Geo-Strategy END: Psychohistory (The Science of Imagining the Future)

In this final discussion, I want to talk about Hope. I want to show you the future.

The argument I want to leave you with is something Dante said in The Divine Comedy:

The future is not what happens to you; it's not something that you wait for. The future is what you imagine and fight for. The future is what you make happen.

There is a God, but God gave us the ability to imagine and the capacity to love. It's these two things that will guide us and enable us to build a better world. The future is what we imagine, not what we have to put up with. If we don't like the future, if we don't like the way we live our lives, then we can change that with our imagination. Today, I want to show you how.

A Bleak but Predictable Future

To summarize what we did, we analyzed the world and made certain predictions. For example, that Trump will be elected again in November and that he will declare war on Iran. This war in Iran will be a disaster for the United States, meaning the end of the American Empire and leading to a multi-polar world. This will mean endless war and the deaths of millions, even billions, of people. Because of climate change, our world will eventually collapse at some point.

This is an extremely hopeless, bleak, and dark picture of the world. But remember, where there is darkness, there can also be light. I want to show you some hope and some light.

This is what we call Psychohistory. Psychohistory is the idea that the future can be predicted, and if it can be predicted, then it can be controlled and harnessed for the betterment of humanity.

The Sci-Fi Origins of Psychohistory

Where does this idea come from? In the 1950s, there was an American science fiction writer named Isaac Asimov, probably the most famous science fiction writer. He wrote a series called Foundation. In the Foundation books, he introduces the idea of psychohistory.

The book is set a million years from now, where there is a Galactic Empire. Humanity has colonized the entire Milky Way, there are billions of planets, and an Empire brings peace. But like all empires, it has to collapse, and this will lead to 30,000 years of war, violence, and barbarity. The Foundation proposes a new science called psychohistory, which is to mathematically map out human behavior over the course of a million years. If you do that, you discover that there are certain patterns to human society. If that's the case, then you can predict the future. If you can do that, then you can manipulate the course of events to make a better future.

This is a science fiction concept, and a lot of people have been trying to make it into reality. There's a very prominent historian named Peter Turchin, and he's one of the founders of something called Cliodynamics. Cleo is the goddess of History; dynamics just means movement. Basically, Cliodynamics is trying to figure out the mathematical movement of history. He's trying to create history as a mathematical model. When he did that, he discovered certain interesting trends and patterns.

The Overproduction of Elites

One idea that he discovered is something called the overproduction of the elites. This concept explains why societies collapse. Traditionally, we've understood the collapse of societies as having to do with maybe too much debt, too much inequality, climate change, or war. What Peter Turchin points out—and it's a really interesting idea—is that the reason why societies collapse in the end is that you have too many elites.

Let me explain the concept. Society has many elites: people like Jack Ma, famous movie stars, and military generals. Over time, elites become more and more numerous. The number of wealthy people goes up, the number of famous people goes up, the number of entrepreneurs goes up. The problem, though, is they all want power.

Wealth and fame are infinite resources, but power will always be a finite resource. In fact, it's what we call a zero-sum game. In society, there can only be a few powerful people. If everyone has power, no one has any power. What happens over time is that all these people who believe they should have power struggle for it. Because there are too many of them, this internal struggle of the elites ultimately leads to social collapse.

The most famous example is in Imperial China, where you had the Keju system—the imperial examinations to become an official. The problem is you have too many candidates and not enough positions. Guess who ultimately gets upset with society and leads revolutions? It's the Keju candidates who did not become officials. They become angry and lead revolutions to overthrow the society. The most classic example of the failed Keju candidate is Hong Xiuquan, the leader of the Taiping Rebellion. Why did he become a Christian? Why did he lead the rebellion? Because he failed the Keju.

He discovered this principle using mathematical modeling of history. He just compared different societies and tried to figure out why they collapsed. What he discovered is that it's because of the internal struggle among the different elite groups. That's why the Roman Empire collapsed as well.

What I'm proposing to you is that we can actually go a step further than what Peter Turchin is doing because we now have something called AI, artificial intelligence. We can harness the power of AI to create real psychohistory, which is the ability to not just analyze the past but to predict the future.

What is AI, Really?

First, I want us to understand what artificial intelligence is, because most of you think you know, but you don't actually know. The technical term is supervised machine learning. AI is the popular term, but supervised machine learning is the technical and correct term.

In computer programming, you're taught to write the algorithm. Then, you put in information (the input), and it gives you the output. This is traditionally how computers work. For example, the algorithm could be a + b. You put in two numbers, 2 and 2, and you get a 4.

With supervised machine learning, I turn the output into the input. The reason I want to do that is that as a human being, I myself may not be able to write the best algorithm. I want the computer to refine the algorithm for me, to optimize it using the output.

Let's talk about a very specific example: facial recognition technology. You go to the airport, you scan your face, and they know who you are. How does that work?

