Peter Thiel's Keynote - Effective Altruism Summit 2013

Thanks a lot for that introduction, Jeff. And I think I'm really looking forward to having a conversation today with people about really just thoughts on how we can make the world a better place in the years and decades and centuries ahead. I thought it would be valuable to do a presentation where I sort of go through some thoughts I have for where I think the world is today and maybe why I think it's so important to try to be pushing the frontiers of technology in certain ways, both in a for-profit and non-profit context. And I want to frame the overall discussion. I want to start by talking a little bit about how I think the future of progress in the—can everyone see this, by the way? And can everyone hear me?

Yep. I want to talk a little bit about how I think the future of progress in the 21st century will look. And one of the schematics I always like to use is that I think you can think of two basic axes for progress in the world, globalization and technology. And I always like to put globalization on the x-axis and technology on the y-axis to underscore that these things are not really synonymous. They often get used as ideas that are almost interchangeable. But I think if we define technology as doing new things and globalization as copying things that work, or perhaps you can consider globalization to be horizontal or extensive progress, where you progress by going from one typewriter to a million typewriters, and technology is, you go from a typewriter to a word processor.

And so there sort of is this qualitatively somewhat different thing. And there's some balance, I think, of both technology and globalization that's needed to make the world a better place. And I want to focus a little bit on the technology piece. And I think there are very different kinds of things to focus on in certain ways. If we want to ask what is the progress of technology in the next 100 years or 1,000 years, whatever time frame you want that to look like, my claim is there are only four possible charts you can come up with or some variations of them. They're very simple. First version of the future of technology is, and the y-axis is measuring progress on any metric you care to measure it on, and x-axis is time on any time scale you care that to be on.

So I think the first one is that it has some sort of decelerating future where you eventually run into some limits, and that we're in this world of basically asymptotically approaching some kind of a limit. The second version would be that it sort of has its ups and downs. It's cyclical. You gain some, you lose some over time. Third chart is that at some point it collapses. This is sort of the apocalyptic version of technology, where science and technology are these traps that humanity has created for itself, and that maybe a la Battlestar Galactica, we should just get rid of all of it while we still can. The new Battlestar Galactica. And then finally, there's sort of the world of relentless exponential acceleration.

And I think what's very striking, if you go back through these four slides is most people don't really think it's any of these. They think it's something totally different. And all these are obviously quite counter to what we would think. We don't think very much about this one, certainly. The cyclical one, I have to confess, I myself find implausible, because I think somehow there's enough information in the world today that it's hard to imagine us actually losing anything that we've gained, and that there's something about information technology that's hard to imagine us really going backwards. So maybe the cyclical thing was true of various past societies, but at this point, if it really goes down, it'll go down all the way.

But if you sort of are left with, the non-apocalyptic ones are either deceleration or exponential, those are both quite counterintuitive. And we don't really believe either of these are likely. Deceleration seems too pessimistic. People tend to think that's not quite right. One of infinite exponential progress is sort of the world of science fiction authors. It's also the world of economists who believe that you can have never-ending compounding. And it's one that, in many ways, you'd find much of our society no longer quite believes, even though people are not prepared to say that if you don't have this chart, you must have one of the other three. So if you ask, which of these four is going on at present in the world?

I am sort of of the view that it would be best for us to be living in an exponentially accelerating world, but that it has been under some constraint in recent decades, and that maybe it's not happening quite as fast as it once was. And so, one sort of view of the future, circa 1967, the original Star Trek, was William Shatner on the Starship Enterprise exploring new planets and new worlds and new civilizations. And that's changed a little bit in recent decades to selling cheap vacations to the Caribbean and things like that. And certainly, there is a question whether something has changed in the last 40 or 50 years. This, you know, we're going to have sort of the exponential and the collapse chart, and you can sort of drill down on certain parts where maybe it's decelerating, and if you look at

per capita incomes. And the upper right chart is median incomes in sort of, again, looking at the developed world, which I think is always the part of the world that's at the frontier of technological change. You have sort of a happy exponential curve going back, starting the year zero. But the bottom right one has somewhat more of a decelerating curve. And this is basically looking at the U.S. since 1950. And by various metrics, there has been some sort of deceleration. Average incomes in the U.S. went up 350% from 1933 to 1973. They went up only 22% from 1973 to 2013. And again, all these statistics, there's a lot of complicated nuances and other things around that. But there is this question whether we may be living in a time where, in many important ways, things have somewhat decelerated.

