StabilityVsPerformance

ThoughtStorms Wiki

Many dynamic systems have a stability vs. performance TradeOff.

http://globalguerrillas.typepad.com/globalguerrillas/2006/03/journalbigban.html

JohnRobb : If we look at today's global environment we see a relatively high performance system driven by real-time global markets and rapid technological progress. Its performance explains why it is spreading so quickly. However, it is also moderately unstable. In our drive towards higher levels of performance we pursued a path of rampant global interconnectivity that has quickly outpaced our ability to dampen excess. The old dampening functions of borders, distance, government, etc are quickly fading. The result is a system vulnerable to rogue feedback. Even a small amount of it can cause global reverberations. Worse, there are people actively working on ways to introduce this rogue feedback. Iraq is a great demonstration of our inability to dampen excess in the face of active opposition (notice how our goals have drifted from building an allied democracy to stopping civil war).

The long-term solution is to build more stability into the system. The best approach I can think of is a highly interconnected but fundamentally decentralized system (most of the benefits of interconnectivity but with lots of local control). Unfortunately, we are far from realizing that goal, since our current view of the world is based on old models.

...

This conclusion also calls into question the efficacy of the idea that merely increasing connectivity is an answer to our problems. Increasing connectivity too fast, in a system without intrinsic dampening or control systems that work, will only accelerate the chaos (human nature doesn't change as fast as technology). If you need proof of that, spend some time reading Jihadi Web sites and pondering the rapid growth of transnational crime. Also, the complexity of this system puts the lie to the idea that we know how to actively dampen its behavior through centralized systems of control. We neither have the scale nor the CollectiveIntelligence to pull it off. The only real solution rests on redesigning the system itself, to enable it to become more tolerant of rogue feedback.