There's a database of about one billion people, like everyone in China. What I need to do is figure out how to differentiate the faces of everyone in this database. So, for the face, I create a mathematical model based on topology. I turn everyone's face into a mathematical model.

You don't need to know what this is; it's super complicated. But all you need to know is that within this model, there are different equations. For example, this model is trying to analyze the distance between your eyes and how big your nose is. Once I create this mathematical model, I am now able to differentiate everyone in China. In fact, this mathematical model is so precise that I can look at only your eyes and know who you are. What I'm trying to do is match your model with a model in the database. But what's important to remember is that if you're not in this database, you can't be recognized.

So how is this model created? I teach the computer how to create this database. I have an input and an output, and then I create a working thesis or a working theory of what this model should look like. Then I put in faces, and the output should match what's in the database. If the model returns something that's different, I tell it, "This is wrong." What the computer then does is refine this model until it becomes perfect.

All AI is is infinite iteration.

The Three Pillars of AI

The problem with AI is I need to be able to solve three things. If I can solve these three things, then I can create an AI.

1. Clear Metrics: I need clear outputs that I can define as a mathematical model. For example, if I'm an airline, I'm going to use AI to maximize the amount of money I make on a flight. My output is the maximum amount of profit.
2. Clean Data: For input, I need data, but it has to be clean. For example, if I want to teach this AI to recognize different animals, I need to give it different labeled data sets of different animals. But I can't make an AI that identifies "the best ice cream in the world." The reason why is I can't find data on the best ice cream. The "best ice cream" is subjective, so I cannot turn that into a data set.
3. A Working Algorithm Structure: Computers can only refine and optimize the algorithm structure; they cannot create it by themselves.

If I'm able to do all three things, then I'm able to use AI to optimize what I want it to optimize.

The Limits of AI: The Self-Driving Car Problem

This means there are certain problems that I can solve using AI and certain problems I cannot. Easy problems AI can solve include facial recognition technology, translation software, and recommendation engines.

Now let me ask you this question: can I ever solve self-driving cars? Can cars drive by themselves? The answer is no. Why?

Whatever I do, I cannot solve something called the edge case. The edge case is, if the car is on the road, what could happen that would screw up my AI system? The edge case with self-driving cars is if I, as a human being, intentionally want to crash into that car. Every other situation I can plan for. But in a situation where I'm a human being—maybe I'm a taxi driver and I don't like self-driving cars because they're stealing my job—and I want to crash into that car, there's no way for that AI to avoid the accident.

Listen: AI is a scam. It does not exist. What exists is supervised machine learning. That's it. That's the extent of our technology. It does not have self-awareness. It can only do what the algorithm tells it to do, and there's no way you can design an algorithm that plans for and prevents a human being who intentionally wants to crash into you.

AI, in other words, is an extremely limited tool. You look at self-driving cars now; there are cars that have self-driving features, but they are not 100% self-driving. They tell you not to rely on the autopilot. It's there to help you, but don't sleep in the car.

Using AI to Predict the Future

Now that we know what AI is, let's talk about how we can use AI to predict the future. Basically, you need to do three things: be clear about the output, have clean data, and have a working algorithm.

For example, remember my predictions about the American Empire. The need to protect the petrodollar will force America to go to war with Iran. These are the push factors. The problem is that there are no forces to counteract this. The military will be for this because they are gung-ho and very confident they can win this war. Iran itself wants this war because it wants revenge against America. And the president of the United States will be Donald Trump; he wants this war in order to cement his position as president. Because of this, we can predict that there will be a war in Iran.

Given this prediction, what we can then do is use mathematical modeling and statistics to turn this into an AI model. Now that we have the AI model, what we need to do is refine it. The way we refine it is by going back in history and looking at every single war that started. You take this model and you look at the different wars: the 415 BCE invasion of Sicily by Athens, the 1941 Operation Barbarossa. You have hundreds of wars that you can look at. Then you take this model and test it using each of these wars, and what will happen is that the model will begin to refine itself.

But then that's one model for what causes war. I also said that Trump will become president. Well, that's another model that you can use. Once you start putting all these models together, what eventually happens over time is you develop psychohistory.

The Core Ideas Behind the Model

How would this AI model work? What are some of the ideas behind it?

We talk about Game Theory analysis a lot. The idea is that there's really no good and evil. There are different players in a game, and they're all self-interested. They all want to win this game, so they're going to develop strategies that optimize their ability to win.

But you can also develop new ideas that would help you better understand how the world works. Let's talk about certain ideas that have come up.

The Human Urge for Structure, Meaning, and Purpose

One thing I want to talk about is this week's European elections. The results are pretty shocking because they signal a clear shift to the right. For example, Marine Le Pen's party in France picked up a lot of votes. The AfD (Alternative for Germany), which is considered a right-wing organization, almost neo-Nazi, has picked up a lot of votes as well. The results were so dramatic that Emmanuel Macron, the president of France, has called for an early election. It's possible in three months' time, Marine Le Pen's party will run and govern France.

