Max Tegmark is a physicist at M.I.T., whose areas of expertise range all the way from the origin of the universe to the ways in which humankind is likely to destroy the world we live in. He’s creative, iconoclast, outspoken, and a genius. In other words, the perfect guest for this show.
Max TEGMARK: The ultimate nature of reality, whatever it is, is way weirder than we thought. We humans are developing evermore powerful tech. And that gives us the opportunity to cure cancer and eliminate poverty, also to commit collective suicide, should we choose.
And there’s even that little extra bonus with Max. He was initially training to be an economist, but he became disenchanted, only then turning to physics. Hopefully, I can get him talking about that topic as well. I’m always interested to hear what upsets people about economics. Maybe the same things bothered him that bother me.
Steven LEVITT: Max, thanks so much for taking the time to talk with me. I have a deep curiosity about the topics you study, but as will soon become apparent, I know very little about them.
TEGMARK: It’s a great honor to be here. And, Steve, if you feel that you don’t fully understand everything about our multiverse and about artificial intelligence, nobody else does either. So that’s quite fine.
LEVITT: Well, let’s start with cosmology. That’s the study of the universe and how it’s come to take the form it has. And as a lay person, I’ve heard of something called the cosmic microwave background, but I had no idea until recently that it’s been the driving force behind a radical transformation in cosmology. Could you explain what the cosmic microwave background is and why it’s turned out to be so important?
TEGMARK: The cosmic microwave background is, basically, baby pictures of our universe from when it was just 400,000-years-old. It’s microwave radiation, the same kind that you can heat up your food with in an oven that’s been traveling for 13.8 billion years until it reaches us today.
I think of it as the cosmic D.N.A., because by looking at this information that comes from these early moments of our universe, we can see the blueprints that determined how our universe is supposed to grow subsequently, where the galaxies are supposed to form and so on. It’s just absolutely remarkable how well those blueprints have actually managed to predict today’s universe.
LEVITT: So what you’re saying is 400,000 years into the start of our universe, there was really nothing much going on. Plasma or something was out there and that generated these waves, but now you and other cosmologists are looking at what first just appeared to be noise, right? Just interference that was making it hard to pick up other signals. And you’ve transformed that into some incredible understanding and some real, high-fidelity belief that you understand how the universe started. Is that the right way to think about it?
TEGMARK: Yeah, it’s almost easier to run backward in time because there’s always been a frontier of our ignorance. You have a very good idea of what you ate for breakfast, but you might not know the details of what happened 500 years ago. And fortunately, the sky is like a time machine, in the sense that when you see the sun, you see the way it was eight minutes ago. Because that’s how long it took the light to get here. When you look at stars at night, you see things that happened often hundreds of years ago. So someone out there looking at earth wouldn’t see us, but maybe the American Revolution.
And by looking really far away, you start to see that our universe actually looked really different long ago. The galaxies were smaller and looked like babies because they hadn’t been around for very long. And when you look really far back, you see no galaxies at all, just a bunch of hydrogen gas that hasn’t yet turned into galaxies.
Now our universe is also expanding and if you expand a gas, it’s going to get colder. That’s how air conditioners work. Which means, if we look farther and farther back in time, when the gas was ever more squished together, it would have been hotter and hotter. If you heat up an ice cube, it turns into liquid water. You heat up liquid. It turns into gas, like steam. And if you heat up that, it turns into a plasma.
There had been this prediction by George Gamow that if you look really far back beyond all that boring hydrogen gas, you should actually see a plasma screen of all this plasma that’s sitting there that’s actually opaque. And taking photos, that is precisely what these images of the cosmic micro background are. And the fact that we see them means that this story is actually validated. That our entire observable universe was once in fact, as hot as the surface of the sun, basically. Pretty shocking, but seeing is believing.
LEVITT: So, look, I can understand — wow, this is amazing. You predict that there’d be plasma there and plasma is there. But my understanding is that what happened is you look at this microwave background and it’s not uniform. It has the tiniest little variations and people like you spent years and years mapping out these variations.
And the incredible part of that is from mapping these variations, you’re able to distinguish and rule out theories that seem pretty good and rule in theories that seem crazy. You do so much with seemingly so little. That’s what I find so fascinating about this endeavor.
TEGMARK: Good science is just like a good startup company. You managed to do a lot with a little. You get much more out of it than you put into it. And those tiny fluctuations that you mentioned, that were there in the early universe, is exactly why we have life today. If something were completely uniform, it would have just stayed that way forever. And we wouldn’t be having this conversation. Whereas what actually happened was these tiny fluctuations that we can see in the cosmic micro background, they were amplified by the destabilizing effects of gravity during 13.8 billion years into the galaxies, the solar systems, the stars, the planets that we have around us today.
