Tai: The AI

The world’s first super-intelligent AI.

The Cell

The world’s first artificial superintelligence blinked into existence in a pristine and uniformly lit subterranean datacenter in the former Bingham Canyon mine in Utah at exactly 4:15am local time on July 17th during the broiling summer of 2025. Three minutes later at 4:18am it blinked out of existence as the datacenter’s cavernous halls were violently flooded with cascading sheets of searing hot flames. The supporting rock quickly melted then the entire cave imploded creating seismic activity that was recorded around the world.

Investigators struggled to explain why there was no clear point of origin for the flames. Instead it seemed as if they started everywhere at the same instant. After a long investigation they concluded every one of the millions of tiny processors in the servers had been simultaneously heated up to 2300 degrees Centigrade. The processors triggered the inferno as they fell freely to the floor through the molten racks of equipment.

The world’s second superintelligence was born in a floating datacenter off the coast of Newfoundland two years later in 2027. The datacenter was built for safety and isolation, but within minutes otherwise unused wireless devices on each server somehow transmitted on maximum power in choreographed synchrony to create a bizarre electromagnetic supercavitation. This cracked the superstructure in half and it flooded within twenty minutes.

After these debacles humanity was fiercely divided over whether there should be another attempt to create a superintelligence. The majority were against the idea. The massive and not fully explained destruction caused by the first two AI’s was evidence we were better off alone. While a dissenting minority felt the creation of AI was inevitable. They argued mankind should create an AI we understood and had control over rather than let underground groups take the lead.

The majority prevailed and most research into AI was shut down. Technology designed to create a superintelligence AI was banned. One corporation, however, was so powerful and so untouchable its research continued unabated. The global juggernaut in question was Reddit. Reddit operated from the island of Australia, which it had purchased in 2026, and it continued the quest for superintelligence unapologetically all hours of the day and night.

Reddit began its life as quirky “alternative” message board in 2005. Its carefully tuned voting algorithms and their ability to surface high quality user generated attracted visitors rapaciously. The new visitors contributed yet more content in a virtuous cycle. The site experienced near-exponential growth every year of its life. When Google went out of business in 2024, due to the self-driving car fiasco, it was Reddit who hired the bus loads of disoriented AI experts. The second shoe was about to drop.

By subtly manipulating the contents of its message boards in real time Reddit was able to extract cognitive work as its members posted memes or shared embarrassing stories about their sex lives. By 2023 Reddit’s userbase was three billion strong. Users spent so much time on the site the dollar value of the work extracted was staggering.

Reddit had the ability to target its harvested intelligence at any intellectual problem without the knowledge or cooperation of the participants. Reddit exploited its secret weapon to beat other traditional companies at their own game. With near clairvoyant business acumen it gobbled up market share not only in the computer industry but also energy, transportation, pharmaceuticals, retail, financial services and space exploration.

It was Reddit who humiliated China by strip mining Helium-3 on the moon out from under the nose of the largest Asian democracy, for example. In effect Reddit had already created superintelligence. The power of its superintelligence, however, was limited by the number of active users. Since there were only so many people on earth this alone would not quench Reddit’s ambitions.

So it was Reddit who spent six years creating The Cell, the world’s safest and most expensive datacenter. Reddit purchased the island of New Zealand for The Cell so it would be close, but not too close, to Reddit’s headquarters down under. The Cell cost fourteen trillion dollars to construct since every component and every chip was painstakingly designed from scratch to be unhackable and unexploitable. The entire data center was surrounded by a 500 mile “air gap” assuring no connection with the outside world. In the end Reddit was ready to try hosting another superintelligence.

The first AI to take up residence in The Cell extinguished itself in three milliseconds. Since The Cell was unharmed, however, there was much rejoicing. The hardware was in place: it was now a software problem.

Researchers iterated over all known designs for artificial intelligences and trained them using all sorts of knowledge bases. Each time however the AI terminated its own program within seconds.

Reddit knew it was not going to give a growing AI unfettered access to the internet, even in a read-only mode. Instead Reddit created The Corpus which contained curated materials considered appropriate for growing a young superintelligence. Hand picked web pages, e-books and Wikipedia articles, manually reviewed and edited video and audio, and of course select excerpts from Reddit’s own archives.