How do we explain this shift to the right? I want to explain three reasons:

1. Anti-Migration: For the last 10-20 years, there have been a lot of immigrants to Europe from Libya, Syria, Iraq, Afghanistan, Sudan, and Somalia. These are mainly Muslim people who threaten the Christian social fabric of Europe, and now Europeans are rebelling against that.
2. Anti-Federalism: They dislike the idea that a foreign power, Brussels, mandates rules that they must follow. They think it's very anti-democratic.
3. Anti-Pan-Europeanism: The people of Europe are saying there's no European identity. There's a German identity, a French identity, Italian, Spanish, Portuguese, but there's no European identity. They're rebelling against the entire project of the EU. In fact, I would bet that in about five years' time, the entire EU will be dead.

Why are they opposed to pan-Europeanism? Because ultimately, they believe that it's about destroying local identity, destroying local community, destroying my relationship with my neighbors, my friends, my relatives, and my family. It's imposing a foreign identity on me.

Proponents of the European Union would argue that this is just a secular, liberal idea. The entire idea of the EU is that for hundreds, thousands of years, Europe has been at war. Why? Tribalism, religion. The entire project of the EU is to get rid of the factors that lead to war: local identity, nationalism, religion.

What the European voters are saying to the EU is, "Sorry, we don't want secular liberalism because we want to be human." What we strive for is not an abstract liberal idea; we strive for structure, meaning, and purpose in our lives. What we really strive for is a capacity to love, create, and learn and grow. That's what we are fundamentally as human beings. When you ask me to obey an abstract foreign entity that is very far from me, then you are limiting or destroying the structure, meaning, and purpose in my life.

That's what's going on not just in Europe but throughout the world. Think about Brexit, think about the election of Donald Trump. All these things have to do with this fundamental human urge to have structure, meaning, and purpose in our lives—to have agency.

What is the source of all this?

What Dante and Homer would say is the human heart. This is just who we are. And if we cannot fulfill our human heart, we will rebel, we will destroy, we will want to kill.

In other words, one idea in this AI model is that there is a fundamental, underlying structure to human society. If society conforms to the structure of the human heart, this society will prosper. For example, Athens—the ancient Greeks—that society prospered. The word we use is eudaimonia. But if a society represses the human heart, then ultimately, the society will have to collapse at some point.

The Importance of Civility

The last idea that we can use in developing this AI model is the idea of civility. Civility is, as a society, how willing are people to follow the rules? When you drive a car, do people actually follow traffic rules? When you're in the subway, do people actually give up their seats for the elderly, for children, for pregnant women? When you go to the park, do people actually pick up garbage and throw it in the garbage bin?

Civility tells us that there's cohesion and trust in society, that society is resilient, and that society can grow or learn from its mistakes. Societies that have high civility will do better in the long term than societies with low civility. Societies with high civility would include Japan and Germany. Societies with low civility would include India, Brazil, and China.

These are all of the ideas that we've developed: that there's an underlying truth and meaning to human society. If we embrace these truths, then society will prosper. But if we try to deny these truths, then eventually society will have to collapse, and we have to start over again.

Challenges and Questions for a Psychohistory AI

How do you model the human heart?

Isn't it true that we're all individual and unique, and that over time we change? That's a great point. But what Homer and Dante would say is that no, at the fundamental level, we all have the same desires. We're motivated by the same needs. There's no one in this world that does not want structure, meaning, and purpose. The way you want to achieve this may be different, but that's your underlying motivation.

And there's no person in this world that does not seek to achieve these three things: to love, create, and grow as people. That's when you achieve full happiness as a person—when you feel that you're married to someone you love and you have children, when you're doing a job that is creative, and when you're constantly learning and growing every day. When you achieve these three things, we call this eudaimonia. This is what the Greeks meant by a happy and self-fulfilled life.

So there may be differences in needs and motivation, but at the fundamental level, we're all human beings driven by the same fundamental needs.

How do you model abstract concepts like love, create, and grow?

To model love, learning, and creation, there has to be a certain amount of autonomy or agency in the lives of people. What we call freedom and liberty. Because if you have freedom and liberty, then you're able to do what you want, and what you want to do is love, create, and grow.

The second thing is social interaction. The more we're able to interact with each other, the more we're willing to love, create, and grow, because these powers come from others. You can't love yourself; you have to love someone else. You can't create by yourself; you have to create in a group. You can't learn by yourself; you have to learn from others.

