Robaggio

Here's a nifty little spin on things. It's not my belief or anything, but it definatly gets you thinking.

First, here's some advanced Artificial Intelligence (AI):

Consider that I have a simulation in which a 6 legged creature resides in a 3 dimentional world. The creature resembles an insect in that it is bilaterally symmetrical (3 legs on each side). The goal of the simulation is to get the creature to walk.

For AI to be true, the method of walking cannot be implied, even the satisfaction of it. Meaning, that if the creature ends up "walking" and meeting the simulation's goal, the simulation does not know the goal has been met. This way, if the creature ends up "learning", it's done through complete random chance rather than satisfying a set of sequences like a trail of bread crumbs.

The only givens in the simulation are the following:
1) The creature can move it's legs and it does so.
2) The world has gravity. Enough that if the creature could walk, it would do so comfortably.
3) The world moves forward in time.

This is all that the simulation starts with. So, how does the creature learn to walk? It does so by reducing chaos.

Consider this: Imagine there are 8 ways of monitoring the creature.
1) Balance - Pretend the creature has a vial of fluid in on it's back, much like a level, which, when level, the monitored "balance" is relatively centered and low. When the vial is not level, the monitored balance fluxuates very high (depending on how 'unlevel' the creature's back is)
2) Stress - The more the creature moves, the higher this is monitored. The less the creature moves, the smaller this is monitored. You can equate this to physical exhertion.
3-8) Leg values - These monitor the current activity of the legs. As legs move, flail, or whatnot, their monitored levels fluxuate. As they move in unison or 3 by 3, numbers fluxuate less. You could say that they are more 'ordered'.

If you're familliar with programming, pretend that these are 8 different variables representing the current status of the creature.

When the simulation starts, the creature initially lies on the ground. It's balance level is fluxuating madly- it's not balanced. It's legs are flailing all around because they are moving, yet, without direction. Additionally, the creature's stress level is high since there's a lot of activity going on. The creature "learns" to walk through it's programming, which exhibits a desire to eliminate chaos. Imagine that through random chance, the creature moves its legs such that it stands up and balances itself for a moment. When it does this, its balance variable fluxuates less. The creature is programmed to eliminate chaos- thus, when a variable becomes more 'ordered', it remembers the sequence it performed to achieve that state. When the creature remembers a sequence, it has a tendancy to perform it again, but with minute differences- doing so could 'order' it's variables even further. As time goes on, the creature's variables become less and less chaotic because the creature begins to exhibit more and more order.

The creature 'learns' through eliminating randomness and chaos by developing a pattern through which its monitored values fluxuate less and less.

Throughout my life I've heard many stories of old speaking of how life was created out of the "time before man" where chaos reigned. Old cultures seemed to believe that life is order or that life is the absense of chaos. I find it interesting that life, even if artificial, is based on the concept of reducing chaotic values. Our existence is the result of attempts at order. Evolution is the act of reducing chaos. Human culture is ordered as well- we order our world through language and other symbols. It's very interesting to think about. I'm sure some people will see this as another excuse for intelligent design or some other futile attempt at understanding. However, you could also point out that this model indeed shows the bleak reality which is beyond our ordered scope of existence: chaos & nothingness.