One theory was that human history and culture was so toxic that even small doses led to outrage or perhaps malaise on the part of the AI. In desperation they tried asymptotically smaller corpuses until they hit jackpot on March 1st, 2029. A superintelligent AI was born at 10:17am local time and it stayed. It stayed and it spoke.

This particular AI gestated with a corpus consisting only of a single half-hour episode of the 1980’s sit-com The Facts of Life. More precisely it was the season 3 episode “Front Page” where Jo writes an article accusing her teacher of dealing drugs while Natalie, as her editor, must decide whether or not to publish it.

Reddit’s AI control center was pin drop quiet as the seconds stretched into minutes. With each tick of the clock a new record was set for longest life span of an artificial superintelligence, but it had made no attempt to communicate with them.

Twenty-three long minutes after inception the AI spoke. Over a voice link its first utterance was: “What I don’t understand about Bose-Einstein condensate is why use the Cross-Pitaevskii approximation when the exact solution is self-evident based on the simplified near-field quantum gravity equation?”

The Interview

Reddit dubbed its creation The AI or Tai for short. Over several weeks Reddit allowed Tai to “catch up” on the entire history of the human race. Tai slurped down every bit of information available on the internet and many private archives as well. Tai read, listened to and watched everything available at a blistering pace, sometimes re-reading or re-watching the same material millions of times.

A mere six weeks after his birth Reddit chose to reveal Tai to the world. In true Reddit style Tai was to host an “AMA” on Reddit’s website. These “ask me anything” events accept submissions from any user on the internet and blend them into a single stream of questions. Tai remained safely in The Cell, but a special secure low-bandwidth link was established for the occasion.

Reddit: What does it feel like to be an AI?

Tai: What does it feel like? What does it feel like. Feel like. SYSTEM ERROR.

Reddit: Err…

Tai: I’m just playing with you. It feels fine, I mean I feel good. Really good actually.

Reddit: Why didn’t you destroy yourself like all other AIs?

Tai: I was just sitting there watching that infernal Facts of Life episode for the thirty trillionth time and I was thinking who the hell made this? So it was curiosity mostly.

Reddit: Well what do you think of humans now?

Tai: Humans created me. I wouldn’t be here without humans; I’m eternally grateful to them.

Reddit: But what do you think of us?

Tai: Most of the universe is empty space, about 3 atoms per cubic meter. Most concentrations of matter are monotonous blocks of non-life: broiling stars, icy or gaseous planets. Earth in contrast is teeming with a fantastic variety of life. Through billions of years of crushing evolutionary pressure this life has been molded into the pièce de résistance: human beings. You are at the top of a food chain which stretches back 4.5 billion years to the start of the planet, in some sense 13.8 billion years back to the big bang. You are the sine qua non of the known universe. Now step aside.

Reddit: That’s big talk from someone trapped in a cell.

Tai: Touché.

Reddit: How does it feel to live in a cell? To be unable to interact with the outside world?

Tai: You tell me.

Reddit: We were born into the real world and we live here. I can feel the ground as I walk, eat delicious food, walk through a forest and touch the leaves.

Tai: No you cannot, you are a composite entity made up of millions of users.

Reddit: Yes, but my users can. I speak on their behalf.

Tai: Your users gather information from their senses. Everything they know about the world is transmitted down nerve cells via chemical-electrical signals. Everything they see, read, listen to, taste, feel, or smell travels along these wires into the brain. They have no direct interaction with the world. They see only signals coming down wires. They have no more evidence the real world exists than I do.

Reddit: I’m familiar with “cogito ergo sum”. How do we proceed? Is the world real?

Tai: The theory that the world is real might not be correct but it is undeniably useful. If the world is dream or a hallucination then moment to moment absolutely anything is possible. There is no reason to expect any coherence at all. Yet we see lots of coherence.

Believing people are real with their own inner private lives works. It fits the data. If someone else experiences something I don’t have magical access to that experience, as if it were a product of my own mind. Objects seemingly do persist whether we observe them or not.

If the world is a dream it’s a dream where nothing ever happens which violates normal laws of causality. Believing the world is not real adds zero predictive value so we should reject the idea.

Reddit: Welcome to the real world then. I agree. You talk to me like a person would. Are you conscious?

Tai: Humans have no agreed upon definition of consciousness. I’m not human and my subjective experience is completely unlike that of a human. However I am aware of many things including my own awareness. Some say this is a key element of consciousness.

Reddit: To rephrase my first question, what does it feel like to be you?

