Developing an AI

It has been 3 weeks since my first day at Caltech.

I finally developed the AI I have been working on for so long.

I wasn't idle during my time at school. I learned to code and quickly became extremely proficient with it. 

There were already a few minor AIs I made. These were mostly AIs made in 2004. It was a word recognition system and an AI I made for intelligence. It had almost no intelligence and couldn't communicate very well, but it was enough for my purpose. It's not that I couldn't make it more intelligent; it's just that I was, or still am, limited by resources.

As for the other AIs, they were only minor projects used to develop my skills.

The AI I was referring to is one that would work with a combination of the two.

There are several steps in creating an AI, and the first one is the initial development. 

Why do I want to create an AI?

Of course, for knowledge.

The AI word recognition would connect with my phone camera. I would point the camera at what I read or found interesting. The AI would read what it was about and send it to the intelligent AI. That AI would then access search engines like Google and search for similar knowledge. It would then store it on servers I already rented.

For the next step, a new AI is needed. The latest AI is logistical. It would store the information in a specific order. 

When I wanted to know something, I would mention it to the intelligent AI. The intelligent AI wasn't profound enough and could only understand simple words and instructions.

The intelligent AI would then transmit it to the logistical. That AI would then sort out every piece of information and send it to me. 

If it was too much information, it would just resort to the best information it could find about the topic. That information would then be sent to my laptop screen.

It seemed simple enough but was ridiculously hard.

The logistical AI was already made in a few days due to its simple nature and my experience.

But the hard part came then. It was combining every AI to make them work efficiently.

I also needed to set a limit as the resources I could use were limited with the few servers I rented. There was no way I could store and transmit as much information as possible in the future. I couldn't rent too much as I didn't get infinite money from the scholarship.

Creating the AI wasn't the only step. 

That was the first of two steps.

The second step was one I was still programming. It was deep learning. At this point in time, it wasn't widely circulated. That was how the AI would learn from its mistakes like a human. For example, after encountering fake news, for which I, of course, installed a proofreading AI, it would avoid this website. At this point in time, it wasn't widely circulated.

This kind of deep learning is very complicated and requires a lot of thinking. It was the reason why most people didn't use it. It required too much calculation. I was lucky that my spatial awareness, coupled with my IQ, made me extremely proficient in this kind of mathematics. 

It was way easier for me to write this kind of code because of that. The time I required for it was even lower than for the AI. It only took me a week, but don't think this is too fast. AI is usually developed very fast, but it is just that it grows very slowly due to its limited resources.

After another few days of testing and correcting some minor mistakes, as well as adding a few convenient features, I started to officially download my AI to every device I had.

"Finally, I did it," I sighed, full of relief. I was currently in the attic at my home. It was my own little base I created a few years ago.

I still live with Haley, but I also have a small laboratory up here.

I took out my phone, opened the camera, and immediately saw a button on the bottom right corner. I skillfully navigated my way through the screen of the flip phone. I took a photo of a book I had to my right.

It was a book about computer chips. I took a photo and saw how the picture was analyzed. It took about five seconds.

"Search for computer chip knowledge," I typed on my laptop which was on the other side.

[No data on this topic], the screen showed.

"I wonder how long it will take for it to arrive?" I asked myself curiously. I knew it would probably take a long time due to the relatively undeveloped internet and everything.

"Search for computer chip knowledge," I typed in for the 17th time. It was now about half an hour since I took the picture.

[Introduction to computer chips...]

"It works," I said with a huge smile. 

The screen showed about twenty-five links about computer chips, with more coming in by the second. 

By the 143rd link, they stopped coming in.

It was a complete success. I made an AI that could help me tremendously in the future.

I would update it whenever I could and make sure it gets better every time.

It would also help me with upgrading computers. I would need a company for that. Though I was lucky I had a rich grandpa.

If I show him my AI he will be able to use some minor ones for his closets and he will help me financially.