Saturday 7 August 2010

Solitude

This morning I finally received the confirmation from the UMCG hospital regarding the urology examination. It will take place on September the 27th, and according to the included brochures it will at least involve an ultrasound of the prostate and a check-up of the urinal tract and bladder. If there are any other tests they'll want to perform I'll probably be unable to learn about it, seeing as how I only seem to be able to get any real information when I'm there in person. I guess I'll just have to buckle up and ride things out until the 27th next month.

As for my general condition at this moment, I must admit that I have since a few days ago withdrawn from interaction with everyone but a few people I know. I just got so sick of other people a few days ago and this attitude hasn't changed yet. The truth is that I don't think that most people are worth a second of my time, and that my dreams of living a more regular life among other people is just a silly thing, if not stupid.

Until my next appointment at the UMCG and assuming a favourable and useful result I don't expect this to change or improve. I'm absolutely sick of learning things about people I have recently met and discover that I was sorely mistaken about them. I have begun to loathe talking to virtually everyone, as they're all so... I don't know... simplistic? Their desires in lives are shallow, when talking to them it's about as interesting as watching paint dry. It's like I'm talking to the dead or something, as though they're echoes of something without any future left open to them. I'm not really sure how to explain this.

What I desire most is to create my own future, to bend the 'facts' about what is possible and what isn't with my own powers and skills. One of these is people proclaiming that the human brain is something mystical, uttering something about quantum theory and other big words they actually don't really understand, or call in fantasy stuff such as souls. I'm utterly convinced that human-level or better artificial intelligence isn't only possible right now but actually really easy. It's easier than raising a human child for 12+ bloody years, that's for sure.

For our current game I have been studying up on AI game development theory and found many matches between my own theories as well as many disagreements. It is a very fascinating subject for sure, and I must admit that I absolutely love a fairly abstract challenge like AI, in the sense that it totally deconstructs everything which we take for granted. It allows one to really grasp what consciousness truly entails and how humans in general are basically no more than reactive systems, foregoing planning for pre-programmed responses.

Reading up on this has rekindled my desire to finish my own AI project, called Khipu after the Aztec word for the device they used to store information with involving knotted ropes. While it is still a fairly involved project, it nevertheless thrills me to my core to be working on something which is truly groundbreaking and fundamental research. This is the stuff I live for, the true purpose of my existence. I want to take everything that is considered difficult or even impossible and just make it work. Because it is who and what I am as an existence.

Sadly being in possession of such a desire seems to put me in a minority within humanity as a whole, as I imagine visionaries have always found themselves caught in isolation.


Maya

3 comments:

Dirk said...

I think that in game theory it's better to talk about algorithms instead of AI. Game algorithms can't think outside the box, in contrary to humans (in general).

Making something that is truly self learning is quite hard and as long it's easier just to program an algorithm, the later option will be used. Using an algorithm also gives more certainty of what your system will do. Also intelligence mostly means copying behavior, this means your system can pick up bad habits.

A more philosophical question is why should we ever want to model a human brain? Except for the exercise and the ever present "because we can", there is no reason to do it. Humans and computers have their advantages and disadvantages. An algorithm that plays chess is much better than an algorithm that plays chess trying to think like a human.

Maya Posch said...

Well, an algorithm is quite a generic term. What you call an algorithm I would call 'scripted AI', being a form of AI in which every possible response has been precoded. This is also used in decision trees and similar.

One necessity for actual learning is a proper memory, something which can store that which is relevant and forget that which isn't, and in such a way that it is easily retrieved by the other required units.

What self-conscious minds have is something I call a self-reflective component; the ability to consider one's own actions, in the past, present and the future, combined with a predictive component to run through other possible scenarios.

In my maybe not so humble opinion this is all that is required for a self-conscious AI, and it really does not seem that difficult to implement :)

Dirk said...

Software always looks easy to implement :).

While not being up to date with current research on AI, I keep being skeptical about claims on copying the human brain. And don't forget the unwritten rule in IT research: "if it doesn't compile, it doesn't exist". (It is possible that this rule only applies to research on computer languages and compilers ;) )