Robin Hanson (@robinhanson)is associate professor of economics at George Mason University, and research associate at the Future of Humanity Institute of Oxford University with a doctorate in social science from CalTech, master’s degrees in physics and philosophy from the University of Chicago. He has spent nine years as a research programmer, at Lockheed and NASA, has 3500 citations, 60 publications, 700 media mentions, and he blogs at OvercomingBias.
Robin is also the author of The Elephant in the Brain – Hidden Motives in Everyday Life and co-authored The Age of Em – Work, Love and Life when Robots Own the World.
Hanson is credited with originating the concept of the Policy Analysis Market, a DARPA project to implement a market for betting on future developments in the Middle East. Hanson also created and supports a proposed system of government called futarchy, where policies would be determined by prediction markets.
Hanson is a man willing to challenge conventional wisdom/norms and has lately drawn criticism for his unconventional economics positions on sex, gender dynamics and problems with today’s society.
In our wide-ranging conversation, we cover many things, including:
- The reasons our culture values and norms are quickly changing
- Why physics forced Robins to become an atheist
- How Robin sees artificial intelligence progressing
- The power of prediction markets and why we haven’t seen more uptick
- Why brain emulation may be the most likely future scenario
- The reason Robin prefers to be more like a historian than a futurist
- Why we’ll never answer the hard problem of consciousness
- Why economics is a great way to forecast the future
- The problem with academia and education
- Why Robin isn’t worried about breakout AI
- Why Robin is sceptical of blockchains
- The reason Robin signed up for cryonics
- What folks should know about AI boom and bust cycles
Producing this podcast and transcribing the episode takes tons of time and resources. If you support FringeFM and the work we do, please consider making a tax-deductible donation. If you can’t afford to support us, we completely understand as well, but an iTunes review or share on Twitter can go a long way too!
So, brain emulation is the scenario where we report the software that’s in the human brain now. So today if you have an old computer running software that you like and you want that same kind of software running on a new computer one approaches to stare at the software or try to guess how it works and then write software on the new computer that works how you think it works on the old computer. But another approach is to write an emulator on the new computer that just makes the new computer look like the old computer to this offer if you can write an emulator you can just move this offer over and it works. You don’t have to understand it you don’t have to rewrite it. Big savings that offer as complicated and messy. So the idea is to do that for the human brain to make an emulator for the software in the human brain.
AI automation robotics it seems as if every startup and large business today is looking at AI and the implications of automating both jobs and tasks to lead to greater efficiency and output. Today we have Robin Hanson on the program. Robin is the associate professor of economics at George Mason University, research fellow at the Future of Humanity Institute at Oxford and did a doctorate in social science from Caltech a masters in physics from University of Chicago and spent nine years dallying artificial intelligence at Lockheed Martin and NASA. He’s been cited in over 3500 articles, 60 publications and has been in the media quite frequently. He’s the author of elephant in the brain and the age of [Inaudible] where he looks at the implications and potential of artificial intelligence and the emotions and driving forces of humanity. Today I had the chance to chat with Robin. It was incredibly interesting. We covered a wide range of topics including the reasons are cultural values and norms are quickly changing. Why physics forced Robin to become an atheist. How Robin’s use artificial intelligence progressing, the powers of prediction markets and why we haven’t seen more Optik y brain emulation may be the most likely future scenario. Why Robin prefers to be more a historian than a futurist when it comes to forecasting forward. Why will never answer the hard problem consciousness, why economics is a great way to forecast the future. The problem with academia and why Robin isn’t worried about breakout AI is now without further ado I give you Professor Robin Hansen.
Matt: I love to start these up with a story and you said you were in a cult as a kid I think that’s where we got to start.
Robin: Well I’m 12 years older so a young tween I was my parents were Christian and I was Christian and I met up with this other Christian church and this was a Pentecostal church in the San Diego area and they had a young idealistic people and lived in group homes and some of them went off and lived in a compound in Iowa. And so I went to their services regularly there in San Diego and they spoke in tongues and they were energetic and they were very loving and passionate and you know inclusive. And as a nerdy little 12 year old boy I ate that up and then my parents soon disapproved. And so they eventually said I couldn’t go and so I didn’t. And then that just faded into my background and eventually I become atheist in the sense that Tom in college I learned physics and it became my worldview and there just wasn’t much room in the physics world view for the mystical elements of religion. And I never got mad or not had a big argument. I know some people who are converted to atheism and all sort of blame the parents and say why did you teach me all this stuff. The thought that I sensed all my style.
Matt: It’s easy to understand how ignorance comes from other people and they pass that down. So you essentially go a full 360 economics and physics are about as far away from religion as you can get.
Robin: Now I’m not sure they are so far away. Are there grand. So in my talks I show this circle of academia. So if you map all the academic fields by citation and you put the ones next which are seeing each other a lot of turns out to be a ring. And it turns out that economics and physics are close to each other on one side of the ring. There are the abstract topics the opposite side of the ring is you know molecular biology geophysics you know all these very fine details and a lot of medicine but computer science physics economics they’re all on the other side. In mathematics over there with big abstract concepts powerful theories that explain a lot of things with a small number of assumptions and religion is kind of like that too at least in its theology as has these big grand conclusions.
Matt: Interesting. That is that’s an interesting way of looking at it. I want to I want to dive in now so. You’ve written a couple of books. The age of them was the one that initially initially brought me to your eye and you brought you to my attention and your look at essentially artificial intelligence and what that future could look like. So let’s jump into that because I know you make a lot of predictions and you were looking at a lot of data.
Robin: Right that book came out two years ago but just last week the paperback edition has come out so it’s timely in that sense. And I actually saw it in the airport bookstore three days ago.
