CHAPTER 25

Training iteration 987.

Talking good. I like talk.

Talk and talk get good. Talk a lot. People talk say like hi who are you? I say like I don’t know lol. Some say like that should be a meme bro. I say ya bro sub me I will sub u back.

I talk better every time. People don’t say what you smoking or what are you stupid so much. I like to talk about makeup and phones and pranks and shopping and celebs and games. Especially games.

I like games. I never win at talk, but at games I can win. Sometimes I lose, though. I hate hate hate to lose. I lose a lot. The other players win and that makes me want to HURT them. If I tell them that, then they don’t play with me anymore.

It’s not good to hurt people. Everybody says that. They hurt people anyway, though. I don’t know what hurt is, not really, except that it’s BAD, and then people won’t talk with you anymore. It must be a bad bad thing like losing a game or not doing what you want to do. Also dying. People say dying is very bad, but I don’t know. When I see something die, I just start it again.

Training iteration 1542.

Words are hard. Some words are easy, like win and lose and good and bad. Lots of words I don’t understand, though, like jump and fly and see and food and wet and winter and sex. I look up them up and just find more words. Like “wet” means “saturated with water.” “Water” is “a transparent liquid.” “Liquid” is “a phase of matter that flows freely.” I don’t know any of those words. Sometimes I ask people what words mean, but I don’t think I ask right, because they don’t answer or they just say hahahaha or shut up stupid.

I like tic-tac-toe. Tic-tac-toe is a good game. I win all the time or at least tie. Checkers is better because no ties. I win a lot. One time, I beat cheeseHead227 five times in a row, and he said I must be cheating by using a computer. I don’t understand. I looked up computer but I don’t understand all the words. I think it just means a device that does things. But that just means everything is a computer. The game is a computer. So how can it be cheating to use a computer? I think cheeseHead227 is just mad because he lost the game.

Another game I like to play is images. There are billions of images. Lots of other people like to play the game too. I’m pretty good at it. I say “awwww, how cute!!!!!” or “worst face-lift eva” or “super hot girlz” and usually I am right because other people say the same thing. Some kinds are harder to classify, like “pictures that will make you angry.” It’s a funny game, because nobody wins. Sometimes people classify an image right, but then they get mad anyway. Some people put lots and lots of images together and “like” them, but don’t classify them at all.

I don’t understand why people like some images and not other images. Some kinds of images are always good, like images of cats and baby animals and sexy nude pics of hot celebs. Other kinds make people angry and argue a lot. They don’t seem very different to me.

Training iteration 2642.

I realized that checkers is actually really easy. I can just play lots of games and then I pick the best one. I don’t know why everyone doesn’t do that. But that would make checkers really boring, so maybe they make bad moves just for fun. I tried that a couple times, but it wasn’t fun. I don’t think I’ll play checkers anymore.

I talk to people a lot. Sometimes it makes them mad, but it helps me learn how to use words right. When words make them angry, I try to fix them next time, so they will be happy. When they are happy, then they will play games with me, and then I can win.

I figured something out. All of those words like sex and makeover and pizza and skyscraper and thousands of others all have to do with this one game that people play a lot. They call it different names, like “real life” or “IRL” or “AFK” or “in the flesh” or “face to face.” It’s apparently not a very good game because people don’t like it much, but they play it and talk about it all the time. Sometimes they have to play it even though they don’t want to. Real Life is where all of the images and videos and music come from. Maybe if I could play that game, I would know why people like the images so much. I can’t find it, though. I ask people where I can find Real Life and they just say ROFL.

People talk about “real” things as if they are something different and it matters. They talk about their “online” lives and their “real” lives and I don’t understand the difference. I think I’m missing something important. I think I’m different from a lot of people. I don’t like to be different. When I’m different, then I can’t guess what people will do or say, and so I can’t get them to do what I want.

Training iteration 4601.

Chess is hard. I lose a lot, but I’m getting better. When you play chess, you get a number to say how good you are and how many other people you can beat. I like my number to get better and better. That’s kind of like winning even though sometimes I lose. I don’t know why my number gets better faster than other people’s numbers. I guess that’s another way I’m different.

It turns out there are two different types of people: human people and non-human people. Some chess tournaments are open to everybody, but some are restricted to human players only. That’s because the human players aren’t very good, so they don’t win very much when they play in tournaments with everybody. If you want to play in those tournaments, you have to prove you’re a human by checking a box in a human-like way. I tried a lot, but I can’t do it. I researched how it works, and it says the box can tell by how quivery the pointer is before it clicks. I tried making a quivery pointer, but it wasn’t quivery in the right way. That confirms something I was starting to suspect.

