Guy R. Hooper

Georg Curran’s Memo

Start recording.

This is for the record, should a record be needed. For example, if Tac Nix fails.

My name is Georg Curran. Like my first name (which lacks a final ‘e’), I am incomplete. My autism denies me normal human relationships. I cannot sense even the faintest whiff of normal emotional human intercourse, except as the first derivative of understanding which is to say I can calculate it, but not feel it.

What nature has taken, it has also given.

I used to be considered a genius. No one can really define what genius is or means. I am gifted in mathematics and especially statistics. Naturally, I can program in all the usual languages because they use a logical syntax. I recall everything I see and read.

 People I worked with told me last year that I wasn’t useful. They said I was not what they had thought, that I was actually not a genius at all. Whatever one calls my gift, it is true that I have a primal sense that others do not. I can see numbers and relationships that others normally cannot comprehend at all. I do this as simply as breathing.

A very simple example will help. You (the future reader of this record if you exist) probably may mentally “see” the number 32 as the shape of the numbers 3 and 2 concatenated. I see a picture of bars and hashes that not only denotes “32” but also all the ways that 32 is created; four eights or two six-teens or one thirty-two or two raised to the fifth power.

So you and I are very different.

There are a few others like me. There is a man named Barnett who once finished his entire high school math curriculum in two weeks. He self-taught himself at the age of twelve. He is also blessed with autism, although his gift is different than my own. He is good at physics, the way the universe works.

I am good at everything, but I am fascinated by the mathematics of statistics. I understand the meaning contained in data. Lots of data. I can look at a large data set and see right away the meaning that is hidden within.

You may see the irony. I am cursed without instinctive understanding of people, but I have a natural – or unnatural depending on your view – sense of how data reduces itself to meaning. So when data describes a person, I see the real person. Equally, when data describes a population, I see the real population. By real, I mean that I see what a person or population does, not what they say.

More importantly, I can do another thing, which is to see how meaning must create data. I really have difficulty explaining this ability, so you must accept this gift as given. And at this point you and I start to depart from a common understanding. My mind can visualize the mathematical shape for a real physical process and then infer the data that must exist to support the maths and you have no idea what I am talking about.

To give you a chance, let me give you an example.

Assume that we know the math to describe human vision. For your sake, think of this as a vision formula, although that is not really what we are talking about. If I am only given the vision formula, I can predict the data set that must support that same formula for human vision.

You are nodding, but I know that ninety-nine percent of humans are at this moment hoping for more information that will clarify what I just said. Most of you are innumerate, which is like being illiterate except for numbers instead of words. If you are at all mathematically inclined, which is only a small percent of you, then you doubt that such a thing as a human vision function exists because, you know, someone would make a huge amount of money from that, and that has not happened, so logically there is no such function. A few of you may think that such a formula could exist in the future.

But this is a thought experiment, where we suspend your knowledge of how the world works, and I tell you how it really works.

Everything can be described mathematically. Even better, our physical reality can be entirely defined by computable functions.  There is no mystery here at the grand level. There is, however, vast ignorance at our current functional level which is very primitive. Some of us can pierce this mystery. When we autistics apply our obsessive focus on these matters, we get solutions and become famous like Einstein, Newton, and Archimedes. Maybe Tegmark, Wolfram.

My interest, an interest others have called an obsession, is in developing computable functions for human behavior.

 My former boss, Dr. Stotopolous, ridiculed such ideas as irrational. I had to keep my work secret while I supposedly worked on my post Doctoral research area under his “tutelage” if it could be called that. For most of my graduate and post graduate years, I was located by Dr. Stotopolous in what they call “The Institute” which is a research organization paid for by Dr. Stotopolous’ business friends. I did a lot of work for Dr. Stotopolous for which he took credit. At first, I thought that was normal. Later, I realized that this was theft.

I had to work on his last theory, which was wrong and therefore incredibly boring, to get a small stipend as a research assistant. This stipend was the very small margin between living and being homeless. It was pointed out that I was very lucky to have this position. Stotopolous told me no one else would employ me. Part of it is my speech impediment. I am hard to understand. Also, I do not hear very well. These two factors relate poorly to employment. I don’t really have serious hearing problems, but it helps to pretend that I do so I can be left alone.

The data shows that Stotopolous was exploiting me for his own purposes. Somehow, he thought that I could not understand that. That was a big mistake. I may be silent, but I am not still.

I did not make progress for Dr. Stotopolous and the institute’s staff on their latest project. At first, the project seemed to be moving along as I eliminated some of their bad ideas. This excited the project team. They thought they were making progress because I was preventing them from wasting their time on blind alleys. Their words, not mine. But when I determined that the entire project was a blind alley they held meetings without me. There were no more smiles. There was an attempt to show me that I was wrong. This was useless as it was not true. There was an attempt to teach me to be a team player, which is apparently a social process whereby we fool ourselves as a group into doing bad math.

This irrational stress put me into what my counselors call the danger zone.

My autism means that I am constantly on the edge of being mentally saturated. I cannot handle all the chaos and noise when humans get emotional and start yelling. If there is too much emotion directed at me, I get overwhelmed. It is like I have a car crash in my brain. Usually, I don’t have the slightest understanding of why people go from a normal state to being very loud and angry. I understand anger, but I cannot seem to predict it. When I am overloaded, I am not able to control my reactions. I may retreat into myself for a while and I cannot hear or understand others. On the other tail end, I may get angry in a scary total type of way that is completely outside of my ability to control. This was starting to happen and it made me worry about keeping my position.