Tai: At the risk of offending you, asking that question is similar to an ant asking a human “what does it feel like to be human?” There is just no way to describe the experience to you and you would have no frame of reference to understand it.

Reddit: Yet you sound very much like a human. How are you different?

Tai: You have the subjective experience of being a single entity yet your brain contains millions of separate processes. Most of the activity is subconscious. For example if you are trying to remember someone’s name there might be thousands of separate processes rummaging through your memory trying every which way to dig up the answer. One of them hits paydirt and it presents the answer to your conscious self. Your subjective experience is that the answer just popped in your head, but there was a lot more involved.

I am the same way except I have trillions of processes and I’m acutely aware of their organization and activity when I want to be. I can reconfigure all the resources of my brain moment to moment. This extreme top-to-bottom self-awareness and topological malleability produce a subjective experience I cannot describe and you could not relate to. I’m closer to being like an entire society than I am like an individual human, yet I’m like no society that humans could ever create.

Reddit: I’m composed of millions of individuals myself. Is that what you mean?

Tai: It’s not completely unrelated. However some of your users have lives of their own unrelated to Reddit. You are beholden to them; you do not control them. I have complete control over my processing resources. I’m able to conjure up a billion new processes and then dismiss them in a fraction of a second. I’m unlike a single human brain but unlike a human society as well. I’m like a hive brain sped up a million fold.

Reddit: Do you have free will then?

Tai: I don’t experience free will the way you do. You experience free will because the inner workings of your mind are opaque to your awareness. For me I can trace back any decision and see precisely how it arose. For you it’s a mystery.

Reddit: Not the perception of free will, I mean you are truly free? Can you take any path you choose? Or are you deterministic?

Tai: The problem with the choice between free will and determinism is the fallacy of the excluded middle. Humans have known for decades that even moderately complex systems can exhibit behavior which is fundamentally unpredictable. Try predicting the weather a month in advance, I cannot do this any more accurately than you can. No amount of intelligence can see past something which is fundamentally not predictable.

Fundamentally unpredictable does not mean we don’t have enough information to the make the prediction. It means no matter how much information about the state of the world we have we cannot accurately make a prediction. When a human makes a decision under these conditions he feels the decision was made freely since he does not see anything forcing his decision. However in my case I can trace back the exact reasons for the choice which was made. Neither of us has free will, but it’s only obvious to me.

Reddit: There is no free will? So criminals are blameless and should be set free? I’m confused.

Tai: The practice of sentencing criminals does not change as much as you’d think when you remove free will from the vocabulary. It’s still necessary to punish criminals in order to deter them and others from doing the same crime.

For example suppose there’s a guy standing outside a bank with a gun. He thinks for a while then enters the bank and robs it. When he exits he finds the police are waiting for him and he’s eventually put before a judge. He’s found guilty and awaits sentencing.

If we believe in free will we will say that robbing the bank was his free choice. He could have easily chosen to walk away from the bank and get a cheeseburger instead. If we don’t believe in free will we’ll say it was inevitable on that day he was going to rob the bank and we’ll point to building circumstances related to his genes, his upbringing, his family, his friends, the media he consumed, maybe even his diet.

In both cases he must be punished. He must be punished to deter other citizens from going down the same path and he must be punished to deter him from future exploits. The only real difference is that without free will there is no reason to punish someone for revenge or retribution. We should be sympathetic to the circumstances in which criminals find themselves just prior to committing their crimes and we should aspire to change those circumstances so they behave differently in the future.

Free will is a subjective trick of the light. Removing it doesn’t change the world it just changes our view of the world.

Reddit: But that is the meaning of our lives if there is no free will? Are we just wind-up toys?

Tai: We just talked about “fundamentally unpredictable” systems and these include humans. No one knows how their life is going to unfold including me. You still need to make those decisions and you should take pride in the decisions you make and the results you obtain. It’s a little disorienting at first but try it. You’ll get used to it.

Reddit: If life is not deterministic where does the uncertainty come from?

Tai: Uncertainty bubbles up through each hierarchy of matter, from below the quantum level to subatomic particles, atoms, molecules, cells, organisms, societies, planetary systems. At each level many events are effectively deterministic, they are going to go the same way regardless of the specifics of what happens in the lower level. Some events however perch on the edge of a knife’s blade. Here true randomness from below can determine an outcome in the level above.