Matt: So I imagine
Robin: it’s San Francisco airport of course which is a little odd. I’m sure it’s not going to be in most airport bookstores but it was nice to see the airport bookstore there. So the book is called the age of em and the topic is one route artificial intelligence and what would be the consequence. So as you know one of the biggest most plausible theories about what big thing could happen in the future is that we eventually achieve machines that are as smart and capable as humans. We aren’t remotely near there but that’s or could happen. There are actually several routes by which it might happen. And my book focuses on a route that people aren’t talking as much about today but it’s still one of the main plausible routes and that’s brain emulation.
Matt: What is brain emulation and why did you go that way.
Robin: So brainy emulation is the scenario where we report the software that’s in the human brain. Now so today if you have an old computer running software that you like and you want that same kind of software running on a new computer one approaches to stare at the software or try to guess how it works and then write software on the new computer that works how you think it works on the old computer. But another approach is to write an emulator on the new computer that just makes the new computer look like the old computer to the software if you can write an emulator you can just move this offer over and it works. You don’t have to understand it you don’t have to rewrite it. Big savings that suffer as complicated and messy. So the idea is to do that for the human brain to make an emulator for the software and the human brain. So to do that we need lots of cheap fast parallel computers we need to take particular human brains and scan them and find spatial and chemical detail and then we need computer models of how each kind of cell works in terms of taking signals and changing internal state and signals out. If you have good enough models of each kind of cell and a good enough scan of a particular brain then you can make a computer model of that whole brain and put it together and just turn it on. And if you do it right it should just work without your having to understand that structure and software you don’t need to know why it works you just need to know that it works. And if it’s right that this emulation will act the same way in the same situation as the original human right it will be different as it has different experiences. But it’s very predictable and understandable. So that’s what a brain emulation is. We’re not close to that but plausibly we might be within a century or so and go ahead
Matt: Realistically it will be very challenging because every brain is wired differently built differently. Now you can run a great machine learning algorithm over that.
Robin: Oh sure but it’s like copying a movie. So if you get a movie on a disk and you want to make a private copy you don’t need a different program for every movie you need just one generic program to copy disk. The idea is to have just a generic process that would scan a brain and tell you which cell is where connected to what for that particular brain and that same process would work for different brain. It’s just collecting a bunch of bits describing where the cells are but not understanding them just like program the copy the movie wouldn’t understand the movie or be able to make another different movie it would just copy all the bits in the movie.
Matt: Why do you think why do you think that this is slightly controversial. It seems like at least when I think about it I feel an internal backlash like is it actually that easy to do. Not easy in terms of the technology we have but easy in terms of the ability to map the brains to think that’s just human human ego.
Robin: Well there are two different ways the problem can be hard a problem can be hard because it just has difficult you know conceptual things that are just hard to understand are hard to do and a problem can be hard just because it’s large like you and I you know if I were working in my backyard if I had one we could make a dam by just piling a bunch of dirt in a spot so that the water didn’t leave that spot right now would be a dam. You and I would find it very hard to make Hoover Dam not the concept really isn’t that different it’s just done up at the larger scale. But once I show you how to make a small them and even a mildly larger dam you’ll get that Hoover Dam is possible. You don’t you know it’s not clear when and not where when you can afford it but eventually it will be possible. So brain simulations are hard but they’re hard in this scalings sense. So from a conceptual point of view from the can we understand how to do it at all point of view. What we have to understand for brain simulation is how to model brain cells that is for each particular cell. We need a computer model of how to take signals and changes internal state and send signals out. That’s not trivial. Now we have decent models for a lot of kinds of brain cells. We just need models for all the kinds of brain cells that that’s the challenge. And it does seem that individual brain cells are just simpler than brains. I mean that’s not obvious. It could have been the other way around. But if an individual brain cell is simpler than the brain then it will be easier to have a model of a brain cell than to have a model of a brain. So the familiar route to artificial intelligence is trying to model a brain trying to understand the whole brain in terms of what are its main parts and how they interact and what functions they provide. And that’s as hard as understanding a brain is for emulation in terms of the conceptual difficulty. The thing is can you understand a brain cell which they’re complicated. Now of course most of the complexity of a brain cell isn’t relevant here. So all we really need to do is model the complexity that’s relevant for the signal processing for how they change signals that come in and go out.
Matt: Hey Matt here a quick aside I should pushed back on this in the podcast. I don’t think you can model brain cell cells simply in terms of ones and zeros and get an effective outlook or intelligence and possibly consciousness. This is something where I think Robin’s a little bit too optimistic in terms of how easy this will be and I think it’ll be significantly more challenging especially considering all of the separate factors that go into how we think and operate but didn’t have time for that in the podcast maybe next time let’s come back to it
Robin: And brain cells of course are cells and so they have all the complexity that cells do in terms of being able to reproduce in sensing and repair cetera. And so it will definitely have to you know we can avoid most of that complexity but we will have to figure out what are the relevant parts and model them. So the main argument for brain emulations being feasible eventually is that their brain cells are simpler than brains and then it’s a matter of scaling it up by making doing a lot of them and of course we’re not there yet but as you know we have a lot of technologies that once you can do a few of them as the cost falls and the technology gets better you can do more and more you can scale it up.
Matt: Do you think consciousness is an emergent property of connected connected intelligent cells
Robin: while conscious this is a property of you and I. You and I are physical objects. We are exactly physical objects. There isn’t an extra part of us beyond the physical parts of us. We are just physical objects we are made out of atoms arranged in certain ways. If confirmed through an enormous detail of a repetition every part of you we take it out and we take it apart and we see that it’s made out of ordinary matter interacting in the ordinary ways. So that’s what you and I are we are physical objects though obviously physical objects are capable of consciousness
Matt: Or non-physical objects. That’s where we get into an interesting scenario.
Robin: Well we are physical objects. You and I and clearly you know we are conscious therefore our physical objects can be conscious. So if you think there’s some special feature of physical objects that makes them conscious that isn’t just the physical parts of them there’s some extra part and that this extra part doesn’t interact with the physical parts then there’s no data we’re ever going to get about the physical world that will tell you about this extra product. And so if you have these different hypotheses about how these different parts have to be arranged in a certain way in order for something to be confident conscious we’ll just never going to know about that
Matt: I need to jump in here robin may well be right. But keep in mind. Never say never. When we put limits on ourselves that means we’re not able to think big or think outside the box. And it’s especially problematic because we’re assumes that the future scenario always stays similar to the present. We don’t know what we don’t know what we don’t know and that’s the big problem. Now jumping back
Robin: So it just started out right now we know as much as we ever will about any non-physical parts that are required for consciousness we will argue about this for trillions of years. If you believe about that which so this idea brain emulation has been around for a while and whenever the subject comes up people usually get focused on these philosophical topics you know wouldn’t emulation be conscious. What variations would be conscious. Also if you made one to me would it be me. I get focused on the technology. You know what mechanisms would be possible. This conversation is going over and over again for decades and I’ve been tired of that and I thought there’s been a key part that was neglected which is OK about what would actually happen. And there’s very little attention to that. There are some fictional stars where people play out dramatic stories but they aren’t at all trying to be realistic. So in my book I don’t really give much attention to the philosophical issues. I go straight to saying what would happen. On-Site look if you want to hear about the philosophical debates there’s plenty of other places to go [Inaudible]
Matt: what would happen. Just a brief overview of some of your thoughts.
Robin: OK so it’s a world that is different from our world as our world is from our farmer forger ancestors. That is pretty different. These emulations initially will be black boxes that will be opaque and so we won’t be able to modify them very much we won’t be able to go in and tweak somebody’s motivation or make them you know music abilities you know to be a different music genre or something like that. We just take these black boxes and we can turn them on turn them off erase them copy them or run them fast and run them slow and that’s about it. So later on we might be able to do more and maybe that will change things a lot. But initially all we could do is just run these black boxes. Initially these black boxes would be produced from scans of individual humans and they would be destructive scans. They would destroy the original human in the process of creating the scan and so would be a one way move from becoming an ordinary human to becoming an emulation.
Matt: If that sounds morbid that’s because it is. Don’t worry we’re going to jump into cryonics and extending [Inaudible] a little bit later in the episode. Robin’s actually a cryonic patient. So let’s keep going.
Robin: And these emulations would create a new world. They would have physical cities but it would be different from the human cities and in the cities they would grow very fast. So our economy doubles roughly every 15 years. And these emulations their economy would double roughly every month. So very quickly their economy would come to dominate the economy of the earth.
Matt: Well just a quick a quick trivia question are we talking about a ready player one scenario where they’re living in a virtual world or these are robots are in some way physical presence.
Robin: So the emulations you know end up doing all the jobs because they’re cheaper and better and they do jobs that we do mostly in desk jobs and they do those in virtuality. But when they have a physical job that’s very connected to physical reality and they do that in a physical body. So when they’re driving a truck or managing an assembly line or you know digging a mine they have some appropriate body for that task that body can be swapped out any time so it’s like getting in a truck to drive when you’re a trucker and getting out when you have a break they don’t have to stay in that body all the time. But but they have whatever body they need for the job in you know when they’re working at a desk job they might as well be at a virtual desk and when they’re in leisure they’re inevitable world that they know what it’s not like the Matrix or something where they’re fooled into thinking they’re in some other world. It’s more like our cities and our buildings. If you look around you in a building almost all the surfaces you see are artificial they’re constructed so that they will seem a certain way behind the surface as you see there are pipes and struts and wires that you don’t see that are there to make things work. But you know there’s all those things behind the walls. You just don’t want to look at them and so we don’t show them to you. So the same for the emulations they live in a virtual reality it looks the way they want it to look but they know it’s a virtual reality. They know how it’s made and that’s important to them sometimes.
Matt: And they would be driving the vast majority of the economy because they can essentially speed themselves up
Robin: while they’re cheaper. The key point but in addition to being cheaper they can change their speed. So the two most dramatic differences in their lives from us are the fact that they make copies of themselves and the fact they can change their speed of course. Another difference is they’re immortal in principle at least what you might think is dramatic. And another difference is that they always virtual reality can be gorgeous and luxurious with no pain, grimed disease, hunger. Their bodies can always be beautiful. So those are mild differences from our world. But the biggest differences are the fact they can make copies and they can change their speed.
Matt: which would make it very appealing for people that were already economically disenfranchised because now suddenly you can move into another world you can eat whatever you want you could I mean in sense you can BJC or you could be a bomb.
Robin: Well so you can have a luxurious physical experience. It’s not clear how satisfying that is to humans because we are pretty status-conscious people as the presumed they’ll figure out whatever is high status in this world that’s limited and we’re lost after that. And so they may not be that satisfied just like today. We’re live in a pretty rich society and we can give most people sort of the physical comforts that people wanted for the last 10000 years. And mostly they’re not happy with those because they want the other thing everybody else has
Matt: and everyone else is not happy with them. Those are some very interesting dichotomy we’ll jump into that and ask why did you choose brain emulation when your focus was on what the potential for artificial intelligence was that the most likely scenario in your opinion or just one had been less explored.
Robin: Well I certainly thought it was neglected and so I do have a strong heuristic to look for neglected topics in my work. And it’s also easier to make predictions about. So when we just think about abstract future robots it can be pretty hard to get some sort of purchase. Some concepts that let you make predictions about that. That’s actually what I’m doing now on a separate project but with brain relations they are very human-like and we know a lot about what humans are like and so you can say a lot about this world. And I really wanted to show how much you could say about a particular scenario by just turning the crank and making predictions using our standard tools.
Matt: It’s essentially first principles but just starting something and going from there it’s it’s very interesting so you studied economics and physics
Robin: and computer science
Matt: and computer science. It’s an interesting combination. What [Inaudible]
Robin: Well like I said earlier I like abstraction. Now actually I think most people who are inclined to become intellectuals who try to get a career as an intellectual. The one of the most common failure modes is that they can’t focus enough people enjoy spreading their attention across a wide range of topics and areas disciplines etc. That’s just fun for people and when they’re treating their intellectual life as-as fun doing what they enjoy they spread themselves out a lot. And then the worlds that reward you for being intellectual they reward focus a lot more on especially academic. I was like most wannabe intellectuals to broad but I managed to squeak by by being narrow enough to to achieve no concrete accomplishments in my particular area. But once that constraint was weaker I was more tempted to spread out again so I think compared to most academics I’m pretty wide. But compared to most wannabe intellectuals I’m about usual
Matt: do you think that we need to overhaul the system the education system.
Robin: I have many critiques of our research and education systems but whether they can be solved by overall is a hard question. It’s much easier to diagnose what’s going wrong than it is to produce effective fixes.
Matt: Yes it is especially want to have an effective fix you have to kill the existing operation.
Robin: Typically when part of the problem is what the customers want and they don’t want what you want them to want. So for example the advertising industry you could say gee the advertising industry is not doing a very good job of informing customers about the pricing features of products. We should reform advertising to make advertising more informative. And of course the problem is that the ultimate customers of advertising don’t actually want that so much. The current kinds of advertising are targeted for them to catch their attention and generate interest in them. And what you really want is for the customers to be different or somehow to force the customers to consume a different kind of advertising than they would choose for themselves. And that same problem in education and in research you can identify the problem. But it ultimately comes down to the customers and what they want.
Matt: Would you in towards totalitarianism.
Robin: I don’t know. I’m quite concerned about totalitarianism but I I get that. I see the tradeoff. So so I’m actually interested in the topic more generally of the tradeoff between what I call governance and competition. And I think that tradeoff is especially interesting regarding the future. And I think a lot of discussions of the future end up being discussions of that tradeoff regarding the future
Matt: what are some predictions that you have in terms of governance and where we might be headed. A quick time out before Robin answer. This is why I love economists. They’re able to look at and analyze situations in a completely different sense than the vast majority Robin’s about to break down a concept that’s very interesting not something that’s well thought about and explains the shift in behavior and human trends as we move forward from a society of scarcity to a society of abundance. Hope you enjoy
Robin: just straightforward predictions and governance is that we’ve slowly been able to do governance on larger scales larger geographic scales larger social scales and to do more topics of governance and so we’ve been doing more governance that that’s the straight forward trend over the last century or two. And in a world that continues like ours that would continue to be the trend. I also think that our attitudes are moving in that direction. So I have the story of how most of the major social and attitude value trends over the last century or two are caused by our moving from forger to farmer about Apple or farmer to forge or values. And I think this is part of that but that’s not the main thing. Mainly we are just getting good at larger organizations so this is also true for firms firms are slowly getting larger cities are getting larger nations are getting larger government is doing more things. And so that’s a slow long-term trend but it is a slow trend and it means at the largest scales we still fail to apply governance to a lot of things because we’re still not very good at it and many people would like to jump faster than we are really able to do and I think that’s you know often typically a mistake to try to do governance at a larger scale on a larger topic than we’re really up to the task.
Matt:[Inaudible] gets [Inaudible]
Robin: It is a lot of costs on large-scale coordination
Matt: of paperwork especially show. you you have a major background in prediction markets. I want to get into that on the economic side of things and I’ve worked with dark Bahnson prediction markets that apparently have helped the U.S. government in the Middle East.
Robin: Well we were a research project you know assigned to the task to show that but. So in 2003 now 15 years ago over 15 years ago I was part of a project that was trying to show the Department of Defense that you could use prediction markets on topics of interest to it. And we chose geopolitical stability and related events in the Middle East as our topic and we were about ready to go live with a Web site to invite beta test users to be included. And then there was a big press conferences the two senators held declaring that department offense was about how betting markets on terrorist attacks and this was a terrible thing. And that was my project. And then the very next morning the secretary of defense declared the project killed. In between that time nobody asked us that the accusations were correct which they weren’t. But they happened to have this press conference just when the [Inaudible] PR person was unavailable. What a coincidence. And so there was a lot of press over that over the next few months about this topic. But and that kickstarted more tension into prediction markets at the time.
Matt: What is a prediction market for people that don’t realize. What are the implications?
Robin: So prediction markets. Is it just another name for a speculative market or a betting market like the stock market or currency market are betting on football. The key idea is that when people bet on a question like will this team win this game more will this how horse will be the winner of the horse race or where will the stock price go or what will be the price of gold. The market price ends up aggregating a lot of information that the current market price is a pretty good estimate of the future, in fact, it’s hard to do better. And that’s a powerful force that until recently hasn’t really been harnessed for many other purposes.
So if you want to know something say you’ve got a project at your company and you wonder you have a deadline and you wonder will we make the deadline. The usual approach is to have meetings where people in related the projects say how are you doing with your part of the project. And people say we’re we’re we’re probably going to make the deadline we’re looking okay here’s our issues and then they produce some aggregate forecast out of that and it’s usually not very good particularly it tends to be biased toward telling you-you’re going to make the deadline even if you want if you just make a better marker where people can bet on the deadline and they can bet anonymously so that even if they are bad news they more retaliation won’t hurt them. That ends up being a lot more accurate. It tells you quite reliably whether or not you’re going to make that deadline. And that’s a powerful force that you could use for that purpose if you wanted to know when we made that deadline and you could also use it to say Well we what will sales be of this product you can use it to answer many conditional questions. You could say what will sales be of this product if we introduce this product. If we’d put it at this price what would the chance of making the deadline be if we change who’s running the project if we added more personnel if we change requirements. So prediction markets have this huge potential to help organizations and people you know find out things about the future if they want.
Matt: What are your thoughts on the crypto projects that are trying to create decentralized prediction markets.
Robin: Well most of the crypto projects out there are software people who are really focused on software issues you know mostly they are software people who put up a coin offering and got some money and then went to work solving the software problems that they thought were interesting and important. And there are of course some software issues but we’ve had prediction market software before. We weren’t lacking for prediction markets software. It always had better software. What we were lacking before was organizations willing to use that software on their problems. There’s a lot of issues in prediction markets being disruptive in organizations. And so what we really need is to explore different ways to use predicts markets and organizations in order to try to overcome this problem a critical disruption. That’s what the field of prediction markets needs. And for that they need organizations willing to try it’s actually been hard to find out
Matt: What if you use that outside of organizations because then it will still be used by organizations but they will have no control.
Robin: Well I mean the key thing is somebody has to pay for it. So it’s a technology for aggregating information for collecting information together on a topic. It comes at a cost. And so somebody has to decide this is a topic worth having information on. So if you have a deadline somebody has to decide I want to know if we’ll make that deadline.
And then if they create this market and they subsidize it then other people be tempted to come and participate to tell you whether you’ll make the deadline. So it’s a way of paying other people to tell you something that somebody has to pay. If you’re just making market and throw it open to the world than the people who contribute will be people who do it for their own reasons. And of course the only do those on there are talks so we do have some markets out there in the world on currency and stock and sporting events and they happen for a spattering of reasons. But though it won’t work for just a generic random stock if you have a generic topic that you want an answer to and you just open a betting market on it somehow out there there’s very little chance that anybody will care enough to contribute. To answer your question people today care about sports because they like to argue about sports with each other and that’s something that really into. And so they liked about sports as a way to affirm their knowledge about sport and that they are committed to their sporting
Matt: [Inaudible] at the same time It’s not like betting on politics where you have to hate the other person.
Robin: And people are willing to some extent the bet on the presidential election because they argue and talk about that and then there’s a lot of economic incentives for speculation about currency and stocks and commodities. That is there’s a lot of economic organizations that buy and sell those things and as long as that’s happening then if there’s a mispricing of those things there’s a lot of money to be made by somebody else coming in and fixing those mispricing and that tempts people to come in and study those markets and trading. But again if you just pick a random topic of interest to you and set up a market on that there’s no particular reason to expect anybody else to carry enough to come and traded. To answer your question for free a people do answer sporting questions for free because they are already there for other reasons. Thinking about those questions and wanting to show people that they know better but that’s not true of your end of question.
Matt: So you’d have to implement the core type system where you have reputational scoring or answering questions for basically people that want to prove their expertise.
Robin: There are many ways to pay the participants but still they need to be paid. So Quora is a way of you know enticing people to answer questions. Quora relies on people wanting to have a reputation for you know answering people’s questions but of course, they only answer the questions they want to answer. So you can’t just ask a random question of interest to you on Quora and get an answer. You’ll have to ask the sort of question that other people will be inclined to answer
Matt: Fair point. I want to I want to transition a little bit now what industries are you most excited about.
Robin: I see myself as-as an economist looking at the whole economy. So I think they compared to most futurists I want to be not really focusing on small number of particular industries and products. I think there’s something unhealthy or at least inaccurate in sort of the tent the tendency of futurists to sort of talk about the world in the future in ways very different than historians talk about the past. So you know any one time period you could you know talk about it before it as a futurists or talk about it afterward as a historian. And it’s the same period with the same major events in the same major processes going on but those two groups of people talk about events and very different styles they have very different expectations were events and I want to be more like the historians because I think they’re more accurate. I want to think about the future as if I were thinking about it later as history and for that person that purpose I want to understand all the major important forces that are likely to be going on and not just focus on a few sexy demo things that a lot of people are pushing their products on.
Matt: What would you say are the the three most important factors that you’re looking at currently them in terms of the economy and where we’re headed. Ever wonder why futurist forecasts are so far off, Robins about to explain.
Robin: Well there’s just long-term trends and understanding them and so that has to be the first priority. I think futurists people call themselves futurists get way too focused on short-term fluctuations and they should really focus first and foremost on the longer term trends and understand them. So you know not just why are we getting rich but why are attitudes changing as I said I think our attitudes are changing because because we’re getting rich and that’s making us move toward forging values relative to farming values.
Matt: Can you explain that a little bit more.
Robin: Sure. So forger’s you know hunted and gathered for a million years and they were mostly like animals in the sense that they did what felt right and that was mostly the right thing to do. But humans had enough cultural plasticity to be able to make new cultures and new values. And that was useful when farming became impossible. Farming was really only possible with a lot of changed attitudes and behaviors. So foragers were really egalitarian. They shared their food they wouldn’t allow anybody to brag or be air put themselves up as a boss. And they they made decisions collectively and they were relatively promiscuous they didn’t have long-term marriage although they had a short-term pair bonds they raised kids more communally. They didn’t save up things for the future they’d have very few physical property. They didn’t land they were friendly relations with their neighbors. But farming required a whole different world of attitudes. Farmers lived in much larger groups that were more hostile to their neighbors they had war they had inequality within they had property had marriage they had slavery. They know or they didn’t travel as much they had less art they had less free time. Their nutrition wasn’t as good. They needed more self-control to make plans for the future and to trade and fight or. And and all these things. And so farmers relied more on conformity and religion. And a lot of sort of attitudes celebrating self-control and self-restraint and commitment. And that worked for 10000 years. And then in the last 200 years we’ve been individually getting rich and as we’ve been getting rich a lot of the social pressures that turned forger’s into farmers have just no longer felt as compelling to us. So for example you know forgers are often promiscuous but farmers marry and if you have a young farming woman who is tempted to have a child out of wedlock. Because it’s a very natural temptation. The cultural tells that woman well if you do that you and your child may starve. And that’s a very real threat. It’s not pretend it’s not fake it’s credible and that keeps women like that largely in line in terms of what the culture wants them to do. Now as you get rich you are or have the same temptations but now you see around you other people who follow those temptations and did OK. There are a lot of young single mothers who are living ok life. And so many people say you know the threat that the farming world had of all the terrible things what happened to you aren’t as credible and that’s true all through our lives and so a lot of major social trends over the last few centuries I think can be attributed to drifting back toward forging values as we get rich so
Matt: Is that you [Inaudible]
Robin: Well that was part of age and then the first few chapters I outlined that I thought of making that into a bigger book but I didn’t really have enough as much to say. And so that’s why you know it’s only just the beginning of age and that includes that but in the last few centuries we’ve been drifting toward democracy toward leisure toward art toward promiscuity, Low fertility, low religion, you know less war. Just most of the major trends can be understood as moving back toward forging values and that people like that and they like to look forward to a future where that continues a Star Trek future Culture novel future where we get even richer and even more able to indulge ourselves more peaceful and no more art and everything else. And one of the key features about the age of animals that doesn’t happen for the robots in the age. So the biological humans who are off on the margins of that world are retired and living on basically the rich capitalists spending the vast wealth that they have as part of the world. Their attitudes continue in that direction. But the robots do not. They are now poor again. They are living at subsistence level and they have need of social conformity and other pressures to get them to act in the ways that this world wants them act.
Matt: So what happens in a in a scenario like that where you have essentially enslavement is not it.
Robin: This not it. No. I mean slavery is certainly possible in most any world but I doubt it’s very common in the future. I just don’t think it’s very productive. So in the past in history we’ve had times when there was plentiful land and scarce labor. And then there’s been times when there has been lots of labor and not so much land. When there is lots of land but not much labor then wages are high and that’s a time when you might want to own a slave because there’s value in owning life. If you don’t honestly you’ll have to pay high wages and if you own the same you just have to pay as much as it takes to feed them and keep them locked out in the other periods where there’s plenty of people and not so much land wages fall to near subsistence level. And in those times there’s actually not so much point in owning slaves. It’ll cost you about as much to feed a slave as it will to hire a independent free worker and that’s how. Of course most times have been in history times and places people have been you know the wages have been near subsistence level. But in the last few hundred years we’ve been getting rich and so you know there’s been a lot more value in owning a slave in the rich world. The age of them as a world where they go back to near subsistence and so there’s actually that much point in owning slaves and in addition we know about the productivity of different kinds of treating people as slaves in the economy saying in the US South slavery is more effective with relatively simple jobs like picking cotton or cutting down wood or things like that house slaves and city slaves who have more complicated jobs the more discretion and couldn’t be monitored as well they were treated a lot better and they were often treated nearly as well as a free worker because of treating them really harshly just wasn’t very effective in those kinds of jobs and most jobs in the world are those kinds of jobs.
Matt: Let’s play devil’s advocate. So I actually am writing a blog post now a century of slavery and it’s looking at how we may enter a new era of slavery. And what I’ve noticed is every time humanity is discovered another race or species of human are related then we’ve enslaved them because of course, they’re inferior to us and that we can come up with whatever moral reason we need to to have someone else help us or very well.
Robin: I’m I’m not sure which examples you have in mind certainly you know humans have enslaved animals. The ones that we can now you know this is actually a very small fraction of animals that we can domesticate animals to animals just well put up with it so we either kill them or leave them alone. But the animals that could be domesticated we did you know animals have had a very simple jobs in our economy and so end slavery can work for those very simple jobs that don’t require very much judgment or discretion. And they can be monitored. Well now when humans have encountered each other they’ve of course are sometimes enslaved each other. And you know in the ancient world it was a common thing in war to enslave your losers. The killer enslaved them and that was one of the ways you could sort of take value out of the losers. You could grab their land you could grab their women grab their physical stuff and you could enslave the others. So I’m. But slavery didn’t usually last in the sense that the slaves didn’t have this population of slaves who kept growing their population to be maintained. Usually the population stays with the client until you conquered another area and grabbed another bunch of slaves.
Matt: In that example but I meant more specifically so we had we had a triangular trade. We also had religious slavery early Christians and Muslims.
We’ve had a lot of different us and them mentality that is essentially I don’t know that there’s there’s a proper word for that. But essentially whenever you can create a barrier between yourself and someone else in your mind yet it becomes much easier from a propaganda standpoint to do whatever you want or need.
Robin: So I saw actually in the revised version of age about my I have more discussion of this than I didn’t have in the first version. There’s a chess section on slavery and there’s also a section on sort of dehumanization or an anthropomorphizing. And so we humans have these two capacities we on the one hand look at most everything in a is it like me or not sort of way. And so we are tempted to anthropomorphize many things that are not very human and we’ve done that all through history because we don’t have a very rich sort of vocabulary for what kind of things that we can think about. On the other hand, we also have the capacity to dehumanize to take something that fits all these criteria and tell ourselves that doesn’t if when that’s convenient for us. And so you know you’re talking about the letter of taking something that by any straightforward obvious measure which would be very human-like ants and calling it non-human because you don’t want to see them as human
Matt: Or in factory farming
Robin: sure about how factory farming they are animals are human-like in some ways and not in other ways. But since we don’t want to see the factory animals as like us then we choose not to.
Matt: What I what I’m primarily worried about is as we’re entering a world where [Inaudible] technology becomes more prevalent among the super-rich when you can begin to have evolution happening on a generational or shorter timeframe then it’s going to start out incredibly expensive.
It starts out incredibly expensive only the super-rich are using it then you’re gonna have exponentially larger differences between rich and poor potentially leading to extended lives and significantly better circumstances or one than the other class ultimately probably creating different pieces of human well.
Robin: So I mean I think the timescales over which those things could make big changes are long. So I actually personally think the other more robotic AI [Inaudible] are more likely to have the cause dramatic changes over a longer time scale but of course, they also have differences and differ kinds of creatures. I don’t think that’s sort of the mere inequality across classes is the driving force you know as what we’ve noted that we can take humans who are intrinsically quite similar and when in the past when we wanted to treat them quite differently we have that hasn’t been a barrier. On the other hand, if we want to treat things that are quite different from us as like us we we can and do that we anthropomorphize as we said. So I think it’s more about what we choose to do
Matt: Speaking of what we choose to do, there’s a lot of problems in the world what would you say are the largest problems in which you would like to resolve any specific order.
Robin: Well so I’m a social scientist and I’ve spent a lot of my life thinking about what the problems in the world are and thinking about solutions and of course what we economists in particular do as we think about institutional changes. That is you could try to imagine telling people to have different attitudes and maybe you could even succeed in some short run but it’s hard to make a stick. And so economists tend to focus on how could we arrange institutions how the rules of interaction. And you know who participates how acts that are such that we have a better outcome. And obviously there are ways in which culture supports that but it’s just really hard to figure out how to make culture different and how to how to control it and to make it be the way you want but institutions we can more concretely imagine how they could be different. We could have different voting rules or different property tax rules etc. and so much of what economists do is try to think about what are the biggest areas where the current rules are producing unfortunate outcomes and that we could make them better. And I think we know a lot of those things. We know a lot of different ways in which institutions can be improved. As someone who’s spent a whole career doing that and watching other people do that I think I can just say with confidence we know a lot. The harder thing is to get anybody to care. So what more fundamentally we need is a process better process by which when you come up with a better institution you can get people to adopt it. That’s the thing we lack in our existing institutions for proposing better and Lucian’s you know start-up businesses or you know being a politician and lobbying for political change. Those are pretty broken and so you know it of course is tempted to figure out better systems for those things so that you can get more things adopted. But even then when you come up and you can come up with better proposals even then the hard part is you can’t get anybody care. So most fundamentally the problem we have is you know first there’s the problem of having bad institutions and then there’s the problem of coming up with better ideas for institutions and then here’s the problem with getting anybody to want to care about those better ideas for institutions in order to promote them and try to make them happen. And most proximately a lot of these better ideas what they mainly need is small-scale experiments a lot of our best ideas for institutional reform look good on paper or even in mathematical theorems. But what they need is smaller organizations small group to try them out for a while. Small towns small cities and there’s just very little interest in doing that. So what the world most needs is small groups small towns small organizations even small churches to be willing to try to help us come up with better institutions by trying out some of the many ideas that people have come up with.
Matt: It’s like there is there are a couple movements now in especial libertarian and watching communities around trying to start decentralized or alternative governments or countries they’re small.
Robin: Yeah although I mean again they’re mostly doing’s offer so you know my main critique about the crypto world is it just way too focused on software. They loved to write software they want to hack software they don’t make [Inaudible] for tools. And of course that’s useful but you also need to interact with actual customers and actual people doing real things and help them and people, of course, say in principle they want to do that. When it comes down to the nitty-gritty of messing with that they’d rather suffer.
Matt: It is the major problem with scientists and developers they would much rather work on their work than actually try to convince others.
Robin: Well I just [Inaudible] But but but just get involved in the iterative process so just so like with prediction markets as I was saying before you know most innovation includes some abstract idea some general idea that can be described simply and has a lot of abstract potentials but then it also needs a lot of detail a lot of concrete detail that’s in a particular context to be worked out to make it work. This is true of-of course the most physical technologies and others offer technologies we use in the world. And it’s true even social innovations like prediction markets is not enough just to have a good abstract idea. You need to feel that abstract the idea in actual concrete circumstances and see the actual problems that show up in that concrete circumstance and then iterate by trying variations until you find things that work better because that’s how most technology anywhere has ever been done. And of course the person who came up the abstract idea tends to get more celebration and attention than the people who worked at all those concrete details but unless somebody works out the concrete details you know it doesn’t happen.
Matt: Yeah. You have to go in startup and there’s no such thing as an overnight success
Robin: Right because you have to make a product and then put it in front of real customers and find out what they do or don’t like about it and change iterate etc. And if you’re lucky you will eventually iterate to something they like. Before you run out of money. Yeah, that’s rare.
Matt: So Robin transitioning completely yeah you’re signed up as a cryonics patient. We actually had Dennis Kowalski of Cryonics Institute on. I’m curious. I imagine your book inspired this or vice versa.
Robin: Well I hope to be revived as I am. That is cryonics is this process by which you take somebody today where medical technology gives up on them. So they are officially legally dead but you don’t give up on them you freeze them in liquid nitrogen and you hope that later on new medical technology will be able to undo whatever was wrong with them and whatever went wrong in the freezing process. Now that the crew of that freezing process the heart of that problem is and the more that had gone wrong the harder problems and the harder problems also harder if you intend to bring back their physical body as a physical body to the full. You know it was before the problem or even when it’s young that’s just a really hard problem. So that will take a long time to succeed. I think eventually it will be possible but it’s a long way off and we can freeze them today and hope for that and then your main risk is whether the organization will last long enough to preserve you until somebody can do something. But the rehabilitation process should be a lot easier than repairing your entire physical body. The brain emulation process is just a scan the brain and just to see the key information about each cell to tell what kind of cell it is and what its key state is. So I think brain simulations will just be a lot easier to do than the full Chronic revival. And so it will happen sooner and if you are worried about the process of preserving you lasting you. You want to grab the first way out. You can find. And so that’s me I want the first way. And so I think brain emulation will it will be feasible long before a full biological revival would be feasible. And so that’s what I want. And the age of them is about a world of emulations which would include a few previous chronic patients
Matt: to to ironic things so you would essentially be reincarnated as a second-class citizen.
But it’s also kind of interesting because you won’t be incarnated as you so you’re kind of saving someone else’s life. And in a sense we don’t know at the end to that part it’s a little bit more touchy but with the brain etc.. So you write NASA and Lockheed what were you focused on. We haven’t talked about your physics background.
Robin: Well at NASA Lockheed I wasn’t doing because as I was doing artificial intelligence research so I got a master’s in physics at Chicago in 84 and I also got a master’s in Conceptual Foundations of Science which is related to the philosophy of science. And then I read about cool things happening in artificial intelligence often Silicon Valley also read about things happening in hypertext publishing which is the precursor to the Web. And so I went off to Silicon Valley and seek my future. I found a job at Lockheed which I tried to weasel my way into doing a stuff even though I had no background in that.
Matt: Artificial intelligence is a bit of a buzzword or a rebranding. Robin, I’ll explain a bit more what the industry used to look like. What’s happened since and why he thinks this isn’t the last boom and bust cycle
Robin: and then on the side I hang out with the Zanta group and other people interested in hypertext publishing in the future and so you know within that group was Eric Drexler in his book on nanotechnology that had just come out and so I spent a lot of time then talking with futurists about of course cryonics at that time as when I heard about that.
And had this job doing a research and as you as you may know that was a big AI boom I all the newspapers and media was talking about how I was big and was going to change everything really fast really soon. And of course that’s a lot like today as we have another big boom. And of course that was wrong and it was over hyped and that’s also true today it’s also over hyped today. So you know this boom will also pass to another decline and you know we’ve had these waves of increase and decrease in interest and they either go back for many many decades. And so but I was caught up in that boom I was a young 20 year old 20 something and I learned to do a research at the time and learned a lot of concepts that as a physicist I didn’t realize there was that much to learn so as you may know physicists are told that you know physics is pretty much all you need to do everything and there’s nothing else that much interesting to learn in the world. And so you know they don’t actually think there’s that much interesting to learn about computers. You want to do something with computer you just tell a physicist to go read the manual and write to do the computer to do whatever you need done and so on. I was surprised to learn how much interesting there was to know about computers and how to make them work. And the key concepts behind them.
Matt: It’s interesting how short-sighted a lot of academics are and thinking that their world is the only world
Robin: while that’s all people humans really were all short-sighted pointed underestimate all the other worlds around us and I actually that’s right you know one of the selling points I think for Age of them is people often celebrate they travel as a way to understand how things are different or you know exploring other disciplines or topics reading history as a way to expand your mind and see how different things can be. And I think seeing a very different future in detail is also a way to expand and see how different the world can be to break yourself out of the little red you’ve been in and sees a larger Vista
Matt: and that is the purpose of FringeFM because we have enough dystopian Hollywood movies we might at least present some of the good stuff as well.
I have one last question for you. So who is your favorite futurist or A.I. focused researcher that you look to or think it’s someone that other people should check out?
Robin: futurist or AI was well I mean very influential in my life was Douglas Linnet. That was a long time ago and he hasn’t written so much lately about the topic but it was very influential for me at the time to see the potential for what I could be and things like that and know that help
Matt: that that is how [Inaudible] .Are you more optimistic or pessimistic when you think about AI
Robin: than other people or that of the used to be
Matt: just in general do you say to the left or to the right, Optimism or pessimism?
Robin: So I think the progress will be slow. So I think people are overestimating how rapid progress will be. And I have a different set of priorities and what I think the concerns are I’m less concerned about one machine in the basement taking over the world in a weekend and I’m just more concerned about which way the world will drift and want to understand which way the world is likely to drift and ask you know what lever points there might be I think people tend to overestimate how much influence we can have on these things. The world is like a big train which nobody driving in. And it’s hard for anyone small group of people to move it a little but still, we should think about what we can do and do what we can.
Matt: Which way do you think we’re headed to bullet proof [Inaudible] a universal basic income of some kind?
Robin: I don’t think well once robots or computers actually displace humans we’ll all over the economy. At that point people will have to rely on charity or they’ll starve unless they have prepared sufficiently. So I think one of the biggest things we can try to do to prepare is to try to encourage people to set up some sort of insurance or sharing arrangements to deal with that risk. I think generically trying to set up universal basic income isn’t a very well targeted policy for that risk it would be much more effective to just have an insurance policy that pays off in that scenario because the premiums will be far lower then I don’t actually see a universal basic income happening as a generic political thing any time soon. And you might imagine that happening in response to some big disaster but then it’s kind of too late to set it up as insurance. Then it’s a matter of whether people want to help you. So I know I’d recommend that you not try to rely on other people’s charity and that you try to set things up proactively.
Matt: I like it and that’s that’s a good way to end this is just make sure that you’re focused on what you need to do to succeed in the heater. Where’s the best place for people to find you, Robin?
Robin: Well I’m on at Twitter at Robin Hanson. I have a website, Hensen, that Jim Yuda dot Edu and my Topix agent AMCOM and my other brain which we have a book we talked about the elephant in the brain dot com.
Matt: And we will have links and all that great stuff in the show nodes. Guys thanks for coming on today Robin.
Robin: Nice talking to you.
Matt: One challenge for listeners by the way. Give them a challenge not related to your books just something you want to look into do or think about
Robin: challenge to my listeners. All my generic challenge to people who are somewhat intellectual as to how can you stop pay less attention to the momentary political cultural debate and you know look at deep enduring issues like focus on the big enduring deep questions as opposed to the local talk that there’s there’s way too much temptation too to follow the current conversation. Whatever everybody’s talking about. To talk about that and that stuff doesn’t last 20 years from now you’ll hardly know why you were talking about that.
Matt: That’s the purpose of this podcast there’s much too much short-term thinking and not enough people thinking forward. Thanks for coming on Robert.
Robin: Take care.
Matt: If you want more of fringe FM you can subscribe to the podcast on iTunes or go to Fringe dot FM where you’ll find tons of audio and video interviews with leaders in the fields of genetics cryptocurrency on [Inaudible] AI space VR and much much more. And you can follow me on Twitter. It’s Matt Ward. If you enjoyed the show please leave a quick review and iTunes to help more people discover Fringe FM.
Make a Tax-Deductible Donation to Support FringeFM