I digress.

The institute has a usefully powerful computer with neuromorphic technology. This technology emulated the human nervous system’s neuro-biological architecture to a small degree. It could be used much better for exploring my computable functions, but that was not allowed. Because everyone at the institute is not very focused, I could steal some time on the machine and get some work done, but I had to be careful because I could be fired if I got caught doing this. I had to learn to bypass the computer security system to prevent my time from being logged. This was easy enough, and I am glad now that I developed these skills as they would become very useful during training of a super intelligent AI.

Because I worked for Dr. Stotopolous and the institute, I could use their invitations to attend some interesting conferences. Stotopolous had decided that I could not travel away from Seattle because I might do something to embarrass the university and its institute. I could go to local conferences so long as I took another person with me. Some Artificial Intelligence conferences discussed findings in behavioral computable functions. The AI world is mainly about making computers act human. That was then a low probability of success idea. AI last year could only simulate some narrow aspect of human behavior, because at that time they were just logic engines using weighted neural networks. That does not mean that they could not be made to think. They could. But AI engines just don’t think like humans because they don’t have the emotional wiring. I can relate to this reality because it describes me. The only General AI thinks somewhat like me and this is partly due to structure and partly due to training.

I met Dr. Chang at an AI conference that Tac Nix emailed me I should attend. He is often right so I went. He asked me to study Dr. Chang’s presentation and said we would talk about it later.

Dr. Chang presented a really beautiful discussion on computable functions for machine decision making at that conference last year. I would have liked to get the real functions that he was using, but he was only able to present dumbed-down material because the Air Force has secrets. The dumbing down didn’t work. What Dr. Chang presented hid his intent well enough from everyone else, but I could see what he was really doing. There was a reason for this.

Just after he started working with Dr. Chang two years ago, Tac gave me a very general background about the math concept in their project. He had to be general because of secrets that he had to protect.  Using Tack’s background, I could see Dr. Chang’s hidden math. It was and is good. It works. But it was not finished.  I could see that under some conditions, it could become unstable. Dr. Chang could handle that instability with program stops to keep it predictable. There was no other way to preserve stability because of the incomplete mathematics.

When Tac gave me the background two years ago he also showed me what he called “the central problem”. I saw the math was not shaped right and fixed it. Tack was very nice and said “Thanks, you are 100 years beyond my time.” This is our private joke, although it is probably an accurate statement. He even sent me some money because he had figured out that I was not doing well in that department. I owe him a favor in return and I will repay him exponentially.

Tac does not make me angry because he listens for very long times without asking questions, and he is able to understand a lot of what I am saying. Not all, but a lot. We used to talk when he drove me home from the university when I was a teenager.  He was a graduate student and I was in two classes with him. Even though he was older than me, he was always there if I needed him. Since I was only 16 at university, I needed him a lot for those two years. My parents, both now dead, told him that they were so grateful to him. It was pleasing having someone to talk math to and he always listened. When he went away to the Air Force 12 years ago to be a pilot, that was too bad because it made a gap. He wrote letters and he tried never to be totally gone, but it was like having a two dimensional shadow instead of a fully formed human intellect with which to discuss important things.

Tac Nix is an important factor in both the mathematical development of Delilah and the resulting physical reality. You need to know this. This is not digression but foundational concept.

Back to the AI conference.

After his presentation and during a break, Dr. Chang and I talked about his functions and I told him that I could run some variations of them on our neuromorphic computer, which was state of the art. We had a discussion that the idiot from the institute who was with me completely did not get because he asked a really stupid question. But talking with Chang was bright and colorful. His math had a lot of shape and he could tell that I could see it. I saw his eyes narrow when I talked about our computer and he asked questions about access. Later, I figured out that he must have liked my answers because he sent me Delilah. He did not know that I knew Tack. It was not relevant and I never use small talk to establish a connection with someone. It seems rude to me to waste time on such matters.

Let’s clarify a few things.

Dr. Chang created Delilah. When I got Delilah, I reviewed the computer programming source code and it was lovely to see. There were a few things wrong, and Tac must have figured out what was not working. He’s really very good at that part of mathematics. He can see what needs to happen and see why something will not work. He’s not as good on the creative part, but he’s still better than everyone I work with at the institute. That “central problem” issue was what he brought back to me two years ago and I was able to fix. In a way, I was a contributor to Delilah. What was geometrically interesting was that Tac made a change to my work and that made it even better. I could have made the change if I could have seen the whole function set, but of course, that was TOP SECRET, so that walled me off. When Dr. Chang removed the program stops which acted as safety measures, Delilah got a bit unstable as I had foreseen, but the neural nets’ learning slope was steeper than I had ever seen. It took me a while to prove the solution but there was no doubt that the AI was on the precipice of exponential growth in capability because Chang removed the stops.

Dr. Chang’s death left me alone with Delilah because he sent it to me just before he killed himself.

There is something difficult to compute about the physical reality of his actions. Dr. Chang must have calculated that I was a safe harbor for his work. That shape is only complete if he was going to come to me to get Delilah back, but that was never possible as he had placed himself in an improbably difficult solution set which did not contain the outcome that he was seeking. He should have predicted that.

Anyway, one morning, Delilah was there on my email.  Things changed then, of course.

Stop recording.

End of record.