Mike: Hello and welcome to episode number 30 of the Future Thinkers podcast. This one we’re going to be interviewing Robin Hanson, the author of the Age of Em. Age of Em is about this small period of time that happens when humans [00:00:30] figure out how to emulate the human brain. He talks about the economics, the science, the physics. He talks about a lot of different things in this book and we are interviewing him today. Robin is an associate professor of economics at George Mason University and a research associate at the Future of Humanity Institute at Oxford University. His blog is overcomingbias.com.
Euvie: Before we get into this episode, I wanted to thank our recent donors and patrons for supporting our show and helping us make it better. [00:01:00] Thank you Connor, Martin, and Donald. Woohoo. For the rest of you guys listening, if you want to help us by sending a small donation or becoming a patron, go to futurethinkers.org/support to find out how you can do that. For any notes from this episode or links to books or anything like that, go to futurethinkers.org/episode30.
Mike: Thanks for listening and enjoy the show.
Euvie: I read that you’ve been working on this book for quite some time. How did the original idea come about?
Robin: [00:01:30] Well, I wrote a paper on this in my first year of graduate school. I started my PhD program in social science in 1993 and my first Christmas vacation, I went home and tried to apply the basic economic concepts that I’d been learning to this unusual scenario. I wrote a paper on that and I published it the next year in an obscure journal. Then I mostly set it aside while I worked on other things for the next 20 years.
Mike: Can you tell us a bit [00:02:00] about your background?
Robin: My background is excessively diverse for an academic. I started out in engineering in college, I switched to physics. Then I went to graduate school in philosophy, then I went back to physics and got a masters in philosophy and then a masters in physics. Then I did nine years of computer research and finally decided to try to take this hobby I had of institution design and turn it into a career. Went back to school to get my PhD at Cal Tech. Then I did a post talk in health policy, then I finally got me [00:02:30] tenure track position here at George Mason in 1999.
Mike: Wow, that’s pretty diverse.
Euvie: Yeah, very diverse.
Mike: What drove all the different fields? What drove you to experiment in the different fields?
Robin: Initially, I was just very self-centred. I was interested in studying what I was interested in studying and when I learned the basics of a field, other fields started to get more interesting. For a while, I just kept studying whatever fields I had the biggest questions in. It took a long time before I finally realized that [00:03:00] if I was going to try to make a career out of ideas, I have to pay attention to the usual career ladder and collect the usual contacts and credentials.
Mike: Can you tell us a bit about the book?
Robin: Sure, the book just out from Oxford is called the Age of Em: Work, Love and Life when Robots Rule the Earth. It’s a lot like a science fiction novel, except there’s no plot and no characters and I’m trying to be really serious and careful about the setting. I take an [inaudible [0:03:26] scenario of technology that might appear, [00:03:30] called Brain Emulations. It’s been around in the science fiction and futurism for many, many decades. I try to take it really seriously and assuming that that technology shows up and is cheap, I try to analyse what in fact the world looks like after that. I try to go into great detail. I fill almost 400 pages full of detail applying everything I know in physics and computer science and economics and other social sciences and human sciences, to fill out a detailed picture of what this [00:04:00] would be like. Applying our standard theories.
Euvie: What made you choose to put economics at the centre of all of it to kind of tie it all together?
Robin: There’s a lot of what’s called hard science fiction out there where people try to take at least the technology seriously and try to get black holes right or starships right or computer viruses right, things like that. There’s much less where people try to get the social science right. [00:04:30] Most people who are interested in tech futurism or science fiction know a fair bit of hard technologies like computers or physics, they don’t really know much social science.
Mike: What about the time scale of this universe you’ve created here? I’ve found that to be one of the more interesting things is it happens in a very tiny scale. What is that scale?
Robin: The scale is roughly a year or two in ordinary clock time.
Robin: We’re talking about emulations of humans that can run at different speeds from human [00:05:00] speeds. So, in fact, I estimate the typical speed of an emulation in this society is roughly 1,000 times human speed. That means to the typical emulations, this era lasts 1,000 or 2,000 years.
Euvie: I’ve seen you describe why you arrived at this 1,000th of a human speed kind of average, but can you explain it to our listeners why you chose that particular number?
Robin: Sure. It’s a trade off between [00:05:30] two factors, both of which are discouraging, are things you want to avoid. One thing you want to avoid is learning a whole career and then very quickly having the world change and having your skills be irrelevant. What you’d like to do is learn a skill and how to do a job and then spend a whole subjective career, perhaps a century or more, doing the job roughly the same way. After that century or two, your mind probably becomes fragile and hard to learn new things and you have to be replaced [00:06:00] by younger workers anyway. You might as well retire at that point.
My best guess for how fast this economy groups, that it doubles roughly every month, and the doubling time of the economy is roughly a time scale on which jobs change, tasks change, and skills need a change. So, what the Ems want to do is fit their career into that month. If it lasts a century, that means that they need to run roughly 1,000 times human speed or faster [00:06:30] in order to fit their career and the doubling time of the economy. If they run much slower than 1,000 times human speed, the jobs change very quickly after they learn how to do the job. That’s a trade off on one side. Of course, that means they could run anything faster than 1,000 times.
They can run probably up to a million times or even faster human speed if they pay more, because in order to run twice as fast you have to pay twice as much to build the hardware and pay for the energy and things like that. That’s on the one side. The other side [00:07:00] of the trade off is, just like us, emulations gain big economic advantages by concentrating into big dense cities where they can then quickly interact with a great many other Ems. But the speed of light is a limitation on how easily they can interact with other creatures how far away.
Today, at the speeds we run, it turns out that our reaction time of a 10th of a second is enough time for light to move anywhere on earth. So, we can be in virtual reality interacting with anyone on earth and not really notice where they are [00:07:30] if we had a good enough communication line. But as Ems get faster, that distance, where they have to be closer than a certain distance so that they don’t notice where they are, get smaller. When it’s 1,000 times human speed, it’s more like 40 kilometres. If you get much faster than 1,000 times human speed, then if even if they have a city of 10s of kilometres around full of Ems, they couldn’t interact with most of them without at least noticing a delay, which would get in the way of their interaction. [00:08:00] They don’t want to be that much faster than 1,000 times human speed, because otherwise that gets in the way of interacting with others in their city.
Mike: This is one of the most fascinating aspects to me about this book, is the time scale differences and how you said in one of your talks, if you’re running as a slow emulation sticking your hands into the pond of the fast emulation, you’ll get bit. I love that. What is the motivation for Ems to do work for normal meat space?
