It surprised me how similar the way I think about this subject to most guys here (now, who said humans were more random and less predictable ? At this complicated topic, I see myself very predictable; in at least near future, I will be thinking along the same lines as, probably, MikeD). (IMPORTANT NOTE: The rest is very long. This was a 3 day weekend, so it seems I really have written way too much for a post. 1-6 is introduction to how I think (i.e. insanely), the conclusion is 6-10, 11 is about ethical stuff) Before we start to think about what AI we want from computers, and if computer conscience is evil, I think we should first consider this as "yet another action taken by humans". So, the following is my human history and model: 1) With a good DNA codebase to start with, the man has achieved enormous complexity and success at obtaining the resources to continue the DNA. 2) The DNA provided a better basis to extend and virtualize certain functions. Better mating, better nursing, and most obviously, the man got to know to make use of tools; to hit harder using a rock rather than his fist, to ride a horse than run. He also similarly used commucation features of all sorts (voice, drawings, etc), but since the DNA provided more variance, neurons(bandwidth), he became much better at it than other contenders (dolphins, apes), thereby pulling more resources from the earth (more/better mating, more guaranteed survival of the offspring). {Actually this better DNA base may be the argument God gave Satan that Adam was a pretty superior being -- and you''ve got to be God to be able to foresee which DNA base is better. I am not a religious person, but I think I can fit the religious rights and wrongs in this model too -- "good" being the good gene that considers the society as a whole instead of the individual, because a peaceful genepool is another precious resource for the individual himself unless he doesn''t mind a chaotic next generation for his own offspring, where he may easily die despite his superiority. Perhaps dinosaurs were evil creatures which killed each other and got extinct when the last victorious idiot T-Rex died after he was enjoying all the resources that dinos care about all by himself-- who knows) 3) (It is getting long doesn''t it) Anyway, we are basically there and we may be considered the winners despite the fact that there are more bacteria; because if the bacteria have a good shot at destroying us, we will (hopefully) know it and take cautions thanks to science. Yet, we are happy at the moment and our only war is with the microorgs that make us sick (literally). We try to preserve others, because they constitute resources (some people consider seeing lions at the zoo a good thing, they say kids should know them, kids like seeing them, so lions will be there, not as a protein resource, but as another types (entertainment?). Also we like pandas) 4) Now I am finally getting to the AI point (You''re still reading). As I stated above, all the man did was extending his DNA without mutating or mating; thanks to his programmable (geez) brain that could store configurations (like a hand holding a hammer) of re-routed nerve-to-mechanic correspondence. Kinda like java interfaces to sensors to parse and command things other than *default*. Naturally, there came a time where man finally transferred fundamental data structures (e.g. numbers) via a paper-pen-eye interface, thereby increasing the brain storage capacity beyond its physical limit and finally, passing information to the next generations without even being alive (lions pass their DNA to their offspring, then teach them while they can, then die-- while you can still learn from Socrates) 5) Man invented computing machinery (long story here), then it was very soon obvious that he wanted to know IF there was a way that he can create interfaces to certain real-life problems, to extend the "general problem solving capability and speed" of his brain. What he wants methinks is the definition of "AI". He doesn''t want (or care if) computers have conscience/ethics/etc. He wants to state the problem and receive assistance (or more) from the computer. 6) But, it became soon very obvious that he had to reverse engineer his brain, to understand the interfaces and data structures (natural language being a mess by itself; maths is much cleaner yet unusable unless you can do the translation-- while doing the translation from natural languages to maths is considered "engineering" (solving them is maths practice) and takes many human years to master), to be able to feed the pIII with something. 7) Unfortunately, he''s not been successful yet. Language processing is tough. All problems cannot be translated, because the language is not static, it is living and problems are interrelated and your wonderful machine will still have to learn stuff that is new; so many similarities (learning, self-modification) to living things normally brought the idea of AL as the most successful implementation of self-improving computational units. 8) This is a good idea. We won''t be able to know "how" the computer solves the problem (as if we know how a calculator calculates), but engineering will make sure that it passes testing (an interesting test pattern would be the famous "spaceship computer going crazy and killing the crew" test). However, nothing in the world is guaranteed to work 100% (planes crash). But we learned to live with it. 9) My guess on how this problem will be solved is this: man will create an intermediate yet powerful language. He will design the process (not the program-- because it will be running for a while) such that it is customizable, flexible, scalable (we may need mcuh more processing power than we first think), and extendible (perhaps we will need a "woman-to-man" translator to solve some other types of problems). Then, meanwhile, he will slowly develop AL that does attack problems (in the beginning) similar to himself. This AL will (in far future) become the problem-solving-engine for the problems translated by the translator. (I initially think that language unit should not be put in the evolution process-- otherwise we will have to negotiate with the computer on the use of language (the *real* computer jargon will rise) ). 10) In years (centuries?), the engine will converge towards proposing what we consider "viable solutions to real life problems". 11) Ethical stuff: It will be fun to watch conversations when computers start suggesting certain people should spend 20 years in jail, or be castrated, or even be killed. I will probably be dead (by natural causes) by then tho. The point is, you never trust something you don''t know. It took a while for man to stop double-checking by hand critical calculations done by computers; but after a while you start trusting them more. We''ll get over it. 12) File size is 6849 bytes by now. I would write more, but I fear being banned from posting ever again. Anyway, this is how I think it will be. Now, I''d like to hear people''s suggestions/constructive criticism/corrections. I also would like to hear if anybody wants to follow these steps (or some other steps). By the way, if you think I am crazy, I already know that, so don''t waste bandwidth telling me just that. K /* Somebody please create a better programming language */

Oh my. That''s the last time i try to read something like this at 2 in the morning.... ga ga ga ga ga ga ga ga............vision...blurring...... *THUNK!* hkajjfdeeeekkkkjkjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj (etc.)

==============================
\\// live long and prosper; \||/ die short and rot.
==============================