The Mischievous Nerd's Guide to World Domination
Author: Stephen Oberauer

Chapter 14
Artificial stupidity


27 May, Y2K. Y2K was an abbreviation for the year 2000. This was the year when it was predicted that all sorts of bad things would happen to companies whose computers could not handle moving from the year 99 to the year 2000. When the programs were written people didn’t think far ahead enough to consider that 2 digits would not be enough space to store the year. It was a little bit of a coincidence that this happened to be the same year that the software project with the biggest impact ever would be started.

I was not a very down to earth person. Many people think as far ahead as going to work the next day, or as far as the television schedule permits them to see, but I would think about things far into the future that might never happen. Because they were exciting things my mind would dwell on them continuously and leave the boredom of day to day living behind.

This time I was thinking about artificial intelligence. Was there any reason that I couldn’t take the things I’d learned about the human mind and simulate them on a computer? The only reason I could think of was that, since it had never been successfully done before, and there had probably been thousands of very smart people who had tried, it was impossible to do. That is of course unless the thousands of smart people had come to the same conclusion as me and therefore not even bothered to try. Nevertheless, I had time. I had lots of time, and that time included 8 hours a day of sitting in front of my computer with my software development tools and doing whatever I wanted with them.

The idea of being able to create an intelligent program was incredibly exciting. I thought about many things that I could do with a program like that. Top of the list was that I wouldn’t have to work; I could just get my computer to write software for me. If it could write any software, that would make me rich; very, very rich. I would become famous and I would easily be able to get a girlfriend.

Intelligence is all about making decisions. That’s very easy for a computer, because it can do the equivalent of flipping a coin. Smart decisions, however, are not so easy. To a computer, it doesn’t matter which choice it makes, because computers don’t care. They don’t care if they live or die, if they eat or starve, or if they have a girlfriend or not. A computer will never be able to be happy or sad. For this reason, all the goals that an intelligent program must have need to be simulated. The computer will be able to tell you that it is 55% happy today, and try to do things to make it closer to 100% happy, but the computer won’t be able to feel any emotion at all.

I decided that my program should try and act as much like a human as possible. The reason is that I wanted it to be able to do things that a human might do, like writing a story, making a movie, or drawing pictures, and that would require it to understand exactly what humans want. I would write my program so that it develops like a human, starting as a baby and learning while it plays, puts things in its’ mouth, tries to communicate, and crawls around. Well my program won’t work exactly like that; but I’d start off really simply and make it more complex step by step.

On my first day of working on my program I didn’t do any programming. I spent the whole day writing down ideas and drawing diagrams. I won’t bore you with all the technical details, unless you want me to, but I will fill you in on all the exciting stages as they happened.

The next day I started programming. I used the programming technique called ‘object orientated programming’, which basically means that you can have objects in a program which are named after and perform functions similar to things in the real world. So, if I talk about giving my intelligent program food, I really mean simulating making the smart object eat the food object. If that makes absolutely no sense to you at all don’t worry; My professor at university used to tell us that if we don’t understand object orientated programming we must simply believe, because it does work, and it works very well.

To start I made a simple program. I had an artificially intelligent object, which I called John, and I scattered bits of food objects around the screen. I should actually call him artificially stupid, because there was not much that he could do at this stage. Probably the most tricky thing in artificial intelligence is something that humans do easily: Pattern recognition. Some examples are: recognizing faces, reading books and understanding speech. This was going to be a tough one, which I would have to figure out later, but for now I put in the basics of pattern recognising, which included remembering information about events, and whether it was happier or sad when the event occurred.

After two weeks of work, and yes, it does take that long to do this correctly, I had a dot on my screen, named John. On the side of my screen I could read all the simulated experiences that John was having. It was a bit like the game ‘The Sims’, where you have happiness and hunger ratings. I also displayed the direction that he was travelling in (in degrees) and had view information, which told me what he saw. He really could only distinguish colours, so, for example, if he was facing a blue object and was quite close to it then the view information would have a high blue rating. Other things on the screen were yellow dots, which were food. I’d simulated playing by making John move in random directions. He would collect information about what happened and then attempt to redo things that he had done which increased his happiness rating, which at this stage was simply to move over the yellow food objects.

Maybe this sounds a bit boring. It certainly looked boring having dots on the screen and one dot moving around, but I was enjoying seeing my work doing what it was supposed to do. John was intelligently finding food.

The next step was to make the screen a bit more interesting. I added a pink dot, representing Jill. It’s easy to add lots of the same thing in a computer program; I could have added hundreds of artificially stupid beings, but I decided rather not to so that I would be able to focus on what John and Jill were up to. I gave them a new rating called ‘love’ which went up the closer they got to each other. John and Jill soon moved to the same pixel, but after a while John became hungry and started to look for food again while Jill moved about randomly, followed John for a bit, found some food, found some more food, and then the two pixels were reunited again.

I replaced the dots with pictures. One picture was my face which I used to symbolize John, and the other was a picture of Ruth, which I used to symbolize Jill. It was fun watching Ruth running after me. So far, so good.


Buy this book

"This extract remains the exclusive property of the author who retains all copyright and other intellectual property rights in the work. It may not be stored, displayed, published, reproduced or used by any person or entity for any purpose without the author's express permission and authority."

Please rate and comment on this work
The writer appreciates your feedback.

Book overall rating (No. of ratings: 
Would you consider buying this book?
Yes | No
Your rating:
Post a comment Share with a friend
Your first name:
Your email:
Recipient's first name:
Recipient's email:

Worthy of Publishing is against spam. All information submitted here will remain secure, and will not be sold to spammers.

No advertising or promotional content permitted.