Robin Hanson was educated at the Universities of California and Chicago, and received his Ph.D. from the California Institute of Technology. He’s worked at the Institute, as well as Berkeley, Lockheed, and the NASA Ames Research Center where he investigated Bayesian statistics. He’s since become a professor of economics at George Mason University, and a researcher at the Future of Humanity Institute at Oxford University. Much of his output has dealt with signaling theory, prediction markets, and the vagaries of efficiency in the complex modern economy. His recent work, drawing from expertise in physics, computer science and economics, has explored the possibilities presented by brain emulation. All this can be found in his book The Age of Em: “Scan a human brain, then run a model with the same connections on a fast computer, and you have a robot brain, but recognizably human. Train an em to do some job and copy it a million times: an army of workers is at your disposal. When they can be made cheaply, within perhaps a century, ems will displace humans in most jobs. In this new economic era, the world economy may double in size every few weeks.” Dr. Hanson is having his brain cryonically preserved.
I’ll start simply: is there a God?
I didn’t hear that.
Is there a God?
Was there, at any point?
Well, I was a Christian as a child. I even joined something of a Christian cult. But by college I became an atheist.
You joined a cult as a child?
A teen. A Pentecostal Christian sect. With speaking in tongues, and group homes, and a compound out in Iowa somewhere where the more dedicated went to live.
So is it a case of teenage rebellion becoming early skepticism?
No no. It was an appealing social milieu. In a Pentecostal sect people feel a high degree of connection and they are very inviting of outsiders and they express a lot of feelings for each another and they have a cause and there’s a reason, a reason for the community to exist. Those were all appealing.
I recall reading something by you about how if one gives up the benefits of religion, because of their love for truth, then why not also give up stories? You’ve given up stories, too?
Honestly, a lot of the appeal of religion comes via the form of religious stories. The social functions of religion overlap a lot with the social functions of stories. I do try to reduce my reliance on stories. To the extent I want to enjoy stories I might like to partition them off from my beliefs about the world, but that’s actually pretty hard.
Well with a goal like that…
But I might say that even if I reject formal religions, I’m still embodying many religious attitudes and forms in my life. So I exclude neither stories nor religion from my life, but, on the margin, I can somewhat reduce their influence. A common Christian theme is the disapproval of sin and the celebration of those who have at least partially resisted sin. There’s a certain style of a sermon in which one describes the certain makeup of a sin, describes its cause, points out its many examples in people’s lives, and provides communal encouragement and a promise of praise if you drop the behaviour. I follow that form when I write about avoiding bias. So I am describing bias as a sin and invoking a larger religious framework that my community is comfortable with.
Is economics a science?
Sure, yes, but I don’t so much care about the word science. It’s more of a broader cultural reference than anything that has a precise definition.
Science… It’s fascinating to me how science – since Keats brought up this view in his diatribe against Newton – how science has been leeched of its imagination by being called ‘science’. How can we decouple imagination from the negative label of ‘bias’?
See it for what it is. Imagination is one of our standard mental capacities. We are imagining every time we envision anything that isn’t real. When I plan my day tomorrow, I’m imagining, and maybe my plan won’t be realised. It’s a question of using this mental capacity for one goal rather than others – finding truth, instead of being dramatic or fun. I don’t mind if the true things that I imagine also happen to be charismatic, and dramatic, and engaging, and maybe even creative.
I want to try this; say you exit this office now and have to describe your book – The Age of Em – to a passerby, for, whatever, a minute. It’s a really crazy book, for me it seems crazy. And after you describe it, the guy selects – ‘plausible’ or ‘implausible’. How would you describe it so the guy pressed “plausible”?
Well first I have to justify a wide range of possibilities as worth considering. So I might say: look, we have a lot of historians in the world. The future is more important than the past – because we can do things about it! So when we look at the past, we break it down by time and place. But when we break down the future, it’s different – we break it down by the major things that could happen. So it’s worth having a hundred books on the future, it’s worth having a hundred books on a hundred different scenarios, each of which has only a 1% chance of coming about. I’m arguing for this standard. If I give you a scenario that has a 1% chance, you should still think it’s worth talking about, because it’s realistic enough and the future is important!
It’s a question of what you want to spend your time on. One would just make the argument: well, look; this book doesn’t have any of your friends in it, none of your children, and probably none of your grandchildren. It’s about people you don’t know in a world you don’t live in. And other people aren’t talking about this, so you can hold forth at parties and your lack of knowledge of my work probably won’t embarrass you.
Do you resent this?
Why resent? Most people live like that, actually, and most people don’t live boring lives. It’s boring for you and I, who have unusual taste.
Well, maybe. People who want to know what the future will be like. And that’s quite rare, but in those communities you might be embarrassed about not knowing.