Robin: Through most of human history, [00:08:30] through pretty much all of animal history, most people lived at subsistence level. That mean that when they worked most of the time that gave them enough income to survive and not that much more. That’s how it’s pretty much always been. If you weren’t willing to work, you didn’t exist for very long. That’s how it works in the emulation world, too. They live in a subsistence economy. In order to survive, they need to be working most of the time and they need to be in demand, i.e. they need to be one [00:09:00] of the very best at whatever they do.
If they are the best at what they do, and they’re willing to work most of the time, then they can earn barely enough to survive. If they choose not to, then they don’t exist. Then somebody else does. In order for this scenario to work, all we need is some people willing to fill this role. It could be that the vast majority of humans when they faced the Em world and somebody says, “Do you want to work really hard to make barely enough to exist?” They say, “No, I don’t like that, forget it.” [00:09:30] If the vast majority of humans say, “No,” it can still exist if a few dozen people say yes. Because the key point is you can make billions of copies of anyone of them. That is, all it takes is a few really capable willing people and they can fill this entire emulation economy with copies of themselves.
Euvie: What happens when the cost of technology becomes cheaper and cheaper, and eventually, if we have solar [00:10:00] power and transistors that are really efficient, and it costs very little to run these Ems, why would they want to be spending most of their time working if they can only just work maybe a few hours a week to fulfil their energy needs? Why would they want to do anything?
Robin: This is the standard economic analysis of the supply and demand for labour. On a key assumption, which some people have questioned, is the downwards sloping [00:10:30] demand for labour. That is, when there is only a limited number of smart humans around, they’re very valuable, they’re worth a lot. But as you increase the quantity of them, and make more and more of them, this downward sloping demand for labour says that our value for the marginal worker goes down. We use them in the most valuable uses first then, as the supply goes up, as the quantity available goes up, we allocate [00:11:00] them to less and less useful tasks and, eventually, there are just so many of the so cheap and available that you allocate them to really quite low value tasks but still, if the value of the task is higher than the cost, you still go there.
The key assumption here is that labour demand curves slope down and they keep going down to very low wages. That is, when we have six billion humans, each one is very valuable. But when we have trillions of them and we don’t in proportion [00:11:30] expand all of the other things we have in society like roads and buildings and machines etcetera, then having enormous number of workers with very limited other inputs means those workers on the margin are just not worth that much.
Mike: It interests me that you’ve gone with this being one of the assumptions here that we would never have an emulated socialist economy. Is that an impossibility in your mind?
Robin: I’m an economist, so I’m using the standard economy [00:12:00] tools to make the usual standard economic assumptions. So, I want to distinguish what the usual standard assumptions are from what we, as economists, conclude. Our standard first cut analysis in almost everything is supply and demand. That is, we assume, because it’s relatively easy and it’s not that bad an approximation often, that there’s a lot of potential suppliers, a lot of potential buyers, and they each don’t think they can influence the overall price that much.[00:12:30] Their costs are relatively local, in the sense that they don’t need huge organizations to supply everything that takes over the entire market. So, supply and demand is our usual first cut analysis and supply and demand implicitly assumes that the suppliers and demanders own the things they can sell, and that they sell them for something like money so that they can then buy other things that they want. So, it assumes a market economy and it assumes relatively low regulation in the sense that people can sell at whatever prices [00:13:00] they choose. That is the standard assumption and it’s what we do to analyse most everything as a first cut.
Now, it’s not to say that we should not ever have regulation or that there are not substantial deviations sometimes from supply and demand. After you’ve taken your first semester of economics, you will hear in great detail if you go on to all of the complicated details that we go into in terms of how there could be less than perfect competition in many ways and many kinds of market failures, as we call them, and many potential regulations and many potential [00:13:30] ways that you have as alternative production. So, all of that is possible and can be applied to the scenario I’m talking about. I am doing the first cut, which is supply and demand.
I would say we have a lot of literature on, if you want to regulate something, what it takes and what the problems are. If there could be a free market on something but instead you’re going to regulate that market then you have to worry about black markets, for example. Can you prevent people from buying and selling? How well can you monitor their buying and selling? [00:14:00] How much can you control the production, is it something that millions of people can separately produce? Or, is there a big central factory we can go in and put soldiers around the factory and say, “We’ve taken this over and now it has to be done our way.”
You have to think about those things if you want to think about regulating. Now, in general, regulation has a limited ability to change what would otherwise be the market outcome. Free drugs would be relatively cheap and you regulate and prohibit drugs, you can definitely raise the price of drugs, but you can’t necessarily raise it arbitrarily. The higher you raise of drugs [00:14:30] relative to what the free market price of drugs would be, the more the people create black markets to evade your regulation and the harder it gets to find all those black markets and squelch them if there’s really a huge demand for the product you’re selling and the cost is actually relatively low.
This world I’m describing of Ems is so very far from the world we live in that even a highly regulated version of it would still be pretty far from the world we live in. So, you could take this world of Ems and imagine whatever the wages are, now imagine we restrict [00:15:00] the supply of Ems and regulate their wages so drastically the wages become 10 times what they otherwise would be. That’s a pretty dramatic degree of regulation. You would really need global regulation and would have to be quite intrusive to monitor everyone and make sure they aren’t violating rules in order to make the wages go up to 10 times what they otherwise would be. That could still be conceivable though. Even that is a world that’s really different from our world and isn’t that different from the world I’m describing in this book.
The key point is [00:15:30] if you want just a first cut idea of what the world looks like that’s very different from our world, supply and demand is a decent first cut. Then with regulation or other forms of government production or whatever, you could move modestly away from that, but you can’t move it all the way back to where we are.
Euvie: What if we go in the other extreme and we have no regulation at all and these emulations are able to innovate extremely fast if they have the processing power, and they’re able to maybe [00:16:00] mine asteroids or go out into space so they don’t necessarily have to participate in a job economy. I mean, they’re still doing things and they’re able to fulfil their energy needs. Do you see that as a possibility?
Robin: A low regulation and economy would probably grow faster, that is, it would more quickly innovate, as you say, more quickly move into new territories both physically and in product spaces. A fast-growing innovative economy could [00:16:30] more quickly change and become something different. The key point is how fast can the population grow. As long as the population of Ems can grow very quickly and can grow faster than the economy can grow, still the wealth per Em must fall. This has been the case for pretty much all of history again. Pretty much all of history we’ve had growth, the number of humans grew and humans developed new technology etcetera.
It was a slow growth in the past [00:17:00] but it was still slower in the past than we could grow populations. Whenever we invented new things, most of that new wealth that we created went into new population, went into having more of us. And it was only in the last few hundred years that we’ve been able to grow wealth so fast that it’s been faster than the rate at which we can grow people, so that the wealth per person has been increasing. The trend towards increasing wealth is a very robust trend, we should expect that in the future. But the fact that we can grow wealth faster than population, that’s more [00:17:30] a temporary affect of the fact we haven’t just haven’t had good population technology for a while.