80% of the people in this country no longer believe that the next generation will be as well off as the current generation, which is certainly very odd given the story of science fiction utopian progress that we've often been told. There certainly is a history of people thinking that maybe there are some kinds of limits to growth. And this is sort of Malthus, Paul Ehrlich, the population biologist at Stanford in the late '60s, who wrote The Population Bomb. The Club of Rome wrote this book called The Limits to Growth in the early 1970s, saying there was a limit to how much growth there could be in an economic system. And that, in some ways, the world was getting quite close to the point where there could be no more

No more real progress. There was a famous bet between Julian Simon, who was an economist, and Paul Ehrlich in 1980, whether a basket of commodity prices would go down in price over the next decade. And certainly, the bet from 1980 to 1990, Julian Simon famously won. And this was sort of taken as evidence that obviously we have tremendous technological progress and things are sort of steadily getting better. If you sort of define technology as the ability to find ways to do more with less, to make things cheaper over time, this was sort of one way you could measure this. But if you had rerun the Simon-Ehrlich bet on a rolling decade basis, Paul Ehrlich's been winning the bet every year since 1994, where the price of this basket of goods has been getting more expensive on a decade-by-decade basis, which again may suggest that we've hit some limits to growth.

And that there's something that's gone a little bit wrong with the rate of technological innovation. It's not quite generating the returns that people once thought that you might be able to generate. So I want to sort of go back a little bit to this question of globalization versus technology. And if we sort of think about the last two centuries as speculating about the future, but just shifting back and talking a little bit about the past, I think there are sort of—I'll just get this on the chart. Oops, yeah. So there are basically, I think, you can sort of divide it into these rough periods in my mind. The 19th century, from 1815 to 1913, was a point when you had both globalization and technological progress happening in parallel.

And there was sort of a tremendous amount happening on both of those. And I would argue that from 1914 to 1970, globalization went badly in reverse. It started with World War I, and it probably kept getting worse, I would say, maybe until 1971 when the US started normalizing relations with China and it sort of started rejoining humanity. While this was a terrible period for global trade or transfer of technology or all sorts of things, it was still a period of enormous technological progress. And then I would argue that the last four decades, with some important exceptions, have seen much more emphasis on globalization and much less on technology. And we've had this somewhat opposite pattern of these two.

I think that one sort of interesting way to capture the way in which the zeitgeist today is very, very different from, say, 1960, the late '60s, is there was a book written in 1967, '68, The American Challenge, by this French author, J.-J. Servan-Schreiber. It was a huge bestseller. And it basically argued that we were living in an exponentiating technological world. And what that meant was that the places that were ahead were going to get much further ahead. And the differences were actually just going to keep growing. And so that by the year 2000, the United States would have left the rest of the world in the dust. The only countries that would still be roughly competitive with the US were going to be Canada, Sweden, maybe Japan, maybe Switzerland.

And basically, the difference between the rest of Europe and the US would be as big as between Nigeria or Egypt and the US were in 1970. And that in a world of accelerating exponential progress, people in the US would have to compete less with people in other countries. And so by the year 2000, the average person in the US would be working seven hours a day, four days a week, and would have 13 weeks a year of paid vacation time. And for the first time in history, you'd have a society where people would have more days off than they were working. This was basically—we were always tempted to sort of dismiss these kinds of predictions on the basis of something like Servan-Schreiber was French or something like that.