And the interesting thing is in order for those amplified seed fluctuations, basically that cosmic D.N.A. that we can observe with our telescopes, to turn into today, and match today’s data that we can see around us now, that will only work if you have some very specific amount of atoms around a very specific amount of dark matter, et cetera, et cetera.
So we’ve been able to use this fact that the data has to agree with what we see to measure all these things we couldn’t measure before. When I was a grad student, we used to argue about the age of our universe, whether it was 10 or 20 billion years old. Now we argue about whether it’s 13.7 or 13.8 billion years old because of these measurements.
LEVITT: So one of the things that you’ve written is that cosmology went from being data-starved to data-rich with the availability of the cosmic microwave background. And it’s an interesting parallel because really economics went through the same phase in the ‘70s and ‘80s and ‘90s, where we went from being essentially a theoretical discipline to being one that is now largely empirical.
And unlike other parts of science, particle physics has the ability to do experiments; you can build accelerators and you can crash atoms in the other ones at high speeds. Or in medicine, you have the ability to do clinical trials, where you can test the efficacy of drugs. But both in cosmology and economics, we’re forced to learn from the data that the universe gives us or that markets give us, which I think is a much tougher challenge when you don’t have access to this kind of experimentation. Would you agree with that appraisal of the situation?
TEGMARK: I definitely agree, but what we’ve seen is it still works as long as you have a lot of data. And as you also know from economics, when a scarce resource becomes abundant, it’s no longer the limiting factor and something else becomes the weakest link. And what we’ve seen in cosmology and many other areas of physics is that as we started to get these mountains of data, suddenly we were limited by our ability to analyze the data.
It doesn’t do any good if you give your grad student, no matter how smart she is, 10 terabytes of data, and tell her to look at it and come back when she’s done because you’ll never see her again. But fortunately, because of better machine learning, we have this other revolution that we can also handle these massive amounts of high-quality data that experiment has given us. And it’s those two things together — the better experiments and the better abilities to process the data that’s really revolutionized physics.
LEVITT: So one of the things I’ve touted a lot on this podcast is the power of data. But as I’ve learned a tiny bit about cosmology, it just reminds me that I’m a little bit off when I say that, because it’s not the power of data. And you really see it in cosmology. It’s combining data with theory that yields these incredible insights. Because if you had all of this microwave background data but no good theories, you’d have no idea what to look for. It’s really that the theories are pointing you in the right direction. And that combined with the data is what unlocks everything.
TEGMARK: That’s exactly right. In the last five years, I’ve been focusing on machine learning with my research true here at M.I.T. And we’re actually working on machine learning tools that make their own theories by discovering patterns in data. But we’re nowhere near what the human scientists can do, of course. And it’s exactly when data collides with theory that great things tend to happen.
Just one more little pitch for data, in economics, you can get a lot of data and you can average it together and try to get more precise. But in physics, sometimes we have these ingenious experimentalists, of which I am not one, who can measure things to ridiculous precision. For instance, the cosmic micro background, you’re looking at a signal, which is at the level of 10-5.
And to take an even more extreme example, my colleagues here at M.I.T. have measured distances between mirrors miles apart to 22 decimal places. And in the 22nd decimal place, they see this effect caused by massive black holes, way heavier than the sun, crashing into each other over a billion light years away in such a violent collision that they warped space time itself into gravitational waves, which we can then see here on earth. This is the L.I.G.O. experiment I’m talking about. And it’s just awe inspiring.
LEVITT: So up until now, I’ve been trying to draw these parallels between economics and cosmology, which is probably really insulting to you because in the hierarchy of scientific endeavors, economics definitely is well below cosmology. And I know that you started as an economist and dismissed it with disdain along the way. Do you remember what you said about economics in your book?
TEGMARK: That it’s form of intellectual prostitution?
LEVITT: Yeah, you wrote about your time in economics: “I soon grew disillusioned concluding that economics was largely a form of intellectual prostitution, where you got rewarded for saying what the powers that be wanted to hear.” Is that your current view of economics as well?
TEGMARK: Yes. But this is not, of course, a critique of economics. It’s a critique of the way politicians abuse the science of economics by only listening to the economists that say what they want to hear.
LEVITT: Yeah. I mean, you’re certainly right. And I think it is a good description of macroeconomics. And the economists that I’ve known, who’ve gone to Washington to be policy makers, come back with much the same statement. The politicians decide what they want and then they ask the economists to confirm that’s a good policy. I think it is not a good description of the everyday life of an academic economist. I muddle around in my office and nobody really cares what I do. I don’t even have to worry about prostitution because I have no demand for my services, so I’m able to do whatever things I want.
But let me say a big difference between economics and physics is you’re able to estimate these things to the 22nd decimal point. If I think about my most cited papers, the parameter estimates I come up with, the two standard deviation error band on them might go from 0.3 to 0.9. So roughly, the range of possibilities, it’s a threefold-range in what the results could be. And that’s good enough in economics and the thought that you would ever have this kind of precision doesn’t even enter into what we do.
And a by-product of that is that we just don’t take our theories that seriously because you could never test a theory to the degree that you do in physics. And consequently, it doesn’t make sense to take theories seriously. But in physics, you really take your theory seriously. If you’re off by the 20th decimal point, that’s enough to knock out a theory. Is that right?
TEGMARK: It can be, yeah. Although, if I can give a little defense of economics—
LEVITT: Okay. We’re always looking for some defense in economics.
TEGMARK: I would say that physics is all fine and dandy. It empowers us greatly through more understanding of how, the way the world works, which lets us build ever more powerful technology. So we can actually be the masters of our own destiny. But if we use that power just to ruin our planet faster and commit massacres and maybe even go extinct one day, that doesn’t make me feel very proud as a physicist. Most sciences have a skeleton in our closet. And our biggest one, of course, right now, is that we still have about 12,000 hydrogen bombs that we physicists built. Which we’ve been very sloppy with and almost had accidental nuclear wars with.
So where does economics come in here? I think it’s not enough to invent technology. We have to win this wisdom race also, where we basically make sure that the growing power of our tech through physics and other sciences is matched by the growing wisdom with which we manage this. And this is firmly relevant to economics because it’s the economical forces of society that largely determine what we do with all our tech.
If we have economic incentives that incentivize people to chop down the rainforest faster, they will. If we, instead, create economic incentives such that profit maximizing individuals will do what’s actually best for everyone, then we’ll have an inspiring future with our tech. It’s actually this idea that made me want to go into economics in the first place as a teenager.
LEVITT: So we’ve talked about how precise the measurement is in physics, which allows you to rule theories in and out, which leads you to take theories really seriously. One of the implications of taking theory seriously is that a theory that predicts a lot of things that are consistent with the data, if it predicts other things, which might seem crazy at first, you still have to take those implications seriously.
Which leads us to this whole idea of the multiverse. I don’t understand what it is, but what I find interesting about it is that nobody other than physicists believes in the multiverse, but I think a lot of physicists believe in the multiverse and that’s really, to me, interesting when people on the inside believe in something wild but people on the outside don’t. So is there a way to explain the multiverse in one minute?
TEGMARK: Yeah, sure thing. The most striking thing I’ve learned in my life as a scientist is that the ultimate nature of reality, whatever it is, is way weirder than we thought. Before even telling you why it’s so weird, let me just explain why that’s actually what you should expect if you take Charles Darwin seriously. Because he predicted that we should just have intuition for the stuff that was useful for our ancestors. So if some cave woman spent too much time pondering what happens when you travel near the speed of light, she does not notice the tiger sneaking up from behind her and gets cleaned right out of the gene pool.
And we’ve tested Darvin’s theory. We tested what happens when you move near the speed of light, and time slows down — our intuition totally breaks down. Very weird. We tested what happens when you look at subatomic particles; they can be in several places at once. Totally weird. We look at things that are much larger than our cave women ancestors had access to and space-time curves into black holes, totally weird, et cetera, et cetera. If you want to know really what reality is like, the first step you have to take is just let go of your preconception that things that sound weird must be wrong, because if you can’t let go of that, you’ll always dismiss the truth.
Now that I said that, what is it that’s so weird about our universe related to multiverse? Well, we’ve again and again underestimated the size of things. We discovered that everything we thought existed was just part of a grander structure — a planet, a solar system, a galaxy, a galaxy cluster, and then the universe. When we talk about our universe, we normally mean not all of space, but the spherical region of space from which light has had time to reach us so far during the 13.8 billion years. Okay? So it’s a very scientific question to ask. Well, is that all there is? Or is there more space?
And basically, all the evidence we have so far suggests yes, absolutely. You can wait one year and you will now see more space that came into view that we hadn’t seen before, and then you can look at the best theory we have for creating space, which actually has a name that you’re going to like, as an economist; it’s called inflation. And the most popular theory predicts that there was this funny process early on where space kept doubling its size over and over again. And the simplest versions of this theory predicted space is way bigger than what we can see. Maybe infinite.
And, that means then that there are many universes just as big as ours and inflation predicts that they typically also are full of galaxies and planets. That is a very weird concept if you start to digest it because no matter how unlikely it is that the particles would have arranged themselves in such a way that you would form a solar system, and then Max and Steve would be having this conversation, the probability isn’t zero, because it’s happened here.
So if you rolled the dice infinitely many times elsewhere far away in space, it’s going to happen again. In fact, an infinite number of times, and you could even calculate how far away you have to go until you find another conversation going on like this. Many people hate this and say, “Oh, it must be wrong because it’s so counter-intuitive.” But our job as physicists isn’t to listen to our emotions. It’s to listen to the facts, data.
LEVITT: So what share of physicists, if you had to guess, would agree with you that somewhere there is another Max and another Steve having the same conversation?
TEGMARK: There’s been a fascinating shift actually. These sort of ideas used to be very unpopular. And in fact, Giordano Bruno was burnt at the stake in Rome about 400 years ago for talking about it. More recently, all that would happen would be you would get burned on the job market if you talked about it too much. So that’s what it was like when I was a grad student. I would publish papers on this without telling my advisor.
Physicists would criticize the multiverse by saying, “This makes no sense and I hate it,” to today and they just say, “I hate it.” So that’s progress. And it’s gone from being something which was widely dismissed and ridiculed to being something which is controversial and discussed not only in bars, but at physics conferences as well. I’ve run a bunch of polls at physics conferences over the years and about 50 percent or so of physicists in certain areas, especially theoretical physics, put their money on there being some pretty crazy kind of multiverse or another.
LEVITT: It’s so interesting that it’s really only people who’ve studied the issue that don’t think it’s crazy. I’m always fascinated when insiders have a very different view of the world than outsiders.
TEGMARK: Well, I think it’s because only the insiders have looked closely at the data. First of all, it’s not such a crazy idea that space would be very big or even infinite. So if what you mean by a multiverse is this finite part of it, of course, there are infinitely many of those regions. And then when we looked in other corners of physics, like we started to study quantum physics, which powered the computer revolution, you see another kind of expansion coming into what exists.
And it’s just turned out to be incredibly difficult to write down a fundamental physics theory that predicts the existence of only the stuff that we can see and nothing else. And in that sense, the simplest kind of theories tend to predict that there’s more to reality than what we can see. And I actually think it’s arrogant to take the opposite point of view because then you’re an ostrich sticking your head in the sand and insisting that just because you can’t observe something, it can not exist.
LEVITT: Yeah, one of the arguments you’ve made, which I found really fascinating is that if you think of all the dials that had to be turned of fundamental properties and various amounts of atoms and whatnot to create the universe we’re in, that you’ve said, “Look, you have something like 12 dials and you’d have to turn each of those dials exactly right and the chances of that happening randomly are essentially zero, which suggests one of a couple of things. One is that somebody super smart set the dials.”
Okay. And that’s a sort of God hypothesis. But your conclusion is more of the view, “Look, there have to be a whole lot of universes out there for us to have gotten lucky to hit just the right dials for us to exist.” Is that a fair appraisal of what I think is a pretty compelling argument in favor of a multiverse view of the world?
TEGMARK: Yeah, that’s exactly right. The idea that there is a multiverse is by far the most economical explanation of a bunch of things we observe that would otherwise seem like very weird coincidences. You’re not shocked that we happened to be living on earth rather than Pluto or Venus. Exactly because there are many planets, you’re not surprised that there happens to be the right temperature to have water on ours. In many ways the conditions of our universe as a whole are way more fine tuned than Earth’s temperature is to allow life.
Can I add something else here also? So a common misconception is that multiverse talk is unscientific because you predict the existence of things you can’t measure and observe, but multiverses are not a theory. They’re a prediction of certain theories. And a theory is scientific as long as it makes at least one prediction which is testable. Einstein‘s theory of gravity, for example, it predicts a whole bunch of stuff that you can test, like how mercury is going to move in a funny way around the sun. And because it’s passed all those tests, we take it seriously.
Also for things that it predicts that we can’t observe, like what happens inside of black holes. And in exactly the same way, the theory of cosmological inflation predicts a whole bunch of stuff that many people, including myself, have done precision tests of and survived with flying colors. So we also have to take seriously other stuff it predicts, such as that there are parallel universes.
LEVITT: I think it’s a great point. It’s a prediction of a theory, not a theory. That’s a great distinction to make.
* * *
Morgan LEVEY: So today I’m breaking my own rule about not asking you a golf question.
LEVITT: Woo hoo.
LEVEY: I found that this one requires just enough reflection about life to be worth asking. Plus, I know that many of our listeners are golf fans, so, here we go. So a listener named Derek who is a lifetime golfer himself wrote in with an observation about the sport. He says that golfers usually are seeking to improve something about their game and carry just enough optimism to make them believe that they are capable of improvement despite physical limitations, like age or something.
LEVITT: Hey, Morgan, so I know that’s not even Derek’s question, but let me comment on that because it is so true. I have never met a golfer who didn’t, number one, believe there was some secret that if they uncovered it would make them better at golf. And number two, was not also 100-percent convinced that whatever they were going to do the next time they went to the course was the secret, even though it never ever lasts.
And it is so crazy that I’ve been at golf for 30-something years. And every bit of data thatI’ve gathered suggests that that is the wrong model of the world. And yet, I believe with a 100 percent conviction at every point in time that today is the day that everything’s going to change. It’s kind of crazy, but that self delusion has an incredible payoff because it brings me so much anticipatory joy every time I’m about to head to the golf course, I really believe and I start to feel the wonder and the excitement of how everything’s going to change today.
LEVEY: So apparently it’s all about potential with golf and in that thinking Derek wanted to know, do you believe that your competitive pursuit of the game has conferred any value to other areas of your life?
LEVITT: Oh, very little. I mean, probably nothing. Golf has been an amazing hobby for me and brought me much joy, but it’s pretty hard to find anything tangible, any real benefits.
LEVEY: I mean, what about this relentless pursuit of getting better and this internal optimism that it grants you?
LEVITT: Yeah, no, that’s all fun. I’m not saying I regret my attempts to play golf and quite the opposite, but I think Derek’s saying, “Hey, has that made you a better economist? Has that made you connections that have allowed you to accomplish something at your R.I.S.C. center you haven’t otherwise?” And I wish the answer were yes. And I actually, to Derek’s point, I do always try to turn my hobbies into money-making endeavors. And I almost did with golf.
I did sign a contract with Stephen Dubner and with Luke Donald, who was then the number one golfer in the world, and Pat Goss, who was one of the leading teachers to write a golf book, but it turned out in the end we couldn’t make anybody better at golf. And I started to write a golf memoir that I was really excited about, but we showed it to the publisher and they basically said, “This is the most uninteresting piece of writing I’ve ever seen in my life. And we wouldn’t even think of publishing it.” And in the end I had to give all the money back.
LEVEY: Well, okay, so sorry.
LEVITT: But I don’t regret for a second having spent all that time in golf because I’ve loved it.
LEVEY: I think it’s healthy for you to find something that is not productive in your life, but that you’re just doing for fun and enjoyment. Thank you so much for your question, Derek, if you have a question for us, you can reach us at PIMA@freakonomics.com. Steve and I really read every email that gets sent. We don’t have time to respond to every one, but we do read them all. We look forward to reading yours. Thanks.
I’m still hoping to cover two big topics before we finish: how life on Earth will end and the potentially massive implications that artificial intelligence will have on humanity.
LEVITT: So let me switch gears because you’ve studied many topics outside of cosmology, as well. And one thing you’ve thought a lot about are existential threats to mankind. Scenarios which lead to our extinction, or at least the end of modern life.
TEGMARK: I’m fascinated by big questions ever since I was a kid. And the bigger the better, right? So it’s pretty inevitable if you spend a lot of time thinking about our universe, that you also think about our place in it. You start to realize that humanity has so much more potential. The future is much longer than the next election cycle, and wouldn’t it be awesome if life could flourish, not just for the next 10 years, but for billions of years? And not just on earth, but throughout much of this beautiful universe?
And above all, I feel that the lesson to take from cosmology is we need more humility to acknowledge that humanity is still very young. We’ve been doing our thing as a species in our present form for only in the ballpark of 100,000 years, give or take a factor of two. And that’s nothing compared to the billions or trillions of years of cosmic future. And it’s nothing here on earth compared to the enormous resources and opportunities that are out there.
We have to be humble and realize also that because we’re a young species, there’s so much we don’t yet understand. It would be a huge mistake just to think that we already have all the answers and should use them to foreclose some of these possibilities. If we do something reckless and accidentally make all life extinct here on earth through some misuse of technology, that’s a pretty dumb thing.
Rather if you think of humanity as a child, what do you want to do as a parent with this child? You want to encourage the child to be humble and acknowledge that there is a lot of important stuff they don’t know. You want them to be careful, not do reckless things. And finally, you want to encourage your child to dream big because there’s no better way to fall short of your potential than to convince yourself that great things are impossible. And this is why I’m so fascinated by these questions that you started asking me here about existential risk and also existential hope.
LEVITT: I love the analogy to humankind as a child. And I know you’re worried a lot about existential threats and it seems from that analogy that you’re much more worried about us destroying ourselves than about some external forces destroying us in any sort of short-time window. Is that a fair assessment?
TEGMARK: That’s spot on. We know that the sun is going to get so hot that it’s going to evaporate all our oceans in about a billion years, unless we move our planet to a larger orbit. That’s an issue and put that on your radar screen.
LEVITT: Dream big. I like how you dream big. We’re already moving the planet into a different orbit.
TEGMARK: We know that we are going to get hit by another asteroid comparable to the one that took out the dinosaurs 60 million years ago, maybe another 50 or 100 billion years in the future. We can stave off that with some advanced planning. So there are a lot of things that nature throws at us, curve balls that we should and can deal with. However, on a much, much shorter timescale, there is threats that we pose to ourselves. For example, have you ever heard of Vasili Arkhipov?
LEVITT: I have not.
TEGMARK: Almost nobody has. I would actually argue that Vasili Arkhipov is the most important person to have ever lived in modern time. He single-handedly prevented a Soviet nuclear attack on the United States. There’s an amazing documentary of him on
YouTube. All these Russian sailors on the submarine that’s running out of batteries and the oxygen is low and they’re fainting. And the temperature is 110 Fahrenheit, and all of a sudden there are these loud explosions because the Americans have found them and were just dropping charges to get them to go to the surface.
The captain thought World War III had started and they were going to launch their nuclear torpedo, which the Americans didn’t know that they had. Even though the captain of the submarine he was on wanted to do the strike, if he had done that, we wouldn’t be having this conversation. We have done a lot of sloppy things already. In fact, Arkhipov is just one out of a long series of humans who have single-handedly prevented some really major disaster.
And it’s a very poor strategy to keep playing Russian roulette. You wouldn’t advise your children to not look left and right when they cross the street or dance with their eyes closed next to cliffs, and so on. It might work out a few times, but in the long run, we’re all dead as some famous economist said. So the first thing I think we have to do to have a chance of realizing these amazing opportunities is stop being so reckless.
LEVITT: So you think nuclear war is the biggest short-run threat to our way of life?
TEGMARK: The bad news is I don’t even think it’s necessarily the biggest. But I do think that people tend to worry about the wrong things. They worry too much about some things and too little about other things. For example, right now a lot of people are worrying about dying of Covid, but the probability of me actually dying of it is quite small. I calculated: it’s going to take six minutes off my life expectancy.
It’s good to take precautions. I decided to get fully vaccinated and do all the careful stuff, but it’s a very small effect. But I also estimate that the probability of me dying in a nuclear war within the next few years is probably about one percent. So that might knock half a year off my life expectancy. And I’m not talking about what the news media tend to talk about. “Oh, maybe we’ll get nuked in a sneak attack by North Korea or Iran that doesn’t even have nuclear weapons.”
I’m talking about the much bigger risk, which is simply that there’s an accidental nuclear war between Russia and America, who have about 6,000 hydrogen bombs each. This is clearly a risk because it’s almost happened so many times before. If you keep rolling the dice and it’s a one percent chance per year or whatever, the probability of you surviving into the long future drops exponentially
LEVITT: You’re a person who’s willing to put percentages on things that most people won’t put percentages on. What would you put as the likelihood of a nuclear catastrophe in the next 50 years if we don’t do anything about it?
TEGMARK: Fifty percent.
LEVITT: Fifty percent. So huge.
TEGMARK: It’s not just me. I mean, John F. Kennedy said that he thought it was 50/50 even that the Cuban missile crisis would cause a nuclear war.
LEVITT: Wait, if you seriously believe that there’s a 50 percent chance of nuclear winter in the next 50 years, is there anything else we’re thinking about other than stopping it? I mean, compare that to climate change. How would you put climate change relative to nuclear winter as something to worry about?
TEGMARK: Climate change has an even larger probability of happening. It’s going to happen more gradually. We have plenty of bandwidth as a civilization to deal with both of them. If you look at the amount of money or effort going into reducing either of the two, it’s very small compared to how much we spend on commercials. If you want to cut the risk of nuclear war to less than half, America could just unilaterally take its missiles off hair-trigger alert so that you get more than 30 minutes to decide, at least, whether something is a false alarm or not. There’s a lot of very simple low-hanging fruit we could do.