The technology by which we grow more people hasn’t changed much in a very long time. Whereas, the technology by which we create wealth has been changing enormously. The world of Ems is a world where the technology of changing populations suddenly changed. That is, you’re suddenly able to make far more substitutes for people very fast in factories.
Mike: I’m interested to know from the side of the emulation, perhaps the artificial intelligence side, what makes you believe [00:18:00] the brain is something that could actually be emulated in the future?
Robin: Well, as I mentioned, I started out in physics long ago and physicists have a totalitarian view of how physics explains the world really. Physics just isn’t a set of theories for an obscure set of odd processes you could do in the lab. Physics is a theory of everything around us and we have been quite successfully really in understanding everything around us, everything in your office, [00:18:30] everything at your home, everything you’ve ever seen, really in terms of a standard set of physical theories explaining what they’re made out of and how those parts work. I buy that. I think they’re right. Physics does explain the world we see.
Now, there are some strange particles that you can make, strange things that happen very far away that physicists are still puzzling over. But the actual world around you, we’ve nailed that. We pretty much know how that works. That includes not just your cars and your food, that includes you, your body. [00:19:00] We’ve looked inside your bodies and we’ve seen parts and what they’re made of, it’s the same parts as everywhere else. We’ve looked inside your brain and we’ve seen that your brain is made out of exactly the same sort of parts as everything else, interacting in the same sort of ways.
We think we understand how those parts work at the low level and how they interact and what they’re made out of. So, your brain is a complicated arrangement of those things and the hard part in understanding your brain is understanding all that complexity. But there are parts we do understand. [00:19:30] Can we emulate a brain comes down to can we figure out how parts work well enough to emulate them. The traditional artificial intelligence approach is to sort of watch what people do and create a mental theory in our minds of how that works and then write code that expresses that theory and see if we can capture their behaviour. The brain emulation approach is different in the sense that it’s trying to support the software that’s in the brain.
So, what we try to do is [00:20:00] we make computer models of how each type of cell in the brain works, in terms of taking signals in, changing internal state, and sending signals out. If we can figure out how individual cells work, in terms of signal processing, and we have a good enough scan of a whole brain, in terms of where all the cells are and what type they are and who’s connected to what, then we should be able to have a good enough model of the whole brain. Assuming, of course, physics is right. That is, your brain really is made out of these cells and they really do [00:20:30] interact through standard physics.
Mike: Yeah, I think it’s actually brave in the assumptions you’ve made to write this book. I think it leaves you open for criticism but it’s entirely necessary to create any vision of the future. You said, “We need to envision more possibilities of the future.” Why do you think we need more versions of Star Trek and that sort of thing?
Robin: Well, there’s lots of interesting things to study in the world but we have a huge field of history which has many thousands of researchers, [00:21:00] many thousands of books actually if you look on amazon.com for keywords, 20 percent of the books have a keyword history. We have enormous studies in history and, of course, plausibly history is important and interesting to study. But the future is also important, plausibly more important because we can actually do something about the future. The past is too late to do anything about. Plausibly, we should study the future even more than we study the past.
In fact, there are very few people who are studying the future. Honestly, an awful lot of [00:21:30] what goes under the name of future studies is more wishful thinking and inspirational speeches and that sort of thing. Advocacy for values indirectly using the future as a way to talk really about today. Which are all fine things for people to do but they mean there’s even less trying to think about the future seriously and the way academics think seriously about the past.
Mike: Yeah, there seems to be a big tendency to latch onto whatever world model works for whatever individuals studying in [00:22:00] a narrow field. It’s interesting that you’ve combined so Many fields and thoughts so deeply about this. I’m interested to hear actually more about the virtual reality world that you’ve imagined.
Robin: Emulations are running on computers and the cost of running animation brain is actually pretty large compared to today’s cost, but the idea is that eventually that’s a low cost. But the cost of creating virtual reality for someone like us is already pretty low, people already experience virtual reality worlds and [00:22:30] video games, and they can tell they aren’t real, of course, but they are real enough that they can be productive and they can interact and work in. Since the emulations are on computers, it’s straightforward to put them in a virtual reality, they don’t need to get dizzy or anything like that they can just have a perfect connection there between their brains and the virtual reality world.
It seems like it would make complete sense to put them virtual reality when they don’t need to interact with the virtual world. Now, most jobs in our economy are desk jobs. So, most jobs in the [00:23:00] Em economy are plausibly desk jobs, that means there’d be little point in giving them artificial desks to work at. Might as well put them in a virtual reality and have them see a spectacularly luxurious office environment. In virtual reality, it’s cheap to make something luxurious and beautiful, it doesn’t really cost any more than it does to make it ugly and grimy, so they might as well.
Mike: Work under the sea today, work on Mars tomorrow.
Robin: Right. They are working most of the time. That is a [00:23:30] key element to this economy. When they’re working, the virtual reality can be beautiful but it needs to not be too distracting. So, that’s a reason why it wouldn’t really be that dramatic if you were to see it. It would be pleasant, it would be nice, it would be comfortable. When you first stare at it you might be amazed and stare at it for a while but it would be the sort of thing that, after a while, you could not stare at. You would be okay with not paying attention to it, it would be in the background and you could focus on your work.
Now, when you got off work [00:24:00] and went to do something in leisure then, of course, virtual reality could be a lot more dramatic and demanding and just make you stare at it and make you pay attention to it because it was so interesting. That would be fine during leisure times but that’s not so good at work.
Euvie: I wanted to move into a little bit of the social aspects of the emulation world. It seems that a lot of the human interaction is based on drives for mating and social acceptance, which is essentially [00:24:30] the more social acceptance you have, the more likely you are to have food in a tribal society, and the more likely you are to have better selection of mates. So, other than some sort of work transactions, what would be the reason for Ems to have these interactions with each other and to have any kind of a social identity or want to be perceived a certain way?
Robin: Emulations are psychologically very human. This is the key point. We take a human brain and we create and emulation [00:25:00] of it and then we put it on a computer, but when we turn it on you have to convince it that it’s on the computer. The last thing it remembered it was an ordinary human. So, psychologically it’s very human. That’s why it’s possible to think about the Age of Em and to think about all the details of it because we can take humans as they are and use those sorts of psychological tendencies to think about how the Ems would behave.