But it is sort of a sense of how very much differently things have worked out, where globalization has happened very much, but because there aren't that many new things for people in the developed world to do. They end up competing far more with people from the developing nations, because we're doing, in some ways, very much the same kinds of things. And you have all these very difficult questions. Can a car factory worker in Detroit really compete with people from China? when they're getting paid 1 tenth as much per car. And then you can say, well, they should just work on computers and do other things. But somehow there's not quite enough to do in those areas. And so we have this world where globalization's been happening, but technology's not been happening.

And we see this in the way that the world is described, where in 1968, or late '60s, you would have described the world as consisting of the First World and the Third World. And the First World was this area where there was a lot of progress. And the Third World was this part that was going to be sort of permanently very poor and very, very far behind, which was a perspective that was pessimistic on globalization, but probably very optimistic on technology. Today we speak about the world as the developed and the developing nations, which is, of course, super bullish on globalization, because it basically assumes that the world's going to converge, the developing nations will become developed over time.

But it is also implicitly, I think, incredibly pessimistic about technology because the developed nations, if you take it literally, are those countries where there will be no more development, since they have sort of no more ideas. And so the sort of question I always like to pose is, you know, how can we go about developing the developed world or sort of breaking out of this sort of a box? And I think there are a variety of reasons to believe that this whole approach may be near an end and that we can no longer have globalization continue without technological progress at this point. And this is somewhat speculative, so I have a question mark whether this period of the last four decades came to an end in '07-'08.

And maybe the financial crisis in some ways can be understood as horizontal progress not actually generating returns. And of course, the way horizontal progress was manifest in the United States was people just built more houses. How do you grow the economy? You build more and more houses. And there are analogous versions of this going on in many other places. There was a quote from Bernanke, the Fed Chairman. This was a speech he gave in early 2007, which of course was superseded by the crisis and people don't talk about it very much anymore. But it was basically to the effect that when you looked at all of these different—people had always thought of globalization as driving prices down, because it meant that you had cheaper labor.

But at this point, globalization had shifted, leading to a world where you had more competition. And so the more globalization you had, the worse living standards would get in the United States, because it would lead to more inflation, which is this extraordinarily—it's sort of a very shocking statement from a Fed Chairman, where you're basically saying that we're living in a Malthusian world, where any more globalization is simply going to be counterproductive, and that you have to somehow find something very different to do. I would say certainly I think there are probably—my own guess is that the next decade will be marked by some very different pattern. My own guess is that descriptively it'll be a pattern where we have less globalization and we have to go back to technology.

I certainly hope we don't get the minus-minus pattern, which would be catastrophic. But I think it's hard for us to go simply back to the long globalization and long technology, because the reality is that we need to actually develop some new technology before more of it can be transferred. And I think you see this in, say, a place like China, which has been the paradigmatic example of globalization. Going back to my first slide with the horizontal X-axis on globalization, a place where you copy things that work. And what now China looks like is you have this massive pollution from coal-fired plants throughout the country. And when people say, 'What's the future in China going to be in the year 2030?'

It's going to be twice as many coal-producing factories as you have today. And that's a point where you start losing a consensus, even in a place like China, for globalization. And I think we're sort of close to a breakpoint in places like Brazil, Turkey. You have sort of this small park that gets bulldozed into a shopping center. But it's sort of, again, this idea that copying things that work has reached some sort of a limit. And probably a very different paradigm is going to be needed in the decade ahead. So this is sort of what I'm focused on. It's more of a descriptive claim of what's going to happen. But I could probably elaborate some more in the Q&A section. Now, it is worth noting that in this period of technological slowdown, the one big exception has been the computer revolution, which has continued apace.

Hardware has gotten faster and faster, à la Moore's Law. And then, of course, you've had hardware, software. We've had relentless progress on the computer side. This is probably, in some ways, a very promising place to look for innovation. I don't think it should be the only place that we look when we speak about technology. But certainly, it's a very natural one to think: are there more innovations that can happen on the computer side in the years and decades ahead, since that's the one place where innovation has continued? And at the same time, it is also, I think, always worth going back to all the areas that were abandoned. I was an undergraduate at Stanford in the late '80s. And I always like to say, people always say we need to have more engineers and more scientists in this country.