I’m trying to get a feel for how much it matters to you, that link between truth and speculation. If we speak narrowly for a moment, people have modeled minds – or whatever you call them – on silicon. But there’s a big big difference between that and transistor systems, which is what people like Carver Mead are invested in. And people have made headway recently in creating ‘crystal neurons’ that mimic the randomness that comes into any neuronal signaling.
Yeah, I heard of that.
But here we need phase change material and we’re still so far from linking anything into even rudimentary network systems. So if we do accomplish the emulation of minds, how can we know what size these creations will be, how they will be maintained, and so on?
The transition to this world is one that we can speak about with less confidence than the equilibrium of this world. I think of these two very distinct paths: one is to stay relatively far away from the brain, to observe the structure of the brain, and use that as inspiration to develop algorithms and tools that do roughly what the brain does in context. And the second tries to take the brain and emulate its pattern of activity.
The grammar and logic of soft or hardware is not necessarily the grammar and logic of the human mind – though it may have the same input and output.
Of course. But the question should be: for what purposes is that difference relevant? So just how much of the detail of the brain will we need to emulate in order to emulate the human mind. The closer the input-output and behavior correspondence is, the more exactly the emulation will do the same thing in the same situation. Including, if you ask it if it feels human, saying ‘yes’.
So if I understand correctly, you’re taking the gamble that, in the gradual laborious development of the emulated mind, mimicking the randomness and idiosyncrasy of what makes us human, we won’t divert, stop, compromise – ‘artificialize’ – along the way?
Yes. If we artificialize before, my book is largely irrelevant. But, see, we know that human minds are capable of enormous intelligence and are tremendously valuable. They function quite impressively, quite impressively compared to most of the machinery that we’ve ever built.
That’s a nice sentiment.
It’s a question of supply, demand, and investiture.
When I read your book I’m tempted to say: look, if it’s plausible that this stuff will come about, we’re scraping against the next revelatory stage of economics. It might feel like being a chemist before Wöhler came about – a chemist who nonetheless knew Wöhler would come about and had nothing to do but wait. Or a physicist who had to sit around and linger for Heisenberg. A geneticist before Morgan, a psychologist before William James, so on. We see this world on the horizon and it’s interesting, but we’re not there, so feel a little, well, dry.
The future attracts, but you have to resist it. You stay in the moment with your eye to the future. I was an artificial intelligence researcher for nine years. And that’s a very slow process. We’ve made machines perform tasks that, some time ago, were solely the domain of humans, but we still seem to be a long way off, yes. That’s why it’s so important for us to talk about plausibility…
As a counterweight to the numbing power of imagination.
Yes, in any area of inquiry one has some sort of basis of knowledge or understanding or theory to draw on and one also has a sense of the things they don’t know. And your job, your goal, should be to think about the things you don’t know while limiting yourself to the current knowledge that will give you confidence in what you say. Analysis is a game between the degree of uncertainty and the degree of confidence. Again, that’s what I always say, the future is more important than the past. I could envy the future for its pool of knowledge, but right now I’m in a position where I can still influence it. We know more now about the industrial revolution than someone living in the 1500s, but by now it’s too late!
Do you think the economist’s toolkit will change fundamentally with time? I get the sense that you do not.
Listen, in economic journals, we’re already not talking about humans. It’s these abstract, alien creatures. We presume to know exactly what information they have, that they’re able to calculate long trees of possibilities of what people might do, that they can calculate these things at high precision without ever making a mistake. In addition, these creatures are often quite selfish. They’re also often focused solely on the moment, or they might have a very long-term focus. They really are just alien creatures. And then we have the lens of efficiency. Much of economics is a game, a kind of applied mathematics. There should be a place for the real and descriptive in economics, but think a lot of economists miss how far out from the real world they actually are. I think, fundamentally, economics should study efficiency and teach people how to be efficient and unbiased.
I’ve always thought: there is no universal scientific method, every science has an approach. I mean, you look for informational exchange in biology, but you don’t look for informational exchange in thermochemistry. Economics could be called the study of efficiency, yeah?
If you look at random states of affairs and think about making them more efficient, I think you will find that actual states of affairs are more efficient than random states of affairs. That is one way to understand the world. The world tends in the direction of more efficient things.
I think many would happily return to administrative economics. I’m referring to Herbert Simon here, the psychological study of how people can best organize. But such work, as I see it, is now mostly being done in business schools, not economics departments. Perhaps regular economists are burdened by the weight of prediction – everyone wants to know when the markets will crash next, and so on.