They will choose to exist and have friends and lovers and all those things for the same reasons [00:25:30] we do and, of course, we’re not always that aware of why we do them. But, as you say, we have these robust patterns that we’ve seen that people do want to have lovers, they want to have friends, they want to have respect. So, they do things to gain these things. Emulations would do that, too. The key is that the world changes somewhat. So, starting with lovers, say, humans have a very deeply engrained desire for attention, respect, sex, pair bonding, and Ems would have that, too.
They might reduce it somewhat [00:26:00] through the analogy of castration or something, but most likely it’s still pretty deep and they still want it. Now, in our world, reproduction has been tied to sex and so even though we enjoy sex and we have a lot of desires about it that have little to do directly with reproduction, that’s the origin of it. So, traditional societies have regulated it and tried to control sex exactly because it was tied to reproduction and reproduction was a key element to the society.
For emulations, [00:26:30] sex isn’t tied to reproduction anymore. That is, they reproduce by making copies and not via sex. So, sex loses its central place in how things are done in society, therefore, there’s less of a need to regulate it. The emulation doesn’t need to control it and make sure things don’t go wrong too much because reproduction is done other ways. In other sorts of reproduction, they probably would have more controls over and more moral sense and more rules and things like [00:27:00] just how they reproduce, but sex itself would not.
Similarly, we like to have friends, and presumably emulations have friends. Humans are very social and that’s one of the key elements of human nature is that our distant ancestors had these large social groups and we had intricate politics and social relations in these large groups and that’s carried over, of course, through all the eras we’ve seen. Farmers have complicated and intricate social relations, so do we in the industrial era. Ems would probably do that, too. In modern [00:27:30] workplaces politics is essential for most workers. That is, whatever your job officially is, one of your sub jobs is to socialize with your co-workers and your boss and others out there to gossip and find out what people are saying and try to help your allies and dis your rivals.
Our human political nature is on full display in most office places and people who just keep their head down and ignore office politics often lose big because they [00:28:00] lose out in the office politics to others who managed to make it work for them. Most likely, emulations will need to be political, too. They can’t just ignore people, they will need to create friends and alliances and gossip and find out what people think of them, say nice things about their allies and dis their rivals, change their alliances as necessary in order to stay on top of local politics. Those are things that humans have done all through history and Ems would probably continue. Right there is a [00:28:30] functional reasons Ems need to be social, because otherwise they get ignored.
Mike: I think it’s fascinating, the direction you’ve gone with the copying of individual Ems and I think some of the opportunities to develop as an individual would be really multiplied from that. You know any time you see a picture of yourself or a video of yourself and you pick up on something you’re doing unconsciously, or you weren’t aware of? But in this world, [00:29:00] you get to actually watch copies of yourself behave and interact in the real world. Like you said somewhere in the book, too, you wouldn’t have to speculate anymore how your life would be different if you had have made a different decision, because you’d be able to watch an Em do exactly that. Do you see other ones like that?
Robin: What happens when there’s other copies of you out there and how you feel toward them has been one of the hardest things to try to envision in this scenario. That’s one of the things that’s farthest away from our current experience. I’ve struggled with that and tried to come at it from a number of different angles and [00:29:30] done the best I can. I’ll say that still I probably didn’t get it that right. One of the things we can be sure of is, yeah, there’ll be a lot more versions of you around that are just a bit different from you but still pretty much you.
That is, they were you as a child, they went through the same training, then, at some point, there was a divergence. They studied for a different career, they moved to a different city, they had a different spouse. Or even more recent versions, they were trained for the same jobs, even do the same sort of equipment, but now there’s a copy of them [00:30:00] a few miles away in a slightly different context and they’re all around available for you to learn from if you want.
Now, you might get too discouraged by paying attention to them because you might not feel unique as much anymore and that the world doesn’t care much about you because, hey, there’s these other versions of you around. Honestly, people are able to ignore that in our world. We aren’t really that different from each other in our world, but we manage to feel unique anyway.
Yeah, you could learn about what happens with other copies of you who have the same spouse, [00:30:30] what do they do and how does that work. Other people with the same job, other people with the same allies or different allies, same hobbies. You could get statistics from these other versions of you who live pretty similar lives and also older versions of you. You could see what someone a lot like you is like five years later because they’re right there to look at.
Euvie: Do you think that the Ems would be able to see their own programming and alter it?
Robin: In principle, today in software we have a phenomena called software rot. [00:31:00] This is a phenomena whereby when systems get old, they’ve spent a long time adapting them to particular circumstances and adding new modules and changing features etcetera. In that process of changing and adapting to circumstances, they often just become very hard to change. Now, with software, you can go in and look at any one line and you can type in new characters and change that line. It’s not that you can’t change any part, it’s that you don’t know how to change it usefully.
The older the software [00:31:30] gets the more ties there are, we often call it spaghetti code. The more that different parts just get tied together, the more you try to change any one thing the more other things you’d have to change in order to make that first change work. Eventually, it just gets a big pile of spaghetti code that’s very hard to change. Today, what we do is we just throw it away and we start over from scratch. That’s what we do when code gets too old and fragile like that.
The code that’s in your brain is pre-existing [00:32:00] spaghetti code that evolution put together. It didn’t document it, it didn’t try to make it understandable to you, so it’s pretty hard to understand. The first cut assumption is that when we first are able to make emulations, we just don’t understand this code and mostly, we can’t change it. Now, we’ll just try turning some knobs and we’ll probably find a few hundred or even thousand like little things we can change that makes some sort of difference and that’ll give us a menu of different ways we can change you to make you more focused or more optimistic or whatever else [00:32:30] we want to do.
But that’ll be a limited menu of changes. Beyond that, it’ll just be this opaque mess that we don’t understand how to change initially. Eventually, of course, we’ll figure it out. Then, eventually, we’ll be able to make more substantial changes. But plausibly, there’s an era before that I can talk about. My book is focused on this early Em era when we don’t know how to make these changes, because once we do know how to make these changes it’s just much harder to predict what happens. Until we know what the actual changes are, i.e. what the [00:33:00] modules are and how you break them down, what parts can work by themselves without other parts, it’s hard to say very much about what we would do with that ability to change minds.
Mike: But it goes to say with that ability to shrink time and live out 10 years in the span of a couple of days or whatever, that those solutions would come nearly instantly.
Robin: I’d say the right metric to be using is doubling times of the economy. Our era, and for the last few hundred years, the economy’s [00:33:30] been doubling lately every 15 years. During the farming era, it doubles every thousand years. During the forager, it doubled every quarter million years. But each of these eras has actually seen a similar number of doubling so far. So, that’s plausibly they’ve seen a similar amount of change even though the time scales are vastly different. Also, plausibly during the Em era we also see a similar number of doublings.