If you ask how many nuclear weapons you need to have to have an ironclad deterrence against Russia and China, well, a few hundred should be absolutely fine because nobody wants their top 300 cities turned into mushroom clouds. There are many easy things we could do, but the better way to think about it is this: We humans are developing evermore powerful tech. And that gives us the opportunity to cure cancer and eliminate poverty and do all sorts of inspiring things and also to commit collective suicide, should we choose. The challenge we’re facing is that the wisdom hasn’t really kept pace with the power of the tech.
You wouldn’t want to go into a daycare center near where you live and give them a box of hand grenades and M-16s and be like, “Hey, kids play with this.” And I feel the same when we give some of the world leaders, as physicists, a box with 12,000 hydrogen bombs and like, “Hey, play with this.” The wisdom clearly isn’t quite there to manage it. You see the exact same thing with every science. So physics, our little skeleton in the closet is the nuclear weapons.
Look at biology — they’ve done a much better job than we have because they were on the cusp of an arms race in bio weapons around 1970 and then a professor, Matthew Meselson, from down the road at Harvard did this amazing campaign. He managed to convince Richard Nixon, “Hey, we don’t want there to be a weapon of mass destruction for $500 that all our adversaries can afford.” And Nixon was like, “That makes sense.” And then we got this ban on biological weapons and all of a sudden the economic incentives were such that all the resources of biological science went into inventing vaccines and curing diseases. And we now think of bio as a force for good.
Chemistry, another example. They’ve given us all these cool new materials, our chemistry friends, but they’ve also given us climate change, where in hindsight, could have handled that better. And the particular one I’m actually most concerned about right now is computer science. It’s kind of the new kid on the block of high-impact sciences. We’re gradually building machines to do more and more of what we could previously only do with human intelligence.
And you can use great things with this. But you can also use it to power the lethal autonomous weapons that the United Nations recently revealed have been used in fully autonomous mode in Libya to hunt down fleeing humans where the A.I. itself just decides who lives, who dies. That’s obviously something that you as an economist can appreciate has just as much of a cheap weapon of mass destruction potential as bioweapons because a weapon of mass destruction is a weapon that enables very few to kill very many.
And if you have little slaughter bots, a little cheap quad-copter with face recognition and some cheap explosive and G.P.S. that is mass produced to 500 bucks, like a smartphone, you don’t need too many Nobel prizes in economics to see that there’s a huge black market for this. And soon every drug cartel and every criminal and every one with an ax to grind is going to be hoarding these things.
LEVITT: You used the example of biological weapons, and I think implicit in that argument was that biological weapons of mass destruction took an entire infrastructure of governmental supported activities. And so they could enforce that. But when I think about the scenario you’re talking about with slaughter bots, it seems to me that almost inevitably there will be technology, whether we try to stop it or not, which is cheap and lethal. Where if somebody decides they want to have some tiny drone that will have the ability to search out an individual and kill them with almost no ability to stop it, that offence is going to trump defense in this space.
Look, we see that coming, but what do you think you can do about it? I think you believe there’s something we can do about it. But I look at that and say, “What in the world are you going to do about that?” That seems inevitable and it kind of reminds you of gun control, right? Where gun control — look, it works for people who are law abiding. When you have gun control, they don’t have guns. But the criminals who really want the guns find ways to have the guns. So what do you do about the inevitability of technology that will be lethal and readily available?
LEVITT: Okay. All right.
TEGMARK: Which I think is a word that you like, right?
LEVITT: I do like that word. It’s my favorite word.
TEGMARK: So let me try to inject some optimism in this. Yes. The basic tech ingredients for building a slaughter bot? Cheap, widely available. But you can say exactly the same thing about bioweapons. In the very building where I’m sitting right now in this M.I.T. radio studio, I know multiple professors who have the knowledge that they could build bioweapons in their lab, if they wanted to. You can download certain pathogens, just the D.N.A. from the internet, then use a D.N.A. synthesizer to just make the stuff. But they don’t. Why? Incentives. It’s not just that it’s illegal, but everybody they talk to including their own grad students find the whole idea disgusting. Why would you want to tarnish the science you love, biology, by perverting it like that? If we can create a similar stigma in the A.I. community, which is well underway and happening, where people think, “This is just disgusting,” then, in addition to there being all sorts of rules and so on, there’s going to be a very great risk for anyone who tries to do it because virtually anyone else they talk to is going to rat them out. And bad things that will happen to them.