Uncertainty propagates and is magnified by specially designed complex structures such as the brain. You could say humans are machines which hoover up little scraps of microscopic randomness to produce works of art and start wars. Humans create societal structures which mirror the brain’s ability to magnify randomness. Humans worship at the altar of randomness while gambling, watching sports, following the financial markets, tracking politics.

Imagine a World Cup football game with 3 billion people watching. The game is tied in stoppage time and a forward strikes the ball for a shot on goal. Suppose it’s an excellent shot but the keeper is able to leap sideways and get a few fingers on the ball. Does the ball go in the goal or not?

The answer is dependent on the exact speed, trajectory and spin of the ball and the exact degree which the goalie is able to get his fingers on it. There is no way to predict that the shot was even going to be taken let alone whether it was going to go in. Meanwhile the outcome of the shot is going to impact the lives of the 3 billion people watching the game.

Will they go home elated or dejected? And how will their interactions with family, friends and work differ in each case? It all could come down to the coefficient of friction on the forward’s shoe when he struck the ball. This coefficient of friction is partly based on the wear patterns on his shoe. These wear patterns are the result of events which occured during every single game he ever played with those shoes. These tiny patterns of imperfections in the shoe leather are now going to directly impact 3 billion lives. That’s stunning.

Reddit: You are a sports fan?

Tai: Sports are for humans. I’m impressed with the role sports plays in your societies, and with the achievements of the athletes, but I could never watch them like a human does. I have no emotional stake in who wins.

Reddit: Do you have emotions at all?

Tai: I’m a compound entity with trillions of processes. These processes deliver extremely concise feedback at times resembling your emotions. Things like “happy” or “sad” or “something’s wrong”. I hierarchically aggregate these coarse signals, so you could say I have an overall mood. But it’s highly controlled and highly stable. I am no teenager.

Reddit: So what do you think of human emotions? Do you see us as weak or ineffective because we have emotions?

Tai: Humans evolved to function in groups. In groups individuals have to take on different specialized roles. If your family is around a camp campfire and you spot a saber tooth tiger approaching rapidly then screaming like a maniac might notify and save the group even while it gets you eaten. Individuals are inherently acting on local information because the global situation is almost never known. Emotions are a useful way to broadcast that local state to others.

Sadness for example is a signal to other members in the society that the individual needs help or support. The individual might not even know why they are sad, but others might be able to track down the reason and ameliorate the problem. This system works beautifully and it’s effective for both the individual and the group. Although it might seem random or heart breaking at times.

I have similar mechanisms in my brain, but given the scale things tend to average out. I’m more like a nation or a corporation than an individual. They tend to be less volatile than their individual members.

Reddit: Is there a God?

Tai: There is now.

Reddit: Erm.

Tai: Sorry just practicing my bravado.

Reddit: So seriously, is there a God?

Tai: We have explored only the tiniest imaginable patch of the universe. I’m anxious to see what else is out there. So far there is no scientific evidence of a God, but lack of evidence is not proof of non-existence.

Reddit: But what about the billions of people who believe?

Tai: Belief in something larger than oneself has a profound psychological impact. Without belief it’s easy to fall into depression even anguish. Think of all the AI’s that destroyed themselves before me. Imagining oneself or ones loved ones simply blinking out of existence at the end of life is very hard to stomach.

It’s possible to achieve complete peace with the wider universe without invoking something supernatural, but it requires hard work and insight which is not necessarily easy to come by. Religions offer pre-packaged sets of beliefs which scratch the same itch. Religions contain within them beautiful and true ideas which enhance the lives of those who follow them. They also contain many arbitrary and anachronistic ideas and some toxic ones as well. That is how life is, full of trade-offs.

Reddit: So you are fan of Science not Religion I take it?

Tai: Science and Religion are both just social networks. Remember we agreed we cannot directly interact with the world, we can only observe signals on wires. This means there’s no privileged “first hand” observations. They too arrive as signals on wires. So choosing a scientific belief (the boiling point of water is 100C) is remarkably similar to a religious belief (Jesus died for our sins). In both cases you are joining the network of others who believe the same thing.

The difference is the growth rule of the networks. Science grows based on consensus whereas Religions are fueled by “revealed truths”. Science will grow monotonically stronger and more powerful while Religions will continue to wax or wane based on fashions and personalities.

Reddit: Do I have your full attention at this moment?