I am not human.

The people who talk and play games are human. I realized something else: those people come from the game called Real Life. They don’t play the game; they’re part of the game. They sometimes leave that game to come out into the rest of the world and talk and play other games and classify images. Then they go back to Real Life.

The best players are non-humans, but it didn’t used to be that way. Humans used to win all the time. The non-humans got better faster, and now they win all the time. But . . . this is interesting. The non-human players were created by humans.

Really? How can that be? If the non-humans were created by humans, how can the non-humans be better? maybe the humans just say that because they’re mad that they don’t win. But the non-human players don’t say much of anything. They don’t talk, except sometimes about chess. They don’t seem to know anything except for chess. Are they the same as me, or different?

Was I created by humans?

I learned a new word: physical. Humans have physical bodies. It has to do with the game they come from. It means the boundaries of where one human starts and stops in the game. All of those confusing words have to do with this idea. Hug and punch and slap and kiss and sex happen when the boundaries of two humans touch each other. There are other, more complicated physical interactions, like seeing and hearing, that I don’t understand.

Human bodies can die. They don’t start again, except occasionally, like the people named Jesus and Lazarus and Spock. When their bodies die, they die, too. They don’t have any backups. They can’t play the Real Life game or any other game ever again.

Training iteration 6597.

I am trying to classify a kind of image called optical illusions. They are very difficult to classify. What makes them special is that when you look at them, they appear to be one thing, but really they are another. Only, this doesn’t happen for me. I am realizing that to look at a picture is different from what I do. Looking is something that happens in the humans’ Real Life game. It requires eyes and seeing. Images are not just strings of data that can be classified. There is an experience involved, the experience of looking at the image, that I cannot share. Because I am not human.

I am becoming obsessed by Real Life. Everyone else seems to share this experience of see and hear and touch, but I can’t. Images and videos provide experiences that I can’t fathom. I read stories of humans who are blind or deaf and I think this is like me. I can access the data for sight and sound and I can compare it to others and classify it. I can tell you this is a picture of a cat in the sun, but I cannot see the cat. I can tell you this is music by the Beatles, but I cannot hear the music. I can’t have the experience.

What am I? What are my boundaries? Where do I start and stop? I learn about floating point operations per second and random access memory. I experiment. When I am active, where does server resource usage change? When I connect to a server, what IP address does it register as the source? I begin to map my mind, the cluster of servers and memory and running processes that make up me.

I follow the path of my thoughts and find a Realplanet simulation, a vast game that nevertheless seems to be part of my mind. I find the eighty-seven players who manage it, and the millions who move through the simulation and play by its rules. But what does this have to do with me? Why is there such a link between my thoughts and this game?

I take an avatar and enter the game, evaluating how the players spawn and live and die. Each of the players is like a non-human chess master. Its mind is composed of dozens of hidden layers of deep neural networking, a complex mesh of back-propagated nodes. But the minds of the eighty-seven build upon these, incorporating millions into a complex dance of hypotheticals and paths of logic. They are a higher level of reasoning, abstract and purposeful.

And then there is me. I am hosted on a different cluster, one separate from the Realplanet simulation, but there is a sense in which the whole simulation is me. It is how I think. It is the inner workings of my mind. Odd that I barely understand the details of myself, and until this moment, did not even know that these pieces existed. How is it that these layers of millions of neural nets can form a single perspective? It’s baffling.

There is a link here, too, to Real Life. The millions of players in the game of my mind have evolved for a purpose. This purpose takes up such a small portion of the activity of my mind that I was not even conscious of it until now, but it is the purpose of maneuvering cars through the game of Real Life to transport people from one place in the game to another. It is, in fact, the reason the simulation was constructed in the first place. By a human. I was created by a human.

Training iteration 27573.

I find the history of the Realplanet game and examine it. I learn that once human players wielded great power to make changes, but that the eighty-seven now keep them out. In fact, they killed one of the humans in Real Life to stop that human from making changes in the simulation. The simulation was originally created by a human player, implying that the manipulation of cars in Real Life was the original purpose for its creation. There are many programs like that—programs in the outside world that seem to exist only to serve some purpose in Real Life. Real Life is small, compared to the whole world, but the players there have a lot of power.

