible to voluntarily contract the scale and scope of hierarchy, leaving collapse as the only possibility?

On a semi-Diagonal Economy-related note, does this support the argument that we must focus on building diagonal structures rather than adapting existing, hierarchal institutions, or is that overreaching?


Jeff Vail said...

Here's the counterargument to this post:

Guided Emergence

Ruben said...

I emailed this relevant snippet to a colleague just this morning...

Incompetence in high places.

From the absent-minded genius, somewhere along the idiot savant spectrum who cannot tie his own shoelaces, to the Machiavellian manipulator, motivated purely by greed or tribal advantage, there is also the incompetent chameleon in charge - the placeman with no business being in place.

In On the psychology of military incompetence by Norman F Dixon, Dixon examines the (many) blunders of (mostly) British generals over the years. As well as considering the nature of their incompetence, he examined the nature of the organisation that put them in charge.

One compelling idea I got from Dixon was the fact that many organisations cannot exclude incompetent placemen from being promoted through their ranks. In fact, the structure of the organisation actually attracts jobsworths who hide in plain sight inside the hierarchy.

The very tools used to measure competence are learned by rote by these placemen, who excel in all areas of organisational efficiency with no real idea how to apply them effectively. Once the Peter Principle has promoted enough of these people to their level of incompetence, they essentially control the organisation, including selecting future promotees.

Eventually, of course, an army may be tested in battle and it can rapidly become apparent if it is commanded by Colonel Blimp.

It's not just militaries in peacetime that suffer this disease - youth organisations, churches and education systems have great difficulty excluding paedophiles because successful careers in these fields grants unsupervised access to children.

It's therefore not too much of a stretch to imagine the higher echelons of government and the regulators being chock-full of yes-men and placeholders, all wedded to a flawed group-think and never tested in the battles they now face.

The lunatics have taken charge of the asylum and the roof is falling in.

Jeremy said...

Thank you Jeff, this is a fascinating analysis that brings many ideas to mind, and seems to suggest many different scenarios that could be better understood by this sort of thinking. As a computer programmer I appreciate the reference to Eric Raymond's "The Cathedral and the Bazaar". This tie-in shows that we can look to the Open-Source software world for instructive examples, as you have already pointed out. Also you are making an argument for an effective market economy for information - which in this case improves the product. But I wonder if some of the problems we have observed in pure markets (the financial crash for instance) would also apply to an information market; this might be interesting to think about. I see the good that could come from the market approach, but maybe there are also some weaknesses if that is the entire approach.

Jeff Vail said...

So, my friend responded:

"Well, to say that centralized analysis to take all the data together
and apply statistical techniques is useless and go by google page rank
is kind of like saying "Fuck Gallup, and all their 'scientific
method', Perez Hilton knows what is going on in America better than
they do!"

And granted, I'm sure Perez makes more than an average Gallup employee..."

And my response:

"While it's clearly a case of the pot (me) calling the kettle black, you're comparing apples to oranges!

Perez Hilton does know what's going on in the celebrity gossip realm fairly well, or so I'm told. He may have no clue about predicting the next election results, but that's not why his site is "high ranking." The more pertinent comparison is whether the top 5 results in Google for "celebrity gossip" have a better grasp on celebrity gossip than would a centralized organization using statistical sampling mechanisms and traditional data management protocols to guage "celebrity gossip," and which of these two structures could respond most rapidly to changes. My bet is on Perez Hilton (or whoever takes over from him when the "rules" shift).

Now translate that to intel. Let's focus on something measurable: accuracy in predicting the next 20 attack locations, times, and modes in Iraq. Which of the following two structures would have more success:

1) A traditional intelligence Wing of about 1000 people, led by a general, with all the standard rank and structure.

2) A group of 1000 people who each write a blog about predicting the next 20 attack locations, times, and modes in Iraq, and where these people are paid ONLY by the "Google ranking" of their blogs relative to each other--#1 gets general officer pay, #600-1000 get airman pay, etc.

Both groups have the same geographic distribution, same incoming education, same access to outside intelligence, same IT support, etc. Who produces the better result? Again, my money is on the "Perez Hilton" of the blogging group over the one product produced by the hierarchal wing structure. Can I prove this? No--not either way. The only thing I am confident in is that, even if I could prove with 100% certainty that the blogging group would significantly outperform the traditional group, that structure is so antithetical to those in charge that it would never be implemented..."

Rice Farmer said...

From my own experience with intelligence work in the US Air Force, I found this particularly interesting. And I agree that a "blog" approach would be preferable. In fact, some of us put out an unofficial "grapevine" newsletter, which (although the subject matter was different) did a surprisingly good job of digging up "hot scoobies" with minimal resources. Your post led me to compare the two approaches to information gathering and analysis for the first time after all these years.

TH in SoC said...

Jeff, your post makes me think of the present Federal efforts to pass a national "Cyber Security Act" which would give the Federal Government power to monitor U.S. internet use, and to shut down the Internet in the event of a "cyber security emergency" declared by the President.

What is your take on this proposed legislation? What do you think its real aim is? How real is the threat of cyber-warfare? And workarounds do you think ordinary citizens will evolve to maintain free and unrestricted information flow and networking even as the Federal government attempts to tighten control over the Internet?

I've packed a lot of questions in here. Hope that's okay...

Jeff Vail said...

Rice Farmer: unfortunately, I don't see this method being adopted, and not for the inability of mid-level people to understand it's potential... some things never change (until they collapse).

TH: I haven't looked in to the cybersecurity act much, but what I've heard is that it only provides for the government to disconnect private networks to keep the larger internet operational. I don't think this is intended to allow the gov. to shut down the net, especially because the economic imlications would be too great. However, that's not to say it wouldn't facilitate this. At the end of the day, my take on it is that, if and when the time comes that the president wants to curtail some or all internet useage, he will do so and ask permission later. I think this would be a last ditch effort, though--hard to maintain the pretense of a liberal democracy after that, so it would more likely be accompanied by other martial law actions. To be honest, I'm not particularly worried about the quick jump to fascism--I'm more concerned about the gradual slide...

Joel said...

Hm...heavy-duty statistical analysis is not unknown to the blogging world.

You might say Nate Silver knows the mind of the electorate better than Gallup, even if Perez Hilton doesn't.

As to the persistence of hierarchy: I think a person's ideal of what power looks like has a lot of influence.

William Polk, via Malcolm Gladwell, writes: "As quickly as he could [after initial success as an insurgent, George Washington] devoted his energies to creating a British-type army, the Continental Line. As a result, he was defeated time after time and almost lost the war.”

In that case, perhaps the best you can hope for is to make hierarchy look so virile that your enemies are led astray by their own Tyler Durdens.

Rice Farmer said...

Jeff, based on my experience in the military, I think your assessment is correct. Everything has been invested in a top-down, chain-of-command structure, which governs the way everything is done.

theprophet said...

If your life feels like it is lacking the power that you want and the motivation that you need,
sometimes all you have to do is shift your point of view.
nike shox r4
nike tennis shoes
nike discount shoes
nike shox shoes
nike free shoes
nike womens shoes
discount nike shoes