I’d say however fast you think would be changing now, [00:34:00] scale that in terms of doublings. How many doublings will it take to make some change? Then as the Em era speeds up, you should think in terms of, “Well, because the doublings are coming faster, change is coming faster.” That’s a good first approximation. Now, actually in the Em era growth will more happen because of just a raw increase in the amount of physical capital as opposed to innovation. Whereas, in the past eras most change has been due to innovation. That says that the amount of innovation per doubling is a bit less in the Em era than it’s been, but [00:34:30] still within a factor of two or so. It’s about the same.
Euvie: Can you explain why that is? The less innovation.
Robin: During previous eras we’ve just always had some input that’s been limited where we just couldn’t expand it. In our era, it’s been people. That is, we can make machines really fast, we’ve got plenty of real estate and plenty of raw materials, but we just can’t expand people. So, there’s what’s called diminishing returns to expanding everything except one key input. [00:35:00] You really want to expand everything together and then you can double everything. If you took a world like ours and you made an entire nother planet like ours, exactly the same, then that planet could, of course, produce everything that we produce and then together the two planets could be twice as much as one planet.
If you just double all the machines on earth without doubling the people or doubling the land or things like that, you don’t double the output, because you’re missing some of the key inputs. So, during our era, growth has been limited because we can’t grow people very fast, [00:35:30] that’s what’s been forcing us to grow innovation. Now, during the farming era they could grow people as fast as they wanted and they could even grow animals as fast as they wanted, but they couldn’t really grow land as fast as they wanted.
Land to use was a limited resource and so in order to be able to use land more efficiently, they had to innovate. That was partly what limited the farming economy from growing. The emulation economy, plausibly, at least for a while, [00:36:00] it can grow the machines really fast, it can grow the substitutes for people, i.e. the emulations very fast, and for a while it’s just not limited in terms of raw materials or real estate. There’s plenty of that around and so, for at least many doublings, it can just grow by just making more things in factories.
Now, making and using things does cause innovation, it’s called learning by doing, so plausibly there’ll still be a lot of innovation. It’ll just be less as a fraction because now we have this new way to grow that we haven’t had until recently [00:36:30] or even now, which is just cranking more stuff out of factories.
Euvie: It’s interesting to think about innovation as just another factor in all of this, I’ve never really thought about that before actually. I wanted to ask you about the experience of Ems, how they perceive themselves. If they’re based on a human, they will probably be pretty close to how humans experience themselves, at least at first. At least until they’re able to edit their own code. So, where do you see it [00:37:00] going? What’s the vector here in terms of their personal experience as this system evolves?
Robin: Just in terms of how they compare to us. When we look backward, to our ancestors, we tend to highlight the commonalities. But I think if our ancestors were looking forward to us, they would see the differences. We don’t really have any other ancestors as choices, so we feel pretty obligated to pick our ancestors and identify with them, but when you look forward you can imagine different kind of descendants, [00:37:30] or even descendants who don’t change. Therefore, you’re less primed to just accept whatever descendants you have as like you. I think emulations, similarly, they will look back at us and focus on the similarities. They will see, yes, they have changed in some ways, but they will see how many different ways they have not changed, how they are still in common.
Just like we look back on our farming ancestors or our foraging ancestors and we see how they are very like us in many ways [00:38:00] and we identify them. But honestly, if they looked forward to us, if you could take a subsistence forager – there are still few of them out there in the margins of our world – and you show them our world and us, they will not so much see us as an obvious continuation of themselves. They will more see us as aliens or strangers. We are weird. We are different.
We looking forward to emulations, we see them more as different because we don’t see that we had to become emulations, we can imagine we [00:38:30] might not change so we imagine the emulations as being different. I think that explains part of the different attitudes. I have been describing, and the book describes more many features and emulations, some of them are familiar and some of them are strange. So, I think you and most readers probably are not sure how much you want to identify with them and how much you want to think of them as the same species or as some sort of an alien.
But I’m pretty sure when they look back to you, they will see the commonalities. Now, even if they’re seeing the commonalities though, [00:39:00] that doesn’t mean they think they’re equal to you. Honestly, if you look inside your own heart at how you think about your ancestors, you feel better than them. You think you’re superior to your ancestors. Now, maybe you don’t blame them for learning what they were taught, they did the best they could given their world, but you think you’re better. You think your ways are better, you think your attitudes are better, your technology is better, your society is better.
And emulations will think that way about you, too. They will look back on humans with perhaps some gratitude, some nostalgia, [00:39:30] some sense of origins. But they will think they’re better, and there’s a lot of ways in which they’ll be right. They are better.
Mike: What kind of variables do you see that need to happen between now and the period that the emulations begin and your book begins? What are some of the more important key innovations or things that need to happen in that period?
Robin: There’s three key technologies that are required to create emulations. All of them are on trends to plausibly [00:40:00] be ready within roughly a century, but none of them are close to being ready. One thing we need is lots of cheap [inaudible [0:40:07] computers. There are standard trends and, on those trends, plausibly well within a century we’ll have enough of those. But they’re not ready yet. Another is we need high enough resolution, cheap, fast brain scans. Initially, these will be destructive. You take a brain that’s been frozen in some way fixed, then you slowly slice off a layer, you scan [00:40:30] the next layer and find spatial and chemical detail to see exactly what’s where.
You record that all in a big database and that’s the scan. We actually have decent scans. We’ve done, for example, a mouse brain at a remarkable resolution. It’s not a good enough chemical or spatial resolution quite yet, but it’s still pretty good. A human rate is 1,000 times bigger than a mouse brain, so we still have a ways to go there, but we’re on track of getting there in a few decades. The third thing we need [00:41:00] that’s somewhat harder to predict is computer models of all the types of cells in the brain. We don’t even know how many types of cells there are there really.
We have some decent models for a few types. People have made models in their computer and they’ve compared that to how these cells seem to work in the lab. It looks like fairly some of the cell types we know roughly how to model them. But we’ve got to do it for all the different cells in the brain. That’ll take a while. So, the computer industry is on the first track is a huge industry and there’s just not [00:41:30] much we can do to influence that. The scanning industry is actually pretty small, so you could accelerate that if you put in more effort. But honestly, it’s probably the thing that’ll be ready first so there’s not so much point.
The third input, which is the cell models, is the thing that it’s a part of academia where people do these cell models but it’s still pretty small. We could put a lot more work into that and then that could speed up. That could make a substantial difference, depending on, [00:42:00] of course, how hard that jobs turns out to be. We don’t’ really know and the basic thing about modelling is you don’t really know how close you are to having a good model until you have a good model. When the models are not good enough you just don’t really know how far away you are. So, one big industry and two academic research areas.