But going into engineering was a terrible decision for anybody who did it at the time. All the engineering fields were bad fields to go into. So it was bad to become a chemical engineer or a mechanical engineer, to say nothing of a nuclear engineer or an aero-astro engineer or an industrial engineer. I think electrical engineering still barely worked as a field to go into. Computer science was the one good field that was not even thought of as an engineering discipline. But we've had this world where, in the world of bits, there's been progress. Applying it to other objects, there's been much less in recent decades. And I think this is, in some sense, you can always—you can always narrow the scope of where you think progress is happening and then see a happy, accelerating curve within some sufficiently narrow scope.

But I think what's happened is that in many other areas, we've given up on this in recent decades. And it's worth looking at all these different areas. And that's what I've tried to do both with Founders Fund and with some of the nonprofit things that we've done. So I want to shift gears a little bit and say a few words about how I think we can think about shaping a future in which there's more technological progress and which we somehow go back to the future. And to go back to this technology-globalization paradigm, and one of the ways I always like to differentiate is: technology doing new things is going from zero to one—figure out the way to do the first thing of a kind. Globalization, copying things that work, is going from one to n. And there is something that's very different about these two modes of thinking about things.

And I'd say one that, I think, applies a great deal is that globalization is, in some sense, it's a statistical kind of a thing. You look at laws of large numbers, you look at what's working, and then you figure out ways to scale it. And if you've been able to build 100 McDonald's franchises, there's some way you can go about building the next 100. Whereas if you're going from 0 to 1, it's very hard to assess what's the probability that's going to work, how do we measure whether that's going to make any sense. And if we think of it in probabilistic or statistical terms, we're, we're likely not to be able to get to a very, very good answer. And I think one of the things that's just jumping ahead a little bit that's an enormous challenge in the nonprofit side is I think almost all nonprofits in the world today are focused on globalization.

And very few are focused on technology. You know, I always get pitched and asked for, you know, money in these things and sort of the easiest way I've come up with to sort of decline giving people money is to draw this chart and to explain that I think, you know, there needs to be done more on the technology zero to one side and then ask someone to explain why is your project, your nonprofit in this technology quadrant. And it always works. It's always a really safe way to get out of giving money to people.

I might have to do something else there. But I think it sort of goes to, again, that we have, I think, in many economic and profitable spheres of endeavor, we're somewhat overweight globalization. extraordinarily true on the nonprofit side where almost every single one basically is long globalization, short technology, which again I think was sort of the paradigm of 1970 to 2007, 2008. And I think one of the things that's true in many of these areas is we have this mania for measuring things. And so how likely is it that the money will be spent well? Or how likely is it this will work? And if you're dealing with something statistical, it sort of somehow is much more measurable than something where you go from zero to one.

And so we have to ask the question whether the way we're going about measuring these things is actually skewing the result in a very profound way and precluding us from exploring anything that's on the vertical end of the axis. So a quote from Elon, who is one of our colleagues at PayPal and founder of SpaceX and Tesla: "And the reason an engineer is the closest thing to a magician is because in science and technology, there's a sense in which every moment in history happens only once. The moment when a new discovery gets made happens only once in the world. In that sense, every scientific moment is, if not a miracle, it's at least a singularity, at least a small singularity that has never before happened, at least on the history of this planet."

And it's very different. And if we ask, how likely is it that it's going to happen? How do we model this? It actually ends up being a very, very tricky thing to do. How likely is it that magic is going to happen? So I want to go through some of the different challenges in it. If I had to sort of Some of the challenges of going from zero to one are things like the challenge of exceptionalism. So we live in a society where people think that if you claim that you're going to be a magician and you're going to do something that's never been done before, do you have a right to claim something that different for yourself? People tend to be quite skeptical of that. We have an educational system where we believe that all truth is collective and that the true answers are the answers that everybody knows to be true.

