Figure source. Just some crazy thought, but suppose organizations could make all of their decisions in realtime, just like a flock of birds does? What kind of information would we need to make realtime decisions? Well, if the goal is just to “survive” (“keep flying”) it would probably be sufficient to have information of the innovation speed and direction of direct “neighbors”. Suppose we could facilitate (supported with modern ICT) a distributed network of decision making information. And each node (decision making “entity”) in the network would get the decision making information relevant for making the most important (“survival”) decisions. So what are then the most important decisions? If we compare to the flock of birds, it’s making sure you’re “local” innovation speed and direction matches that of your direct environment. It’s allmost as if we would apply Ashby’s Law of Requisite Variety here. The law states that “variety absorbs variety, defines the minimum number of states necessary for a controller to control a system of a given number of states.” So here you have it: you need to match your own variety with that of your direct environment. Don’t make it more complex or less complex but make sure the complexity matches. And to reduce the decision making complexity you could try to limit the number of decision making connections because that would make the distributed decision making process more complex. That is because the complexity of the network increases quadratically with the number of connected nodes. But also the value of a network increases proportional to the number of connected nodes (Metcalfe’s Law). So we need large networks to create more value. The large networks could consist of interconnected small “local” networks to reduce the total decision making complexity. So here’s a wrap up: use the distributed topology Paul Baran† designed for the Internet as a reference model for distributed decision making and as a reference model for distributed value creation, combine that with Ashby’s Law of Requisite Variety to reduce the “local” decision making complexity and combine that with Metcalfe’s Law to increase the total value of our network. If we would really do that, I wonder how our would world look like?
Posts tagged ‘flock of birds’
Complexity isn’t a bad thing. If we are willing to accept that it is there, maybe we will be able to embrace it. And start taking it as a given instead of something we should try to conquer. It’s the way we try to distribute complexity where things can go wrong. A very natural tendency we seem to have is to consider a complex case, abstract by leaving out all things we don’t want to see and then renew partially. Look at today’s larger organizations: they are often large, complex social systems. And yet we seem to support them predominantly with swiss-army-knive type architectures because that makes things easier for at least a part of the organization. It’s in fact a brute force method for which you need brute force governance. And it most probably leads to suboptimization which is not bad perse but if we could make it better, why not? So that is where a few simple renewal principles might come to help. A flock of birds needs only two principles to stay airborne: know the distance and speed of your direct neighbors and adapt your own distance and speed to that. The flock principles. Now suppose we tried to handle complex cases in our complex social systems based on these flock principles. And suppose we abstracted out all redundant information, we would already be able to “stay airborne” and still renew locally. Abstracting out would imply leaving out all relationships that are not directly connected (or to compare with the flock: no direct visibility). Then you only need to handle the direct relationships (the ones that have direct visibility) but you may never change the global “flock” principles. So if we would limit the number of direct relationships and trust on eachother that the total still would work optimal, this could possibly help us architecting distributed concepts.