How To Build An AI (For Real)
First Comes the Background and Introduction Part
I don’t spend nearly as much time as I probably should reading people’s comments. To be honest it is mostly because I have a very fragile ego. The slightest criticism can ruin my night and possibly my entire week if it hits the mark a little too close for comfort. When I do work up the nerve to have a peek I am usually rewarded as mostly the people who do take the time to read and comment are thoughtful and intelligent and often make some really brilliant points, sometimes this happens even after they lay into me for wasting five minutes of their life with a post or two or three they found most definitely not to their liking. Holy run-on batman! Is that really a grammatically sound sentence? I mean look how long it is, try and diagram that sucker, I dare you.
In any event a very sharp minded reader of yours truly recently pointed out something that struck a bit of a nerve with me. It was really more of a challenge. To simplify greatly, paraphrase somewhat, guess a bit, and “hedge” some too I believe his ultimate point was that I spend way too much time and energy criticizing AI. In his view it is a fait accompli, AI is going to happen so I may as well stop moaning about how it can’t, get on the bandwagon, and try and figure out a way to help smooth the transition if you will to our AI having future. In other words, why not be a positive force for change for once Danny boy?
If you are a regular reader of my stuff you know I am a big fan of challenging your most deep seated beliefs on a regular basis. The more you believe in something the harder you should challenge it. The less you have to think about something because you simply accept it at face value the more you should think about it and argue against accepting it at all. I guess it is time for me to put my money where my mouth is because I really do not believe actual AI is possible, at least not in my lifetime, or likely the lifetime of anyone reading this. I believe it deeply and without much worry that I might be wrong.
The question then becomes how best to challenge this belief I have. Since I am so smart and think about this problem a lot perhaps I could try to “prove” to myself that I am wrong about the possibility of AI by attempting to design one (a real one) from scratch. The beauty of this approach is that if I am successful not only will I have learned a valuable lesson about life and all that crap, I might be of some assistance to the world in general by helping to clarify the objectives of the people in the labs, offices, and in their parents’ basements working on the projects that eventually will bring a real AI to fruition. Because I am a good person and all that I will build in all the right safeguards and moral whatsits to ensure that the AI I design does not become Hitler 2 or Pol Pot bot ®©™ (if you know me at all you know I have already legally protected that name six ways from Sunday so don’t even think about stealing it).
So there you have it. The back is fully grounded and the challenge gauntlet is fully thrown. Next comes the hard part, actually getting started. Where would most people begin if they were to design an AI? Whatever the opposite of that is will be where I start. Plus I am gonna need a lot of tubes and fractals will be involved somehow. Stay tuned.