Whereas if you're going from zero to one, you're coming up with a truth that nobody else knows at that point yet. And then I think there is always this question about how do you explain this and how do you pull people in when you're trying to do something that's very, very new. But I think there are a number of features in our society that have made it unusually hostile to this idea of going from 0 to 1, and therefore made it default very much from 1 to n. And to sort of get at this from, again, one slightly different angle, I've often thought that there's sort of this two-by-two matrix on the future. And you can think—you can be, sort of simplest, this is a McKinsey-style matrix—you can be optimistic, pessimistic, definite, or indefinite.

And so: optimistic, future will be better than the present. Pessimistic, the future will be worse than the present. Definite, you have some very definite ideas of what the future will be. And indefinite, you don't know, and the future is sort of fundamentally unknowable and uncertain. And so if you're definite, you'll be able to act with conviction. Indefinite, the correct strategy is always a strategy of diversification. If you're optimistic, you will look to the promise of the future. If you're pessimistic, you'll be afraid of the future; you'll figure out ways to try to stop it from happening. And sort of one—one frame for putting different countries in these categories is, I think, the United States in the '50s and '60s was optimistic and definite.

It thought the future was going to be better in very specific ways. There would be faster planes, faster cars, faster computers. There would be all these ways that, in very measurable ways, living standards were going to improve year after year, decade after decade. And there were very specific things you could do in that world, like the Apollo program, to make it happen. There were definite goals, and you could direct resources to them. And then when you succeeded in those goals, you got towards a better world. In the sort of optimistic, indefinite quadrant, I sort of put the US from 1982 to 2007, where people would say the future was going to be better, but we didn't really know why or how.

It would be sort of stupid to ask specific questions, but we had processes that would automatically make it better. And so it would be like the stock market would go up, um, in a, um, in a sort of happy northeasterly direction over time. Um, you couldn't actually pick any stocks; it was a statistical thing. You could just invest in the index. Or, um, or there were ways the economy would grow. You know, like, why would it grow? What should—what should the economy be doing? Can't ask those questions. That's too, uh—that's way too definite. No room for planning. It's just this sort of statistical, indefinite process where an economy grows on its own. And it's just people can make small, little micro-decisions, and somehow they magically add up.

China is debatable, but I tend to put it in the pessimistic and definite quadrant today because it thinks that it has a very definite view of the future. It knows what the country will look like in 20 years, and it will be old before it is rich. It will be sort of a poorer version of the developed world and a much more polluted version, but it's very, very clear what that looks like. And then the sort of indefinite, pessimistic version has been probably Japan since the 1990s and Europe at present, where the future is sort of worse, but you don't know what to do, and you have this general malaise. You know, there's sort of different types of disciplines that I think tend to dominate in these contexts.

And so, you know, the definite optimistic one is it's driven by engineering, science, art, specific visions of a better world. And those are the kinds of careers that are seen as the most valuable in a world of definite optimism. In sort of an indefinite optimistic world, I think is one that's dominated by finance and law and sort of procedural careers, not a process over substance in a way. And this was what characterized the US from 1982 to 2007. Definite pessimistic is something like some sort of rationing regime. You say this is what China thinks the world's going to run out of oil, so it's trying to secure the last remaining oil by getting access to all the resources in Africa. And you sort of do things like that if you have a definite and pessimistic view of the world.

And then the indefinite pessimistic one is you just want to be safe. You want to get insurance. Things are going to get worse, but you try to get insurance. And there's a variety of strategies that are insurance-related that end up dominating in the pessimistic and indefinite world. One of the thoughts that I do have on this—and there's a lot more I can say on all these slides; I could probably spend about five, six hours trying to go through—I'm going through much greater length material than I'm going through today. One general thought I have is that of these four quadrants, the upper right quadrant may be a fundamentally unstable quadrant. And so that an optimistic, indefinite world isn't actually possible in the long run.

So if you think the world's getting better, but you have no idea how it's getting better, and if nobody has any idea of how to make it better, then maybe it's actually not going to get better. And that it's possible that in a world which is indefinite and optimistic, it will eventually drift into the more pessimistic direction unless people go back to finding some specific visions and ideas that enable them to plan for a better future. One sort of paradigm in this optimistic, indefinite world is that money ends up being sort of this very dominant thing. And you can think of money as the ultimate indefinite good, because money is pure optionality. You can do anything you want with it. So in a definite, optimistic world, money is a means to an end.

