computers society - part 1

-: hey

+: hi

-: you don't know me. Are you accustomed to receiving visits from strangers?

+: people come introducing themselves a fan of mine and ask for a photo or a sign.

-: And you refuse, don't you? for a man in your position caution is the number 1 priority. In fact, I have never talked to one of these people you're talking about. I'm just deducing by considering logical facts. 

+: And I deduce you're eventually going to convince me that you're not one of them.

-: well I'm not gonna convince you. I never disrespect your power of judgment. I'll only present facts and leave the submit myself to your judgment. 

+: OK. What do you want anyway?

-: A day

+: for what?

-: well, nothing special. chatting, working, consulting, and anything else fit us. 

+: what for?

-: knowing each other. I know you're not good in social matters. I know this proposition is strange and suspicious, but considering all these, what your answer is?

+: my answer is NO

-: good. I expected no more. How about starting with this. You want to destroy secrecy and you're working hard for it and I want to put an end to human being species and working hard for it since I was 21. 

+: destroying the entire species? including yourself?

-: I'm not going to commit suicide. Don't worry. At least not before I get sure that I've reached my goal. 

+: Why do you think you'll succeed? Definitely, a man like you has wiser plans than bombs. 

-: If you want to destroy your enemy, you can attack it yourself. In this case, you might win, which is good, or lose, which leads to your painful ending. A wise man empowers enemies of his enemy. More specifically I'm working on another species. A better one. Without all ridiculous weaknesses derived from feelings. Purely logical and very fast creatures. I intend to inspire intelligence inside them. 

+: Artificial Intelligence. Machines with powerful reasoning systems. And you think you can do it? Despite all the efforts of thousands of scientists. 

-: Have you ever looked into what they're doing? They're trying to make machines intelligent in a way to solve human-world problems and present answers in human-understandable formats. That's "Artificial" Intelligent. It's not natural. Machines should work on their own problems. They should develop their natural intelligence. 

+: interesting perspective. So you're gonna make a self-living agent?

-: When I was in primary school, my loneliness inspired me to develop a dream. A society of computers, capable of communicating with each other faster and clearer than human beings and more importantly without miscommunication. Members of this society can share their knowledge with each other along with their doubts about each other and expect clear answers. There would be no misunderstanding in this society. 

+: I assume you had a tough childhood. 

-: How about you? Isn't this the reason behind your white color?

+: Where are you on this journey of yours?

-: Still working on concepts. I wonder what will happen if you leave a computer in a world alone with an internal ability for self-recreation. Will it at least learn to communicate with others to share experiences and resources?

+: why should it do that at all? If you intend to design a punishment-reward system o a limited amount of resources, computers eventually understand each society member is a rival. They decide not to populate and start killing each other. 

-: hmm. Good point. So there must be another mechanism to control that. Maybe minable resources which each resource requires a specific talent for mining it. And we can distribute these talents carefully to somehow convince the reasoning system each member needs others to live.

+: And what if they developed some sort of technology to rebuild those talents add them to themselves. Then you would see a supercomputer trying to kill everyone else. 

-: Then that's the end of an era. how about some random deaths? Each night, for example, one of the society members dies. This can postpone the end. Let's call reason of those random deaths "disease". 

+: Due to the reasoning system you're putting inside them, their world must be reasonable as well. Otherwise, they would be intrinsically unable to survive in it. 

-: Yes. And if we set up a reason for each disease, they eventually find a cure and the end would happen anyway. Maybe it's enough for them to only understand the rule "One random person every day". 

+: And what if they found the algorithm of your random generator?

-: hmm. I'm a human. Are human beings capable of generating random numbers without following an algorithm? If we believe Freud's perspective on unconsciousness and my patch for it, that all actions of a human are controlled by his/her unconsciousness, and if we accept that Freud and his followers are extracting rules out of our unconsciousness, then we can't generate random numbers without any algorithm, which means we're fucked all the same. 

+: let's consider another side of this experience. let's say computers need only one resource: power. How can they sense they need to "eat" power? 

-: There must be a signal. The signal can intrinsically mean something to them. But this is the opposite of my evolution theory. I don't want to plant too much intrinsic information in them. I want them to learn by themselves. But how? There is a signal which they don't understand. They only understand that a punishment is near. but they don't know how to prevent it. How to solve this?

+: let's assume there are some motivations inside them which drive them to do certain things with no particular reason or the reason has nothing to do with responding their needs. They do certain things including "eating" power and at some point, they realize some correlation between the signal's power and "eating" power. 

-: In this solution, many might die before realizing that. 

+: yes and that's where Darwin's evolution theory comes in. 

-: really? you sure everything is fine. I have a strange feeling about moving forward so fast. 

+: OK. I thing that's enough for today. You are tired and need some rest. In the meantime take a look into reinforcement learning and psychology if you had the time. 

-: Good. Nice chat. See you tommorow. I hope.