Messing with complex systems
Posted by aogTuesday, 30 September 2003 at 19:42 TrackBack Ping URL

Thinking about some cognitive dissonance, I was pondering global warming, transhuman genetic engineering and the welfare state. The common theme in all of these is deliberate manipulation of complex systems we don’t understand (and which may not be understandable at all non-statistically). I tend to be more concerned about this kind of thing because as a systems engineer I tend to be the guy who cleans up other people’s messes when they fiddle with things they don’t understand. It’s frequently quite astonishing how software systems one has designed and built from the ground up still behave in surprising ways. And as clever as I am, my systems are still orders of magnitude simpler than planetary weather, the human genome or modern society.

The cognitive dissonance comes in when various factions freak out over one of these issues but are gung ho on a different one. For instance, most of the people who are deeply concerned about global warming and how we must do nothing to disturb planetary weather because we don’t understand it support massive changes in the structure of society despite the fact that our knowledge of cause and effect in the latter is slightly better than the ancient Egyptians grasp of atomic theory. Or how the people who believe in the harmony and diversity of nature are gung ho for a mechanistic monocultural economy, the latter being far more similar to an ecosystem than a factory.

To hit the other side as well I’m not impressed with those who are aghast at societal engineering but blasé about human genetic engineering. Recent work is indicating that the human genome is the result of three billion years of random hacking, a stunningly complex self-modifying spaghetti code on a not always reliable executor. The practical effect of this is similar to very complex software systems. While it is frequently easy to fix a wrong behavior, it’s much harder to create new behaviour. To a large extent it’s because bugs generally require just a small change to do something that is “normal” in some sense. Shifting the pattern of action is a lot trickier. It now seems unlikely that there’s a “intelligence” gene that can be tweaked independently. Kind of like the complexity of society… While I don’t have the visceral concern about genetic manipulation that some do I do think of it the same why I think of social engineering - something to be done cautiously, carefully and with the expectation of unintended consquences. Fixing little things (especially if we have working models to copy) is far less worrisome than large “improvements”. I can only hope we start colonizing multiple planets before we engage in significant genome modifications.