Posts Tagged ‘Complexity’

Driverless Cars and Systemic Vulnerabilities

December 29, 2012 2 comments

drivlscarSometimes it seems that the uncontrolled and non-systemic is more stable than the systemically controlled environments deliberately aimed at stability.

While the systemic model, with its top-down, consequentialist policies put in place specifically to produce stability, often produces stability over the short run, it at least as often produces unimaginably wild instability over the long run.

This is a problem that crosses platforms. A common analogy is the fires in managed forests. While top-down policies designed to systematically control and extinguish small fires prevent problems in the short run, the lack of fires to clean out the dried brush ultimately leads to uncontrollable infernos.

I am more likely to write about an almost identical problem on this blog: the Federal Reserve. While top-down policies designed to systematically control and extinguish market phenomena like inflation and unemployment often allow for stability in the short run, the lack of market responsiveness ultimately leads to catastrophic crashes, runaway inflation, and heavy unemployment.

This is something that I have been thinking on quite a bit recently, and I have come to the conclusion that we have to embrace the chaos. Ultimately, we’re in for a wild ride no matter what, so why delay the reckoning? Let’s deal with the problems in incremental and manageable ways, preferably at the level of the individual. The only alternative seems to be waiting for intractable problems later.

The reason I bring this up now is because Motor Trend magazine recently published an article called “The Beginning of the End of Driving.” Read the following excerpt and try to tell me you remain unworried about the systemic vulnerabilities.

Continental plans to have autonomous assistance available for limited freeway driving and for construction areas by 2015, says senior vice president Ralf Lenninger. It will add low-speed city capability in 2017, followed by two-lane highway and country road driverless car technology about the end of the decade. The company calls this “the car you can’t crash,” and it will meet the company’s goal for a zero-percent accident rate.
Zero percent accident rate? Kind of like the Titanic was unsinkable? The potential for systemic vulnerability to me seems greater than the individual vulnerabilities. If we go to a top-down system of driverless cars, what is to stop a rogue programmer from redirecting all of our vehicles into each other at once?
This is not to say that the technology is not potentially useful, or that it can never work. But when the impulse in government is emphatically interference, I doubt the likelihood of independent, individual driverless cars. Far more likely, we will be shunted into a system of “rational regulation,” which may very well end up shunting all of our cars into a ditch one day.
By all means, let’s let technology surge forward, but let’s ensure that we avoid the impulse toward systematic regulatory systems to go with it. I would hate to see driverless cars with the same track record as the Fed.

Complexity, the “Do Something” Instinct, and the Sandy Hook Shooting

December 27, 2012 1 comment

I have as yet avoided writing about the Sandy Hook shooting, partly because I dislike rushing to rash decisions, and partly because I believe the narrative has focused on unimportant ideology as opposed to the important ideas underlying the situation and the reaction to it. Now, however, with the benefit of some distance, I will weigh in.

What happened was horrible. There is simply no way around that. On the other hand, frantic reactions, calls to political action, and statements claiming that what happened was “obvious” do nothing but exacerbate the problem.

In reality, while it is easy to deconstruct the event with the benefit of hindsight, it is simply impossible to claim some sort of valid foresight. It is not enough to say “I knew that something like this would happen,” because that is a statement without verifiability until the proof is at hand. And when the proof is at hand, such a statement is already useless. If you had any real knowledge, you would have stopped it from happening. If you had no ability to stop it, you had had no real knowledge – simply baseless speculation masquerading as knowledge.

One can analogize this “foresight problem” to other occurrences. For example, if someone gets in a car accident, he might say later on “if only I hadn’t slept in and left 20 minutes later than I usually do, I would never have been in this mess.” This is the functional equivalent of claiming foresight about Sandy Hook. If either were valid, the problematic situation would never have obtained.

Nor is it possible to claim that, despite the lack of specific foresight, there was enough knowledge to have stopped the tragedy via preventive action. This kind of claim usually follows the “if only we had this law…” form, and it is the functional equivalent of the above. This also fails to account for the foresight problem, and it suffers from a debilitating case of confirmation bias.

The foresight problem can be extended through thought experiment. Consider a chess game.* The number of possible chess games that can be played is somewhere in the order of 10^120, or many multiples of the number of atoms in the universe. Once the chess game is, say, 50 moves in, and the chesspieces are arrayed just so, it is relatively easy to reconstruct what happened. Before the game began, however, there was an infinitesimal chance of predicting just such an arrangement. And remember, this is chess: as complicated as its game-tree complexity is, it pales in comparison to the complexity of an hour, even a minute, of human action.

We have no useful foresight, but neither does our hindsight yield much when faced with the multitude of competing and interacting causes and effects that a chessboard simplifies but real life makes manifest. After all, if the “obvious” solutions were actually obvious in hindsight, this would not be a recurring problem. We would have corrected it after the last time.

For example, getting rid of guns does not solve the problem, as countries such as Russia continue to see mass violence despite heavy gun control, while Switzerland’s rate of gun violence is comparatively miniscule despite enormous rates of gun ownership. Britain found that gun violence increased 40% after its gun ban in the late 1990s. Unlike poorly-informed pundits, I will not speculate as to why. And indeed, a majority of the mass shootings that have occurred in the United States have occurred in states, localities, and even schools that are “gun-free” or gun-restricted zones. Real life does not cooperate with our strictures. (QED.)

Nor is lack of guns an obvious problem, as the just-as-reactive NRA has suggested. What earthly purpose would stationing the equivalent of TSA agents (who, by the way, are not armed) at schools be, when the TSA has shown nothing but incompetence? Why would one expect an armed policeman to stop a shooter when armed policemen are already commonly at schools and haven’t helped? Why would a national guardsman or combat veteran be effective? Remember that one of the worst mass shootings in history took place on a military installation in Texas. No shortage of guns on a military base.

What are we to do with ourselves, then, if there are no solutions to the problem? (And, emphatically, there are none.) We do the best we can to defend ourselves, but we also accept that the real world is a dangerous, messy place and circumstances beyond our control may end our lives abruptly.

Callous? I call it realistic.

Indeed, the possibility of being the victim of a mass shooting is quite similar to the possibility of being in a fatal car accident, except for the fact that the latter occurs with startling regularity by comparison. We seem to have no problem processing the idea that, while our lives might end on the road at any moment, we have balanced the risk and accepted that, dangerous as driving may be, we will carry on. We seem to completely lack the capacity to accept the idea that, while our lives may end in a mass shooting at any moment, we are capable of balancing the risks here too. Apparently, when guns are present, rationality goes out the window. This despite the fact that cars can be at least as lethal as guns.

And capable we are of balancing the risks inherent in life. After all, we do it literally all the time. Short of locking down the entire citizenry in individual padded cells, there is simply no way to prevent the next mass shooting. That may be hard to hear. But think of the risks you take every day that are greater. Driving is a good example, but also note that you are about as likely to be killed by lightning than be killed by a mass shooter. Shall we live our lives entirely indoors?

Or can we accept the fact that there is no solution?

Once we have accepted that the risk is intractable, we can begin to accept that calls to action, proposed laws, gun controls, the “do something” instinct – all are useless in the face of a tragedy like this. We can avoid the rush to judgment. We can avoid the poison of baseless ideology. We can begin the healing process. We can resume our lives.

While the complexity of the world denies us the ability to see into the future, we can see that the fruits of the “do something” instinct, be they legal or otherwise, are poisonous to liberty. We are faced with a choice. We can keep our liberties and our dignity intact, and live with a future where mass shootings are possible. Or we could pass laws taking away fundamental individual liberties, making the people servile and impotent, and thereafter live with a future where mass shootings are possible anyway.

These are our only two choices. We should choose wisely.


*I have previously used the chess example to illustrate complexity here.

%d bloggers like this: