Distributed AI

[Big Hairy Jimbo]

I've been thinking about distributed computing of recent, and wondered what might be possible, if someone wrote a distributed AI program. One that reported back to a central server and could learn from each of the host PCs using it. The system could come up with questions and send them out to participating PCs, with the hope that the HOSt user would answer appropriately. After sometime, and many participating users/PCs, the system might become "intelligent". The resulting AI, would truly be a citizen of the world, as its "intelligence" would be formulated by many people from around the world.
The same question could of course be sent to many people, so that the system could determine crap from real, as real info would be more frequent than crap. (I hope, as CRAP would be unique). Anyway, I'd have no idea how to make an AI program like this, but I would particpate wherever I could. So if there are any young programming guru's out there, who havn't been disallusioned by what is possible, and what is not, then get to it. Start writing the first global internet intelligence. Just remember that heavier than air flight was impossible once, and that distributed computing using the internet gives you one hell of an amazing supercomputer to use FOR FREE!!!! (well sort of).

 

[malisha]

You know that would be a great idea, except for the part about it making up questions, unless some initial questions where hardwired into the program, then it could probably build a knowledge base after that but asking a good question is an A.I problem in its self

 

[ImaHamster2]

Dear Big Brain,

Enclosed is a question only a genius can answer. Please email your brilliant response. Could you also forward this letter to ten of your dimmer colleagues so that others can laugh at their answers?

Sincerely,
AIspam

 

[kmguru]

A very good idea under the right ecosystem. Humans did not show up on this planet after the jelly fish. So it will take some time. If I have my way, you will see the beginning with some clever programming in a very short time (may be like 5 years).

Here is the problem. There is a serious need for such a system (a precursor) at certain department. The people at that department are using old methodology to acquire the talents. They have to follow the procedures, you know. I am afraid to tell them that creating dinos (bigger and bigger hardware) is not the answer. It only looks good in movies. I am not sure how to be diplomatic yet effective. There are too many powerful people at the dinosaur tent. I have a very few young whipper snappers in my tent.

Rest assured, the project will be done the old way. When, it does not work, may be my group will fix it. We shall see. It is very difficult to jump ahead of the race.