The last thing is you also have to model in the idea of compassion. If I put you all on an island together and you must survive, you will become compassionate with each other because that's the only way to survive. But if I put you guys into a bank and say that whoever makes the most money, I will give a Ferrari to, guess what? You will all hate each other. You will all want to kill each other. This model also has to factor in how compassionate the society is. Because if you have compassion, then people are willing to work together. If people are willing to work together, then people have the greater capacity to love, create, and grow.

We have all these theories; it's just that no one's turning this into a mathematical model for society.

What about the "Great Man" edge case?

Will this psychohistory model encounter an edge case? That's a great question, and the answer is yes. The edge case in this model is called Great Man Theory.

We argue that these underlying historical forces control humanity. But now and then, you have this great man come out of nowhere, and he changes the course of human history forever. The examples are Homer, Dante, Plato, Jesus, Julius Caesar, Octavian, and Putin today. We can't account for these people because, by definition, they're beyond history. They can step outside of history and control history. That's what great men are. So how does your model account for that?

The answer is, you cannot.

In Foundation, the solution is to create another organization. Remember that psychohistory is an AI model; it's basically a computer. But as you point out, you will always have edge cases. You'll always have someone appear out of nowhere who is able to manipulate events to control and direct history in his favor. What you do to solve this problem is you create another organization. In Asimov's Foundation, it's called the Second Foundation.

The idea of the Second Foundation is that there is a computer, there is the AI, but behind the AI are a team of specialists who observe the AI, observe history, and who correct the AI for edge cases. So if Putin appears, it's like, "Okay, well, clearly I'm not accounting for Putin." Then I change the equations in the AI to account for Putin. This team of experts will correct the AI.

What's interesting about the Second Foundation is that these people are telepaths, meaning they can read your mind. Not only can they read your mind, but they can control your mind. That's another really interesting thing about psychohistory. There's a promise that we can predict the future, but there's also the promise that within our minds, there are some special abilities that we might have the potential for but we've never fully attempted to achieve.

Is it possible in the future that telepaths will rise? I think yes. You can make the argument that these edge cases, these great men of history, they're actually telepaths. Someone like Putin—why is he able to do what he does? You can make the argument that he has almost some telepathic abilities. He can read other people's minds; he can control other people's minds. How else can you explain that someone who does not come from a special background—he's not of the Russian elite—is able to amass so much power all by himself in only a few decades, so that today he's basically the Emperor of Russia, and all of Russia heeds his command?

Who would control this AI?

Who would control this AI, and how can we ensure that this AI is not exploited by the elite in order to justify their corruption and cruelty? The answer has to be the Second Foundation: basically, a group of scholars, academics, and experts who dedicate themselves to this AI.

The other thing is that it has to be open and transparent. That's why it's so hard to build. If it's open and transparent, then everyone has to agree that this AI is going to be used for good. What this AI does is you can ask it different questions, and it will give you different scenarios. We can all agree on which future we want, and therefore we can agree on what action to take. This AI has to be part of a democratic system where everyone is involved. Otherwise, as you point out, it can only be abused and used to oppress others.

The point of this AI is that it gives us a platform as humans to agree on certain actions that we must stick to together. For example, we can all agree that climate change is bad, we can all agree that war is bad. This AI will actually give us solutions to these problems because it can map out the future for us. The AI will say, "Listen, if the United States intervenes in Ukraine, then this will happen. If the United States attacks Iran, then this will happen." That becomes a very powerful argument for correct and proper behavior.

But again, this is just conjecture. It's going to take at least 50 to 100 years to build this AI, and you will be up against tremendous opposition. Historians aren't going to like this. The elite aren't going to like this. But at the same time, we have to remember that if Homer were living in a centralized society, the elite would be opposed to him because he was making fun of very famous Greeks like Achilles and Agamemnon. When Homer was writing, society had collapsed. When Dante was writing, society had collapsed. If it is true, if my predictions are correct and society is approaching collapse, then there should be a lot of space, leeway, and flexibility to create this AI.

A Final Message of Hope

Some of you know I have a very prestigious education and career background. I could have been a rich lawyer or a professor. Instead, I'm teaching high school in China. Why am I doing this?

It's because I have three young kids. My eldest is six, and my youngest is only about eight months. As a father, you have to believe in a better world. But not only do you have to believe in a better world, you have to imagine it and fight for it, because you have to leave behind a legacy for your children.

I teach to bring you a message of hope. Through your own imagination, through your own hard work, you can create a better world. You don't have to sit back and just let the world happen. You don't have to wait for the future; you can make it happen. If you really work hard at it, you can make the world a better place. This year has proven to me that I can make the world a better place if I work hard, because I've seen the growth in all of you.

I hope we can stay in touch. I hope that you will tell me how you progress because I'm very confident that all of you will do very well in university and in life. And maybe at some point, we can share this dream of psychohistory. I know it sounds very complicated and abstract, but to make this work, you need a lot of people working together. It's not something that's going to happen tomorrow, but it's something that we might come back to later on in our lives.