Slow time to a halt, and look at the particles of a dust cloud.

First of all, the particles are rock solid and cannot be moved, because time doesnt flow, hence no movement.

Then think about Heisenbergs uncertainty, what does the particles REALLY look like?

Is this state what the scientist call Bose-Einstein condensate, or is it just a smeared blur of undefined radiation?

Even if it would be plain radiation frozen in time, it would be absolutely predictable, if you knew the exact volume of space it fills and its exact energy at every point in this space.

In theory, this is impossible, but is it really?
Isnt the theory or the ways of interpreting it plain inadequate and in the end actually predictable?

Even so, the extremely small sizes of the systems makes any attempts of calculation/prediction almost absurd. Therefore, we can generalize into a ''Heisenberg'' theory where we use ''randomness'' as a means of measurement of complex systems.

It works, but its not perfectly predictable, just like a blurred image isnt sharp enough when we zoom it.

Randomness is just a way of measuring complex systems.

Einstein never believed in randomness.
He always said: "Gott würfelt nicht.", which translated means: "God doesn''t role dice."
However, modern theories can''t be calculated any longer. Stephen Hawking even went as far and said: "Even God cannot know the exact position AND the exact speed."
It is true, random numbers appear.
When rolling a die, there are factors the change the output of the number, like "how fast did the hand move?", "how rough is the table?" and everything, but in the world of physics, there are completely random things. If you would rewind time, the same thing might happen again, but the probablity that something else entirely happens is just as great.

It is impossible to create randomness in our current enviornment. It is entirely possible however to create close to random. The only real question that applies to game programing is how close to random does it need to be. The idea of close enough is demonstrated in simple math. For example if a middle school kid is working with a circle, and he sees the symbol <pi> chances are he will substitute that symbol with 3.14 . For that kid in middle school 3.14 is close enough to the actual value for <pi> that he has no need to continue on with 3.14159... and so on. If you don''t agree with my very first statement about the creation of actual randomness, that''s all well and good. Simply put, it doesn''t matter if you can create a random enviornment. All you really need to create is an almost random enviornment. Adjusting how close to random does well to add to the development of a game. A computer controlled oponent''s skill level is based on the predictability of his movements and the possibility to counter.

I have a point to add to this, but I have to go to school. Basically, "A Brief History of Time" by Steven Hawkings alludes to some experiments that relate output to input, and input to output. If I have time later, I''ll clarify.

I have access to the perfect random machine. It is impossible to predict its output. Many have tried and failed.

She''ll be home from work shortly in fact..

;0)





Stimulate

I''ve been trying to train my random machine to act as a dishwasher too. =)


Best Regards,
Jeromy "Maverick" Walsh
------------------------
"The question isn''t how far, the question is do you possess the constitution, the depth of faith, to go as far as is needed?" -Boondock Saints

What is with all this talk of randomness in the human brain. The brain is an EXTREMELY complex thing. It has uncountable variables to calculate. You seem to be concluding, because you cant grasp the complexity, that the brain is random.

Humans seem to make quick assumptions in the state of misunderstanding. If you want to make AI using randomness: fine. But the human brain isn''t necessarily random.

The title of this thread is "True AI". That is something which could possibly be like a human brain or possibly not like a human brain. One thing is for certain though: discussing randomness is of no value to furthering one''s understanding of what a true AI is comprised of.

If I had all of the mechanisms for a true AI, then adding randomness would not increase its value at all. Conversely, if I did not have the complete mechanisms necessary for a true AI, adding randomness would not get it any closer to being a true AI.

Pay attention to that last paragraph I just wrote.

I too, had this idea of a self-programming program, but after reading the first few chapters of AI: Modern Approach, I abandoned the idea. There are simply much more effective ways to do the same thing, and much easier as well. the reason we don''t have AI as smart as humans is because we don''t have hardware powerful enough to handle human-like AI. The concept of having a program that can reprogram itself, even with the most advanced optimization tools and fastest hardware, would defeat its own usefulness in a matter of days. It would simply not be able to run itself. And so, the AI community has simply settled for a mediocre performance. As soon as the hardware is available to handle extremely advanced AI processes, human-like AI will be commonplace. Until then, I''m afraid there''s no immediate and easy solution.

Scientists first wondered how human sized things worked. A few thousand years later, they figured out it had to do with smaller things called atoms. But how do atoms work? Well, a few hudred years later, we find subatomic particles are the driving force of atoms. But how do they work? Say we find an even smaller force that dictates how particles work. But then, how does that smaller force work? Reminds me of a fractal image. The closer you look at it, the more you realize you''re nowhere near understanding how things ACTUALLY work =-P