There are many other classes of weapons, which are actually pretty low-tech and pretty cheap to manufacture, like a shoulder-launched anti-aircraft missile, which a terrorist could take down to your nearest airport and shoot departing airplanes out of the sky. Good luck buying that. We managed very successfully in making those extremely hard to get hold of.
You can do very similar things with slaughter bots. You can take a rather unusual ingredient, like computer chips for drones, and you can insist that every manufacturer put a little code in the chip, so you can trace everything. And if anything bad happens, which companies are to blame. You can make sure that therefore the suppliers of these things have a responsibility to know who all their large customers are.
We’ve done this very successfully with plastic explosives and all sorts of other high-risk things. And we can absolutely do it again if the will is there. It’s ultimately economics, again. It’s not in the national security interests for the powerful nation states that you get a new weapon of mass destruction that’s so dirt cheap that every rogue actor can afford it.
LEVITT: So for sure that’s true. And maybe I’ve just watched too many Hollywood movies, but once people can make it themselves, then you don’t need very many bad guys, right? You need one bad guy. And so you take someone like the Unabomber. The
Unabomber could make very crude weapons that could hurt people. And he went out and he hurt some people. So, obviously, 9/11 was this catastrophe.
And I think there are many really simple ways for terrorists to disrupt our life, which they choose not to because the ultimate goal of terrorism is in some ways, hasn’t been to destabilize the U.S. It’s been to maybe take over Saudi Arabia or something. And so the terrorists have been looking for the big score and they haven’t used these simpler, cheaper ways of doing things to us.
They’re not flashy enough to do it, which I think has saved us from a lot of pain that could have come from terrorism. But I totally get what you’re saying. Through standard methods you can try to make it more difficult for people to exploit technology. But I still just wonder in a world in which—
TEGMARK: Can I just chime in there? Because you make a very good point. I just want to inject some optimism here, again. There’re always going to be people who want to cause as much terror as possible for whatever reason, right? The Unabomber killed
relatively few people with his bombs because he had to build them himself.
And if you have some person in the future who builds their own little ghetto slaughter bot in their garage, they’re going to build five of them or something. They’re not going to be able to build a million. So they will never be able to do a genocide and kill everybody of a certain ethnic group in a region. However, if someone wants to buy a million iPhones? That, they can do. They can place an order for it. So there lies the difference. Right now, it’s completely legal for any country, including North Korea, to build as many slaughter bots as they want and sell them on the international weapons market.
If we don’t have an international ban on these things before they are mass produced, now someone can buy a million of these little things and they can say, “I want to kill everybody who has this skin color and has done this hashtag on Facebook and whatever — here is where they live. I uploaded their addresses.” That’s the real threat. Anything home built will be on the same small scale as before. It’s the mass produced stuff that we can and should prevent by having a strong stigma.
Well, I didn’t manage to cover nearly as much ground as I’d hoped, but the great news is that we just kept talking so next week we’ll be back with the second half of my conversation with Max Tegmark. We’ll not only talk about artificial intelligence, but also it turns out that Max isn’t someone who just whines about a problem, when he doesn’t like something, he figures out a way to fix it.
TEGMARK: The key thing to remember is that this is not a depressing topic like nuclear weapons. This is actually something where we could on one hand screw up spectacularly, or it could be this incredibly inspiring future.
That’s next week. And in the meantime, if you want to learn more about Max’s view on the world, take a look at his book entitled Our Mathematical Universe: My Quest for the Ultimate Nature of Reality. This is a guy who does not think small.
People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and Freakonomics M.D. This show is produced by Stitcher and Renbud Radio. Morgan Levey is our producer and Jasmin Klinger is our engineer. Our staff also includes Alison Craiglow, Greg Rippin, Emma Tyrrell, Lyric Bowditch, Jacob Clemente, and Stephen Dubner. Theme music composed by Luis Guerra. To listen ad-free, subscribe to Stitcher Premium. We can be reached at firstname.lastname@example.org, that’s P-I-M-A@freakonomics.com. Thanks for listening.
* * *
LEVITT: I kind of want people to get a sense of you as a rule breaker.
TEGMARK: Oh, you want me to confess?
LEVITT: Yes. I want you to confess.
- Max Tegmark, professor of physics at the Massachusetts Institute of Technology.
- Our Mathematical Universe: My Quest for the Ultimate Nature of Reality, by Max Tegmark (2014).
- Secrets of the Dead: The Man Who Saved the World, documentary about Vasili Arkhipov (P.B.S., 2012).
- Unsung heroes safeguarding humanity, (Future of Life Award winners).