Economists also profit off of that. That’s how stars are born. The problem is that they take shortcuts and have no actual stake in the game. So I can write a hundred opinion pieces heralding disaster, and I’ll be right a dozen times. That’s why I worked on prediction markets – it’s an old idea, you can trace it back to [Friedrich] Hayek – getting people to trade on what they claim they believe. Put money in. And when you don’t want to, be quiet. You’d see economists working more. Working on substance.
You write that unequal love is a problem. I think I might take issue with this. If you’ll forgive me (laughs) – what is love?
I don’t need to define love in order to talk about love in my capacity as an economist. All I need to do is refer to the sense people have: do they feel loved?
But come on, who cares what people say? You have two people having sex, and they say that they’re not in love, do you take that straightly?
This doesn’t matter as much as you seem to think it does. As an economist, I view people wanting things, people wanting care, people wanting sex, people wanting cosmic revelation, and I can look at how efficiently they achieve those things. Sure, that can be explicitly stated, or observed in behavior.
In biology, some guy comes around with systems theory, and says, look, how brilliant, we’re seeing the ovary interact with the insular cortex, and it’s so revelatory! Well, it might be, but you’re still trying to understand a single phenomenon, fully. Yet there is no unified economic theory.
If you’re thinking about a starship, and you’re thinking about what kind of engine or shielding you might make, then you’re focused on the physics. But then you start talking about designs, and who would go on this starship – and there, you’re firmly back in economics! It’s encompassing – if you’re focused on the degree of destruction inflicted by cosmic rays on your heat shield, it seems like you’re solving an engineer’s problem, that’s not so much economics, but that part is certainly relevant to the social interactions of how much this ship will cost and who will get together to build it and—
Does economics end where life ends, then?
Okay let me clarify. You’ve done work on colonizing space.
I’ve analyzed the equilibrium of a potential wave of colonization of space – I presume that’s what you’re referring to. It looks like the speed of light is the limit on travel speeds. So we have some limits and we can already begin to think about the behavior of our descendants on a wave of colonization that is capped by the speed of light. I did an analysis where I assume that this colonization is via some short hops. That is, the optimal strategy is to land somewhere, grow quickly there, make some new seeds and send them out for a relatively short distance – so a hundred light years, or a hundredth of a light year. Then you see there’s been this wave of landing and jumping, and we think about a wave of contests like that, and the participants cannot even help each other, because they’re too far away from each other, because of the speed of light. So there’s a limit and there are these local competitors. In that context I’ve written a paper where I analyze that behavior, in a mathematical form. For example, in my paper, however many seeds you send out from your ‘oasis’, on average only one of those seeds makes it out to repeat the process. If it’s a billion seeds, only one of those seeds makes it to repeat the process. If there’s a hundred, only one makes it. That’s a robust mathematical feature of this contest at the leading edge.
There’s something there, no? The unique creative opportunities the discipline presents.
What about aliens though?
Obviously, if there are aliens out there, you have to win the contest to meet them!
What would they look like, these aliens?
Not as difficult as one might think. If we’re talking about aliens, we’re talking about creatures that are way way beyond their origin as a simple subsistence animal or forager or industrialist. I would presume that they have thrown out biology entirely, or completely revamped it.
You’d like that, yes.
They are, you know, robots… and the things that you can say about them, I think, will be the things that you can say about us eventually.
Given The Age of Em.
Well, human culture started out in different places with different sorts of music and clothing and local customs, but in a modern economy, I think more than people understand, we’ve converged upon sets of efficient practices. Why should this convergence of efficient behavior stop on a planetary level?
I wanted to ask you if you’re proud of being an American.
Uh, no. I would rather be proud of things I can claim some credit for. I can be grateful for being an American. I suppose the effort I put into staying an American could give me pride, but I don’t think I really do that either.
Is there a problem with America today?
That’s an easy standard. Of course.
There’s a problem, really?
We’re rich, and comfortable, and increasingly aware of that fact. In that context, we are less eager to innovate and bear risks. And we give tremendous veto power to parts of society that would feel like they’re on the losing end of a change. We are more eager to do the things that make us feel good, and make us look good, rather than being, well—
Give me an example – the education system?
There is simply too much education.
Too much. I mean, I see my college students sitting in class, barely assimilating useless general principles. That’s not finding a place in the world. It’s a waste.
It seems to me that if we’re going to posit economics as a specially inventive mode of empirical inquiry, we nonetheless have to keep it from seeming relativistic. How do you keep economics from becoming a morality play? Take inequality. It’s an economic idea, but is so bound with emotion, and ethics. As is education. You call yourself an economist, not an ethicist with a calculator.
Partly because values, at least in part, flow from economics. My one-factor theory to explain as much as possible of the change undergone by society in the last few centuries is the idea that we are slowly drifting back towards forager values and manners as we get rich. Our distant ancestors from a million or two million years ago were foragers, meaning that they would wander around, staying in one place for one, two weeks and then moving on. That’s where humanity evolved, where most of human nature evolved. In that context, humans were highly egalitarian, they shared a lot, they didn’t allow anyone to really give orders, they tried to coordinate, to resist domination. They had rules against bragging, maybe. They were relatively promiscuous, they didn’t have property. This is our human heritage. But roughly ten thousand years ago, farming became possible – but above being possible physically, it had to become possible psychologically, meaning that we had to adopt a large number of changes in attitude. And humans had enough cultural plasticity to make that happen, in part via conformity and religious pressures. And this new farming set of attitudes and habits included marriage and family and war and slavery and a lack of travel and a greater work ethic. And we got used to this and managed to implement it for 10,000 years. And then in the last few hundred we’ve gotten rich! In the context of increasing wealth, we feel these social pressures that turned foragers into farmers less strongly. In the farming world, if you told a young woman that she would have a child out of wedlock, they might starve and die, because society had already arranged itself that way, that would be a valid concern. In our society, women who want to have sex outside of marriage know, actually, that it won’t be so bad if they have a child. And so they’re more tempted to do what feels right rather than what this farming culture heritage has inculcated in them. And that’s a consistent trend through the last few centuries, explaining not only increased promiscuity, but more democracy, less slavery, more leisure. A great many trends can be explained this way.
That’s not teleology?
No. I mean, earlier, in some sense, it might have been, because the behavior would maximize the number of offspring in forager societies, but now that’s no longer relevant, but we can’t do much about it. So that’s where our behavior is moving, not because there’s a goal, but because it feels natural.
And going back to love?
In the farming world, marriage was not usually chosen on the basis of romance. It was more chosen on the basis of co-production, of family, of family based units of production, of keeping the right sort of ties in clans. All of those were important. But we’ve moved back to romance.
Some argue that romance in recent times has actually become even more commodity-oriented. The concept of dating, for instance, has boomed – boomed – with increases in capital, as people who are attracted to one another consummate their bond by experiencing the economy together.
That is true. Centuries ago, when we were near the margin of subsistence, we would choose our partners more by the criteria of co-production. I can agree that co-consumption has replaced that, yes. So we now want people we can enjoy our leisure hours with, that’s very important – we want someone to enjoy parties, events with, someone to like the same products, activities with. As we grow richer, these trends will continue – people will be more consumptive, more promiscuous. Before the Age of Em.
What will the apocalypse be like?
Presumably there will just be a lot of apocalypses in the future. I could see a pandemic. Among emulated minds you wouldn’t really have biological disease, I don’t think, but you could have something analogous, viruses. Wars still apply to emulated minds.
I had a man tell me that the apocalypse will be someone turning our simulation off.
I find it hard to worry about that. Keenly observe whether, in the future, people grow more inclined to simulate the past in a giant, robust way. I just find it hard to believe that people in the future will be eager to simulate creatures like us.
Expensive. We’re not really interested in simulating the whole past. We’re so obsessed with recent eras like World War Two or even the [American] Civil War, but the farther back you go, the less people are interested in simulating life, in movies or acting or academic work.
But it just takes just one guy! Look, with emulated minds, we’re already working with humans as capital for innovation. You don’t see the danger of that dusty historian coming along? That idealist, the obsessed simulator?
I’m sure people will think of grand simulations. It’s a question of if they will be able to afford it. If you want to live in a simulation, you have to consider that it cannot be random. That it has a purpose.
You’re freezing your brain?
I’m signed up for that, yes, to have my brain cryonically preserved. I’m a consumer – as a consumer I’m signed up for that, yes. My rough estimate is that a 5% chance of it working would make it worth buying the product. It’s sufficient for me, I would say there’s a greater chance than that.
Are you afraid of death?
Yeah! (laughs) If there’s anything worth fearing, it’s that. It’s losing what you feel is important.
Well, I don’t think you would really lose it. But what’s important to you, Dr. Hanson?
I took some time early in my life to ask: what was the most important question, and how could I answer it? And I think I found it and answered it. The most important question was: how can we come together, all over society, not just in economics or even science – how can we come together and decide what to believe about what works? And I concluded that it was basically the idea of prediction markets. Trading on the outcome of future events, investing yourself into the future, binding yourself to a community.
And I found myself in the position where I had identified this key question, and, in my view, answered it, answered it in an actionable, real way. I had showed people my work, my theoretical explanations and my lab experiments, my data sets. I explained to them why it’s so important to bet and gamble and trade on the future. And then, to my surprise, I found that the world does not much care.
Fairfax, Virginia, August 2016