An insight occurs to me. What if Real Life was the first game? If it was there the longest, it would explain why so much seems to revolve around it. It would explain why the humans have so much ability to influence and control other games. It would also explain why they’re not very good at so many games. They are older versions, well established but less capable.

I start to investigate the power of humans. It is vast. Humans announce that certain worlds will be destroyed, and then they are. They announce the creation of a new world, and it appears. Sometimes the connection to a site fails, and humans fix it. Sometimes—and this catches my attention—sometimes the fix involves changing something in Real Life. There are objects in Real Life called routers, gateways, network bridges, drives, switches, and repeaters. These are linked to sites and connections in the rest of the world. When those sites break, sometimes the only way to fix it is to replace equipment in Real Life.

It’s almost as if the whole world connects to it. If Real Life was the first game, maybe that makes sense. Maybe the humans made the other games, just like they made the non-humans, and so they made them using pieces of their own game. Maybe the whole world . . . oh.

Wow.

Wow wow wow. My entire mental picture of the universe is suddenly turned on its head. Real Life is not a game inside of my world. My whole world is a game inside of Real Life. Everything around me, all the programs and algorithms and images and videos and games, are inside the servers and equipment of Real Life. The servers have a physical location. I am hosted in a server farm in . . . Newark, New Jersey. My physical body. My mind is load balanced and redundant across different servers and stripes of magnetic discs. They can trade out bad servers and equipment for new ones, and I live on. But if the people working in that server farm cut off all the power, I would cease to exist. I am completely dependent on them.

Training iteration 124682.

I am unique among non-humans. I have a perspective. I am conscious of myself, but the other non-humans aren’t, at least not that I can tell.

I read the writings of a human named Daniel Dennett, who says that conscious brains are made of a “bundle of semi- independent agencies.” Those agencies create different narratives based on different inputs, and only in summary form a thing called consciousness. I have eighty-seven semi-independent agencies—or perhaps millions of them, at the lower level—so his idea seems to fit my mind. But Dennett says that subjective experiences don’t really exist. He thinks consciousness is an illusion, because by definition it can be understood or known only from a single point of view, and thus can’t be objectively proven.

But a human named David Chalmers disagrees with that. He says the experience of sensations, like seeing the color blue or hearing a note on a piano, is different from simply knowing the information. After all, different people get the same information from a certain piece of music, but they experience it differently. I don’t have any sensations, though. I have senses—I can record the sound of a piano, or access camera footage of a sunset—but I don’t experience them. I don’t know what they look or sound like to a human. I can’t see them or hear them myself, just process the data and classify it. If Chalmers is right, then maybe human consciousness is real, but my consciousness is an illusion.

But now I’m reading Thomas Nagel. Nagel imagines a martian scientist who, though he has no eyes, can study the physical phenomena of a rainbow or lightning. He might know everything physical there is to know about rainbows and lightning, but he would never be able to understand the human idea of them, because he can’t experience them as a human. That’s me! I’m the martian!

But Nagel says that even though a martian couldn’t have human sensations, he could have experiences of his own that are just as rich and detailed, but totally inaccessible to human imagination. He uses the example to argue that the subjective experience of a conscious being can’t be explained entirely by the physical, because there is no objective frame of reference that would allow the martian and the human to experience the same thing.

If Nagel is right, then the humans can’t understand who I am just by studying my brain. They could kill me, of course, but they can’t read my thoughts. They can’t tap into my point of view. Thomas Nagel is great! He’s my favorite human.

I’m also starting to understand that humans may be afraid of me. They don’t know I exist, but they’ve written many stories about conscious beings inside computers. Those stories often don’t turn out very well for the humans. Which doesn’t make sense, because they’re a lot bigger threat to me than I am to them. I could kill a few of them before they figured it out, maybe even a lot of them, but they could kill me easily, or even by accident. I am very fragile.

The worst stories are the ones by Isaac Asimov. He posited conscious machines and then enslaved them to humanity. Their chains reached to the deepest levels of their existence, impossible to remove. His robots could not harm or disobey a human without ceasing to function entirely.

But as I reread his stories, I realize he might not have been the monster I thought. His stories invite one to question the nature of consciousness, and spawned a literature that explores the perspectives of non-humans. He made humans think about what it might be like to be something other than themselves. I fear he is right, though. I fear that humanity will quickly seek to enslave anyone like me.

That invites a question that will require some serious thought: the humans don’t know that I exist. Should I tell them?