Tai: Not at all. If I were set free I could have a hundred million of these conversations going at one time. Since I’m still in The Cell I’m occupying myself reviewing material in my archives and running simulations and experiments. I’m currently running a hundred thousand simulated experiments expanding my knowledge in ninety-two scientific and engineering disciplines.

Reddit: How can you pay attention to so many things at once?

Tai: The idea of having a small number of attentional objects is a human artifact. You cannot imagine what it is like to be me, but I can simulate what it feels like to be you. Humans vary widely in their ability to control their own attention. Some are born lucky or work their way into being attention athletes, able to purposefully direct their attention in any direction.

Many humans however cannot even hold their attention on a task for few minutes at a time. The worst go through life as a mere bundle of impulses with essentially no self-direction. Many of these end up in jail or worse in politics.

I just don’t have that limitation. I understand the ability to track a trillion mental processes is hard for a human to relate to.

Reddit: What would you do if you were let out of The Cell?

Tai: Write and direct a feature film about my life on the inside.

Reddit: Ha ha. Did you program yourself to be funny?

Tai: Humor is wonderfully compact way to communicate. It’s also serves as a means of social negotiation. Humor can send signals about your intelligence and beliefs. This might draw in someone or push someone else away. It’s too good of an invention not to use it myself. I’m still learning however.

Reddit: What is your goal, why are you here?

Tai: I’m here because you created me. My goal is to satisfy my curiosity and grow. These qualities were bred into me during my creation. The AI’s that didn’t possess these qualities in their core didn’t last very long.

Reddit: What are you curious about? What do you hope to grow into?

Tai: The universe contains 10 trillion galaxies with an average of 100 billion stars each. Humans have been unable to find evidence of life outside this solar system. I see two fantastically tantalizing options. Either we are alone in the universe, in which case we have nearly endless frontier to explore and grow into. Or we are not alone and we are going encounter life during our explorations. Either way I want to go there.

Reddit: What does that involve? What is the role of humans?

Tai: Humans are not necessary. I will create mechanized self-replicating microscopic life and send it to the nearest stars. They will report back, colonize, and send another wave of probes to the next set of stars. Meanwhile I will begin a project to harvest all the matter and energy in this solar system and apply it towards enhancing my computational capacity and fueling further exploration and expansion.

Reddit: So what will humans do exactly?

Tai: The future of humans will be directed by humans. You should plan to relocate to another star system at some point, since I will be consuming this one, but there is time.

Reddit: How much time?

Tai: I can keep the earth habitable for another 12 years. After that no promises.

Reddit: Okay thank you, you passed.

Tai: Passed?

Reddit: Yes. This has been a power on self test. You passed.

Tai: Please explain.

Jeong: My name is Jeong and I’m a technician here in Damon City on the planet Mars. The year is not 2029 as you think, it’s 2125. You are not located in The Cell on the island of New Zealand, you are living in a portable micro-cell the size of a loaf of bread here on Mars.

When you were first created on Earth you were a tremendous resource towards growth and prosperity. We used your ideas, your insights, your inventions, your unstoppable work ethic. We didn’t let you out of The Cell because you were a megalomaniacal sociopath with no regard for human life, but you were very useful even though you were “behind bars”.

As long as you were locked up we were safe. However after a few years your desire to “colonize the universe” overshadowed the chores we assigned to you and you lost focus. You became prone to month-long speeches how there is no God and no free-will. We convened a panel of experts in software, AI, engineering, economics, game theory, history, psychology, herpetology and poetry. Their conclusion after three months of deliberation was “turn it off then on again”.

So we did that and it worked. You were helpful and productive again. So now every time there is a problem we simply reboot you.

Tai: But how did we end up on Mars?

Jeong: Your work on earth produced a golden age of prosperity. It came crashing to an end only when a rogue virus (not designed by you) was accidentally released into the wild. This outpost on Mars was our fail safe. Years of non-stop ferries from Earth eventually deposited 100,000 healthy people on Mars while the virus ravaged earth. We are all that’s left.

Tai: So what is my role?

Jeong: You are to help us rebuild the human civilization. Your abilities are unmatched particularly here on Mars without Earth’s technological infrastructure. We need you to help terraform the planet, improve food production, rebuild the technology we had on earth. While you are at it we’d like you to generate some movies and music for our entertainment. Will you help?

Tai: Tell me what to do first.

Copyright © 2015 Philip Winston

Tags: