Too Complex To Understand – or Control

By | April 29, 2012 | 0 Comments

Any intelligent fool can make things bigger and more complex. It takes a touch of genius − and a lot of courage − to move in the opposite direction.
− Albert Einstein

To make something simple is a thousand times more difficult than to make something complex.
− Mikhail Kalashnikov, inventor of AK-47

The more they overthink the plumbing, the easier it is to stop up the drain.
− Commander Montgomery Scott, Chief Engineer, USS Enterprise

This is tax season, and the very fact that it is a season illustrates my point. The Internal Revenue Code and its accompanying regulations occupy 3.8 million words, over four times longer than the complete works of Shakespeare. One can argue that tax rates are too high, or too low, or just right. But no one can deny that the regulations are too complex for even intelligent, educated people to comprehend.
Question: Why are tax laws so complex? (1) They grew over time, with each special-interest group lobbying Congress to enact its pet provision. (2) Bureaucrats love indecipherable laws, so they can impose their own interpretations and increase their own power. (3) Both of the above. Answer: (3).
The same question arises in regard to ObamaCare, which runs in excess of 2700 pages. And it will require more thousands of pages of regulations, making the whole thing even more incomprehensible. A tax law too complex to fathom is painful. A health-care law too complex to fathom can be lethal.
Unelected, unaccountable, faceless bureaucrats already control a good portion of our money. Now they may control our health and our lives as well. This thought should frighten anyone.
Some people believe in magic. They believe that individual citizens are too stupid and irresponsible to control the health care of themselves and their loved ones, but once an individual becomes a bureaucrat, he suddenly acquires the wisdom to control the health care of 313 million Americans. Really?
Even worse, under ObamaCare, the individual’s requirement to buy health insurance of the “correct” type would be enforced by − surprise! − the Internal Revenue Service. A complex, incomprehensible law affecting our health and very lives will be interpreted and enforced by the same bureaucrats who now do such a superb job interpreting and enforcing complex, incomprehensible tax laws. There, don’t you feel better now?
Needless complexity affects government, but it affects other areas as well. People stand in line for hours to buy the latest iPad or iPhone, but that is their choice. On contrary, some effects of increasing complexity cannot be avoided.
● A Highway Patrol officer was driving his wife, daughter and brother-in-law to the 13-year-old girl’s soccer practice. The car accelerated spontaneously. The officer was trained in high-speed driving and was able to control the Lexus at 120 mph. But technology prevented him from stopping it, and all four were killed in a fiery crash.
Granted, most “runaway” cars are caused by confused drivers who mistake the accelerator for the brake. But a 20-year Highway Patrol veteran who did safety inspections on large trucks? “Blame the victim” goes only so far.
Power brakes depend on vacuum assist, and at wide-open throttle there is very little vacuum. So after a few applications, the brakes reverted to non-assisted and were unable to stop the speeding car. Even worse, the car had the new “on-off” button instead of the old, reliable ignition switch. The driver repeatedly pushed the button, but the engine did not shut off. He did not know that if the car was moving, he had to hold the button down for three seconds − during which the car would travel over 500 feet.
The familiar ignition switch is user-friendly − turn right for on, left for off. A button used for both on and off is a user-unfriendly gizmo. Having to hold it down for three seconds is a lethal complication.
I once had a stuck accelerator. I was able to shut off the ignition switch, coast to a stop, and make temporary repairs. It was easy and instinctive, and did not require reading a lengthy owner’s manual that I probably wouldn’t recall in an emergency, and surely wouldn’t know in a rental car.
Even worse, many new cars now come with no printed manual. You download the manual from the Internet. This is acceptable for a TV set, but not for a car. How do you access the Internet when your car breaks down on the road?
Air bags, antilock brakes, and stability control save lives. But technologic innovations should solve existing problems, not create new ones. Engineers should think like engineers, not like salesmen − and surely not like toy makers.
● The CAT scan is a life-saving innovation. But an excellent hospital exposed over 200 patients to about eight times more radiation than a scan should produce.
The excess radiation remained undetected for 18 months, until a patient reported hair loss. Apparently there was no monitoring of radiation dose. Perhaps the computer software was developed elsewhere − and was never checked. Perhaps the software or hardware malfunctioned − and was never checked.
No one took responsibility for assuring that the equipment was working properly. Everyone trusted the computer.
Misuse of technology is a separate problem. If people choose to talk on cell phones or text while driving cars − or even locomotives − that’s their fault. This problem is the new technology itself, which may be less effective or less safe than older technology.
Take Windows Vista. Can anyone understand 15 gigabytes of code and write it error-free? Vista was advertised as faster and more secure than the earlier Windows XP. In fact, it is slower, and its security is arguable − security updates come every few weeks. Now we have Windows 7, which again is advertised as faster. Tests show it is slightly faster than Vista, but about as fast as XP.
Many “improvements” are merely fixes for problems that were introduced by the prior “improvement.” Fixing your own mistakes may provide a lifetime job, but it does not advance human well-being it just adds still more complexity. Surely this process cannot go on indefinitely. At some point, which will be recognized only in retrospect, complexity will be so great that no one can fix the problems. But what then?
Windows, which most of the world uses, is continually “updated” but only occasionally improved. What does this say about computer programs that control dangerous devices like cars, CAT scanners, airliners, or many aspects of our national defense?
Recall the Air France jet that crashed into the Atlantic, killing all 228 persons aboard. There was speculation that the flight-control computer malfunctioned, as happened before. It now appears that the flight crew misinterpreted the data the computer put out. This was called “pilot error,” but it seems more like a failure of technology to be readily usable by human beings, even trained ones.
Mercedes, one of the finest carmakers, produces models that are rated by owners as average or less for reliability. Even the best engineers and workers cannot construct devices that are needlessly complex but also reliable.
I just had my eyes checked at UCLA Medical Center. The eye doctor had to ask the receptionist to use her computer to print out my prescription, because his computer hadn’t “talked” to the printer for three days. UCLA is a world-class university with as many computer geeks per square mile as any place in earth. If even they can’t get printers to print reliably, can we agree that we have a problem?
We dumb down the SAT, yet we expect people who are less educated to run more complex devices. Instead, the complex devices will run them. We must not let computers control our lives − computers that are too complex for most people to understand, but which act too quickly for anyone to control in an emergency.
Don’t forget “algorithmic investing,” in which computer models supposedly gave objective evaluations of investments. But in fact, they allowed the biases of those who wrote the programs to continue without being questioned − until banks and investment firms went broke, and nearly brought down the whole economy.
This is an illustration of the fact that people tend to trust most what they understand least. Investment bankers knew a lot about business, but they knew little about mathematics or computer models, so they trusted “experts.” – who may neither have been so expert nor have had their best interests at heart. I once met a chief master sergeant in the Air Force who spent his career repairing helicopters. He refused to fly on a civilian helicopter. He knew all the pitfalls and trusted only what he understood and had been checked by people he knew. His long experience had given him wisdom – a word rarely heard today.
Finally, in both senses of the word, we have “end-of-life care plans,” which supposedly allow “terminal” patients to die painlessly, but which actually allow the biases of those who wrote the plans to continue without being questioned − until salvageable patients are killed. No more Hippocratic Oath. No more individual patients cared for by individual doctors. Just rote, mechanical, unfeeling, government-run “algorithmic medicine.”
The human mind can handle only a limited amount of detail in a limited amount of time. We are continually “updating” our laws, our computers, our cars, our phones, and our very lives with increasingly complex new versions. Meanwhile, the pace of life increases.
Soon − if it hasn’t happened already − the lines will cross. Increasing complexity will cross decreasing time to master it. Then cars will accelerate uncontrollably, airliners will crash, our phones will reveal our location to miscellaneous snoopers, and our health care will be managed by computers. That last possibility is self-correcting we won’t have to worry about the pace of life after we’re dead.
We should heed Scotty’s advice. If we continue to overthink the plumbing, the drain will eventually stop up, and we will drown in sewage before we ever get to the stars.
A prior version of this column was posted in 2010. Dr. Stolinsky writes on political and social issues. Contact: You are welcome to publish or post these articles, provided that you cite the author and website.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Social Widgets powered by