All three have to progress to sufficient levels before Ems are possible. They’re still a long way away. Probably far enough away that it isn’t worth any business venture trying to start it up now. But at some paint, say, within 10 years [00:42:30] before this is feasible, it will be worth making a business venture and at that point people will spend billions of dollars to try to pursue this because you can make trillions of dollars by being one of the first to be able to field this technology.
Euvie: Being an Em farmer.
Mike: Yeah. How is studying this subject and writing this book changed how you plan for the future and your personal life? How have they affected that?
Robin: I’m not sure it’s changed my life that much. Again, I wrote this [00:43:00] first paper back in 1993 and some of the main policy implications were just obvious then, one of that being people need to diversify their assets so that they own something other than their ability to earn wages. That is the main recommendation for most anyone trying to survive this sort of transition. Plausibly, it doesn’t happen for a while but when it does happen it may happen relatively suddenly. Trying to wait until you hear about it and then doing something may be a mistake. May be just wiser [00:43:30] just to set something up and then not have to deal with it for yourself, your children, your grandchildren, just be in the habit of being ready for it.
That was obvious decades ago. If you want to succeed in this world, which is a high bar because most emulations are going to be copies of the few hundred best humans, so if you want to be one of these few hundred best humans, you want to think about what does that take. I had some initial ideas, again, 20 years ago, but I think more recently I do understand better what [00:44:00] the demand will be for. Honestly, very quickly the demand moves from the most productive in the human economy to humans who are just very capable and young and flexible and have the potential to learn how to be the best in the emulation economy.
Because this is likely to be based on destructive scanning initially, we have the scenario that early in the emulation world, the emulation economy starts to approach human parents [00:44:30] with promising young children and saying, “Could we please destructively scan your child so that we can turn them into an Em because they have the potential of being a successful Em in our new world.”
Robin: That’s a pretty dramatic conflict. So, you should think about your position on that and be ready for it, because probably most of the successful Ems will be those descendants of these children who were scanned in early and then turned out to be [00:45:00] well adapted and well suited for this emulation world.
Euvie: Yeah, the ethics of all of this is a whole other ball game.
Robin: Right? Ethics is awkward because most societies, the ethics has adapted somewhat to their circumstances. That is, our ethics in our society is somewhat dependent on the world we live in and is different from the ethics that our ancestors had in their societies. That makes some sense but then it makes it harder for us to think about the ethics [00:45:30] of a different society from ours. We can say, “Well, by our ethics, what they might do is something we disapprove of.” But, of course, they might change what they approve of and they might think it’s okay.
So, in my book I’ve tried to mainly focus on what I think they will think rather than what I think about whether what they’re doing is right or wrong. That seems the right sort of thing to do for the purpose of my book, which is just to describe this era as faithfully as I can. It’s much easier for me to present myself as [00:46:00] an expert on the positive consequences, i.e. what’s likely to happen if we did little than it is as an expert on what should happen. It’s much harder to be an expert on what should happen. I may have opinions about that but I don’t want to confuse readers and listeners with thinking that that’s the expertise I’m presenting.
I want to separate and distinguish that. Emulations probably, for example, will be more okay with what we might think of as dying. [00:46:30] For us, death is a terrible, huge cost and it’s a terrible moral outrage, so we are greatly offended by anyone who would not only kill someone else but even kill themselves. We’re not comfortable with suicide. Emulations though, they’ll have huge financial and other incentives to be used to an easy come easy go. For an emulation who lasts a whole work day, they may work for 8 to 12 hours and then they need to rest for 12 to 16 hours before they’re ready to work [00:47:00] again.
That means only a half to a third of their daytime is spent working. But at the beginning of the work day when they’re reading and rearing to go, they could make a lot of copies of that version of themselves and then those copies can work for a few hours. Then if you end that copy and then erase it, for that copy, pretty much all of it was spent working. So, it’s a factor of two to three more efficient in terms of productivity. Now, it doesn’t remember or learn as much from [00:47:30] that day’s work, so when learning is important then you don’t want to do that so much. On the other hand, it doesn’t get older and more fragile with that experience either.
So, I think and predict Ems will mostly be okay with quite often splitting off many copies at the beginning of the work day and having very few of those copies continue onto the next work day. They might be erased or perhaps retire to a much lower speed. For many people in our world, that sounds like a horror, that sounds like a violation [00:48:00] of their moral rules and something they wouldn’t possibly tolerate and would never want to go in for, but I predict that emulations will be okay with it. For them, it’s not a big deal. They will have tried it out many times before. It seems like it works okay and they’ll just be doing it.
Mike: I suppose when it becomes commonplace, it’s not going to be so much of a big deal anymore.
Euvie: Yeah, if you can backup all the experiences and the data from their day and just feed it into the main copy that goes on, then maybe there’s just not a big deal.
Robin: Right, if you could merge the copies, [00:48:30] so you merged all of their experiences and memories, there would be much less of a conflict. However, I think during the emulation era, it just won’t be possible to merge in that way. What would be possibly is that at the end of the work day, you could just report back. Each of these copies could go back to the original and say, “Hey, this is what I did today, this is what I learned. This is things to watch out for.” They could give a summary report. They could even have recorded the whole day on video, or they could write a summary up.
They could be archived so they’re available if there’s ever any issue. If two weeks later you go back, “What happened [00:49:00] with that client? What did I do then?” You could go back to that archive, revive it, and ask it, “What happened? More detail.” But still, you wouldn’t be fully merging their memory as if you had done it yourself.
Mike: You could totally imagine a narcissist having a relationship with himself or herself. Having this whole circle of friends and relationships and stuff.
Robin: I actually expect that there’ll be somewhat of a disapproval of people who are too into themselves. You can imagine [00:49:30] a whole city of George’s where George does all the jobs in the city of George and George has lots of relationships with himself of many sorts. It’s all about George. I expect other Ems would find that a little creepy. In most work groups, there’d be just one copy of any one person like George or Tim or Sally. You would know there are these other versions of yourself out there, but you wouldn’t actually interact that much with them most of the time. You would be more in a familiar interaction with other people who look and act different than you.[00:50:00] But you could, as you needed and wanted, go visit and talk to these other copies of yourself, because you trust them more than you trust other people and they are a unit then of training and finance and law and politics that you’re aligned with.
Mike: What starts to get blurry near the end of this book where you’re not predicting so much anymore, feel it gets too hard to predict?
Robin: The book is ordered in terms of kinds of disciplines. So, I start with physics and computer science, [00:50:30] then I move onto basic economics, then I move onto organizations and sociology and psychology. In terms of that order, I’m moving more away from fields that I have a strong education in towards other fields that I have to brush up more quickly and then perhaps just less understand the fundamentals of. I’m also moving from fields where we have really solid deep theories to fields where we just have a large patchwork of overlapping theories, none of which we rely on that heavily.
So, that’s also [00:51:00] places where it’s harder to draw conclusions. The deeper the theory, the more you can extend that theory well outside of the range of data you had to support it. So, there’s more worry when we have this patchwork of little theories that they’re more tied to the context of the world that we live in now and wouldn’t work in this next world. Of course, at the end of the book I’m also talking about the harder problems of how do evaluate all this, what other variations might be possible, where have I gone wrong. That sort of thing. How should you try to survive and succeed [00:51:30] in this world? Those are just harder questions but I figured I shouldn’t ignore them.
Euvie: Before we finish up this interview, I wanted to throw a curveball. Since in this early emulation era the emulations would act very similarly to humans, they would feel very similar to humans, they might not be aware of their code, they might not even realize that they’re in a virtual reality. So, recently there’s been a lot of key scientists coming up with the idea that the simulation hypothesis is actually very likely, [00:52:00] that by certain calculations it’s quite probable that we are currently living in a virtual reality. What’s your view on that?
Robin: I actually have a paper from very early on when the very first paper by Nick Bostrom came on and got a lot of press about the hypothesis, I immediately wrote something at the time about how to live in a simulation. That is, assuming that there is a simulation, how should that change your actions? Somewhat remarkably to me, there’s very little interest over the years in that question – [00:52:30] what to do different. There’s enormous interest in could it happen and is it true? I think that’s analogous to the emulation era book in the sense that for many decades people have talked about, “Are emulations possible?”
And if they were possible, when would they happen? Would they conscious? Would they be me? There’s been almost no attention to, “Yes, but what would the world actually be like?” Again, in the same way for simulations, there’s very little attention to, “Yes, but okay, if you’re in a simulation how should you act different? What is your world like in that sense?” So, I think it’s an interesting hypothesis. I think on that, [00:53:00] it’s hard for me to really believe that the future’s going to be that interested in making simulations of us. I have to get a mildly low probability that I am in a simulation, or you are.
But it’s not zero. So, I certainly think it’s something to consider. But people I think are way to quick to assume that it doesn’t matter, that is, you should still do exactly what you do before, regardless of whether you’re in a simulation. I think that’s just wrong. The key point is that most simulations are not [00:53:30] global simulations of the entire universe. Most simulations we’ve ever made, and probably will make, are affected by the cost of a simulation. The bigger and wider the scope of the simulation, the more it cost to run it. The longer lasting time and the more people that are involved and the more the universe is involved in the simulation, the more expensive it is to run that simulation.
Assuming that our descendants have any remotely similar costs we do, they will then be more willing to run small simulations than big ones. So, the first implication is that if you’re in a [00:54:00] simulation, you’re probably in a small one. What does that mean? That means there’s not much point in saving for retirement, you’ll never retire. There’s not much point in sending donations to Africa, Africa doesn’t exist. This larger world you think you’re in, it’s mostly not that big a world. The small world around really is the real world and the rest of it is illusory.
So, that means you should pay attention to the smaller world and focus on it. There’s also probably a reason for the simulation. You may not be it. It might be some key pivotal event [00:54:30] that you’re near, or a key pivotal person nearby or interesting person. The simulation will probably focus on that pivotal event or person, then to the extent that you move away from that pivotal event or person. You might be dropped from the simulation. You should think around you and ask what are these plausible, pivotal people or events that you could be here because they are here and try to stay near them.
Euvie: Move as close as possible to Elon Musk.
Robin: [00:55:00] If you happen to know him and he’s a plausibly reason the simulation could be run, then yeah. Don’t lose touch.
Mike: Are you familiar with the game that’s going to be coming out in the next few months called No Man’s Sky?
Robin: I’ve seen some press about it, it looks pretty.
Mike: It’s based on the same idea as Minecraft was where everything is procedurally generated and there’s an algorithm that, depending on your location in the galaxy, the whole universe is created around you.
Mike: I thought that was pretty interesting [00:55:30] as it actually fits in some people’s theories about the simulation hypothesis that even in physics when you observe something, its state changes compared to when you don’t. That’s the same sort of way this video game works.
Euvie: Right. Everything is a seed until you, as a player, go and observe it and then it generates based on a certain algorithm.
Robin: Even if our descendants, say, spend an hour a day on average in simulations online. That’s a pretty big assumption, right? They really like to be in simulations, they spend an hour [00:56:00] a day in simulations. The question is what fraction of those simulations time are they spent simulation us. Now, for an analogy, I’d say think about the Roman empire. Today many people play video games and many people play act. They do role playing and they are in plays, they even fantasize.
If we just added up all the fantasizing and role playing and play acting and video game times, what fraction of all that is spent simulating the Roman empire? Ordinary people in the Roman empire. It’s pretty small and it’s honestly probably [00:56:30] substantially smaller than the number of actual people there were in the Roman empire. Of course, our descendants will be even larger than us, but will they become fascinated with somebody in the Roman empire or us? I think that’s just too easy to assume that we’re so fascinating that, of course, the future will be all over us.
Maybe there’s a few people in our society, so maybe Cesar will get enough simulations that Cesar should have thought he was in a simulation. For most random people in the Roman empire, they should probably believe they were in the Roman empire, because [00:57:00] descendants aren’t really that interested in simulating them.
Euvie: What if they’re not pre-programed simulations, they’re just a seed and then it goes and does its own thing? Like an Em. You upload the brain into the computer and then you maybe program it to do certain task, such as think you’re living in a biological world and you have to procreate and you have to work, and then it just goes and does its own thing based on those parameters.
Robin: Remember the emulation world is a subsistence economy, so they cost [00:57:30] of running an emulation is itself a large fraction of any one worker’s wages. Now, the cost to emulate any person in history is a similar fraction to the cost of running an emulation. It just can’t be cheap to run an emulation of a full mind, no matter how far in history that was compared to the cost of running an emulation. Any person is expensive to run. Not, it’s easy to have a seed that has the potential of a person, but it’s expensive to actually use that seed. [00:58:00] Even in this new video game, it’ll have this vast galaxy of potential places to visit, but most of them won’t actually get visited.
Robin: Most places on most of those planets will never actually be seen, they’re all potentially there in the seed but, in fact, because only so many people play the game for so many hours, only so many actual places will turn out to be real.
Euvie: I wonder, this actually ties into the idea of the Fermi paradox. If the Drake equation is correct, and [00:58:30] the probability of other intelligent civilizations out there is actually very high, why haven’t we encountered any of them. One of the answers for the Fermi paradox is that we are living in a simulation, that none of the other alien civilizations just have been simulated yet.
Robin: I actually have a section on that at the end of the book there’s a section called Aliens, because so many people have asked me, “What does this book tell us about the Fermi paradox?” Or, as I’ve called it, the Great Filter. [00:59:00] I basically have to say it doesn’t really tell us much. When you look out in the sky and you’re looking for aliens and you wonder why you don’t see aliens, you should be looking for things that are really far more advanced than we are. We are just at the earliest possible stages of even being the sorts of things you could see. If we’re going to change a lot in our future history, then that’s the sort of thing you should look for out there.
Now, emulations are different from us, they certainly give you an idea how different the things you might look for are different from us, but plausibly after emulations come [00:59:30] something else, and then something else after that. The robust thing to say is that, “Gee, it sure looks like we could grow an expand and take over the solar system and change the sun. We could just do really big things.” Even emulations will be able to do far more than we can. Whatever they turn into next will be even more capable.
So, the robust thing is to say is our descendants will just have an enormous ability to use matter in various ways and to make use of enormous energy and matter. Whatever you’re looking for should be something like that, something that’s growing. [01:00:00] Now, people have said, “Notice how the emulations really come close to each other into a small number of dense cities and in virtual reality they don’t even pay attention to the physical world. Therefore, couldn’t all these aliens be stuck in a bunch of small densities where they pay attention to virtual reality and ignore everything else?”
I don’t think that’s very plausible, because even if that’s what most are doing, this emulation economy’s growing very fast. Again, plausibly, it’s doubling every month and a big chunk of that growth is just raw physical. On a cosmological [01:00:30] timescales, that’s nothing. Even the growth rate of our economy doubling every 15 years couldn’t plausibly last for another million years. Even at the rate that farmers were growing, by doubling every 1,000 years, even that couldn’t last a million years and a million years is a tiny fraction of cosmological time.
The robust thing to be noticing is just how fast total growth has been and how robust it seems to be of lasting through many different eras. So, the main thing to say about aliens is, unless they really destroyed [01:01:00] themselves, they should grow fast and grow big. Then the puzzle is, where are they? Because wouldn’t something growing big and fast be visible out there? Yeah, that’s the problem. Where are they? They don’t seem to be there.
Euvie: What do you think is the most probable solution to the Fermi paradox?
Robin: They never were there. Something very early on in our history was the key filter, the origin of life or the origin of multi-cellular life or some key step long ago was something that was just very [01:01:30] hard and it’s hardly ever happened in the vast universe. Therefore, that filter ahead of us isn’t so hard. But even if the filter ahead of us is Much easier than what’s behind us, still it could still be hard enough to kill us. So, we should still be really worried about what is ahead in our history that could kill us and trying to do everything we can to avoid that.
Mike: I’m actually surprised based off your book and the idea of the emulations that you wouldn’t say they didn’t grow bigger, they went smaller, they went into their own virtual reality [01:02:00] dimensions and never bothered to leave it.
Robin: Well, any one Em gets smaller, but the entire Em civilization gets bigger. Virtual reality is built out of physics and real material things. You can’t be in virtual reality unless you’ve got real computers running, made out of real physical stuff with real energy and real cooling and real real estate. And the actual physical real estate the emulation economy uses might start small but, again, it’s growing so fast that in a relatively short time scale it’ll fill up the earth [01:02:30] and then fill up the solar system.
You might have many trillions, quadrillions of emulations will join the virtual reality but built a solar system full of computers where we disassembled everything else to make those computers so they can enjoy their virtual realities.
Euvie: Like the cover of your book.
Mike: Do you have an audio book planned?
Robin: I’m told when it’s coming but usually I don’t get that much details. I’m also told that Chinese translation is coming.
Mike: Cool. Alright, thanks for joining us. This was a lot of fun.
Robin: It’s been great talking to you.
Mike: Alright everyone, thanks for listening. [01:03:00] That’s it for that episode. If you want to get access to the show notes go to futurethinkers.org/episode30.
Euvie: If you guys liked this episode we very much appreciate you leaving us a review on iTunes or wherever you’re listening to this. Thanks for listening. We’ll see you in the next episode.
Brain emulation, also known as mind uploading, is a concept that has fascinated science fiction writers and futurists for decades. This is the theoretical idea of scanning a brain – or a mental state – and copying it into the computer. Although no such technology currently exists, many prominent people have explored it as a likely possibility for the future, from inventor and futurist Ray Kurzweil to science fiction author Charles Stross.
Robin Hanson: The Age of Em – Whole Brain Emulation
Futurist and economist Robin Hanson takes the idea of brain emulations to a new level of detail in his new book The Age of Em: Work, Love and Life when Robots Rule the Earth.
He extrapolates from what we currently know about human biology, computer science, physics, psychology, and economics, and attempts to predict what will happen once we are able to successfully upload the first humans into the digital form. By his calculations, this is something that we can possibly see within the next hundred years. The events he describes in his book take place 1 year after the first human has been uploaded to the internet.
Robin Hanson himself has a very diverse academic background, which makes his multidisciplinary approach somewhat different from those who came before him in this field. While many other futurists and science fiction writers have focused on trying to get the technology parts of the story right, Hanson focuses more on the economics, psychology, and social science aspects on the emulated humans and their environments. He also outlines the 3 technologies that will be required to create full brain emulations, and how we can develop those technologies faster.
In This Episode of Future Thinkers Podcast:
- What happens 1 year after we upload our brains to the internet
- How brain emulations will interact with each other and biological humans
- Life in VR: what would it be like?
- Why future growth may not driven by innovation
- The 3 technologies that are required to create full brain emulations
- The ethics of emulating humans – destructive and selective uploading
- How brain uploading will change the concept of death
- The simulation hypothesis – what’s the chance we are already in a simulation?
- Brain emulation and the Fermi Paradox
Quote From This Episode:
“History is important and interesting to study, but the future is also important – plausibly, more important, because we can actually do something about the future. The past is too late to do anything about.” – Robin Hanson
Mentions and Resources:
- The Age of Em by Robin Hanson
- The Singularity Is Near by Ray Kurzweil
- Accelerando by Charles Stross
- Zero Marginal Cost Society by Jeremy Rifkin
- Operating Manual for Spaceship Earth by Buckminster Fuller
More From Future Thinkers:
- Individuality and The BORG (FTP006)
- Transhumanism and Technological Evolution (FTP007)
- Blockchain: Building Blocks for a New Society with Vince Meens (FTP033)
- Global Phase Shift with Daniel Schmachtenberger (FTP036)