There are very specific things you want to do with it, and it's a means to an end. In an indefinite, optimistic world, you end up with it getting reversed, where basically various things become means towards getting money. And money is the most valuable thing because it's pure optionality and you always want to have more options. And so you always try to get things back to being in a monetary form rather than it being a means to an end. And so one of the crazy dynamics this can lead to is sort of like a startup version of this is: you have a company. You sell the company. You start a startup. You sell the company. You make money. What do you do with the money? You don't know, because you have no ideas.

We're not going to ask the question how you started the company. So let's just say you had one idea, but you're out of ideas now. So you sell the company. You make money. What do you do? You don't know. So you give the money to a large bank to manage it. What does the bank do? It does not know what to do with it. But it knows, because you don't have ideas, you're always supposed to diversify. So it gives the money to a large number of institutional investors. What do each of the investors do? They don't know what to do with it, so they invest it in a large portfolio of stocks. It has to be well diversified, because if they put too much money in any single company, that would suggest they had convictions and ideas.

And we can't have that, because we know the future is fundamentally indefinite. And then what do each of the companies do? They don't really know what to do with it either, but they just try to generate more cash flows. And one of the upshots of it is that eventually the interest rates you earn on money go lower and lower because people have no idea what to do with it. The money builds up, but people can't figure out what to do with it. And this is the real interest rate, which is the amount of interest you earn after inflation. It's now less than 0% on a 10-year TIPS bond, which is a way you can measure your expected returns after inflation. And that's, I think, sort of what logically happens when you have a buildup of money and people no longer have any ideas of what to do with it.

So I think this need to go back towards a definite, optimistic future in all these different areas is sort of characteristic of this crisis of globalization having run its course and a need to develop specific valuable technology. So there's sort of many different ways to get at this idea. So let me end with a few thoughts on: if we want to go back to an optimistic, definite future, where does one start? And how does one go about this? So I always like framing it as a question of this threefold question one should ask: What is valuable? What can I do? And what are others not doing? Because we're interested in 0 to 1 now, not 1 to n. And so 0 to 1, it's the intersection of this diagram. And it turns out

to be relatively small as a Venn diagram intersection. And I think it's worth exploring a little bit why that intersection is as small as it is in practice. I think there are several ways you can formulate the combined question. And so you could say the key question, the intellectual version of it, would be: 'What important truth do very few people agree with you on?' The business version is: 'What valuable company is nobody building?' And the altruism version: 'What valuable nonprofit activity is nobody doing?' These are roughly analogous questions. I normally like just asking the intellectual one as an interview question. And it turns out to be an extraordinarily difficult question for people to answer.

My slightly flip answer when people ask me, 'How do I answer the question, "What important truth do very few people agree with you on?"' is that the answers always take the form of: 'Most people think X, but I think not X.' Most people think that the top question is a very easy question, but I think it's a very hard question because people are generally not able to answer it. And so there is something about going from zero to one that's quite hard in all these different contexts. And I think there are—there's several different reasons that it's very hard to do. This is sort of a photograph of a two-week-old baby emulating an adult making various faces. And I think one of the reasons that—sticking tongue out, opening mouth, pouting—

And one of the reasons that zero to one is hard is because we are all sort of very deeply mimetic and we end up, you know, the word 'ape' already in the time of Shakespeare meant both primate and to imitate. We're deeply imitative. We shouldn't pretend that we're not, but it is in some ways a powerful constraint on being original. And we all like to think that we're very original, but in many ways the sort of imitation is an extremely important part of who we are as human beings. I think it's how we learn languages, it's how we acquire all sorts of cultural skills, but then there is a question: When we're trying to do new things, we have to be aware that we often find ourselves in these social contexts where we're not actually doing this and we've been sort of programmed to copy one another in one form or another.

So if I had to go back to these questions, go back to the intellectual version: what important truth do very few people agree with you on? One of the reasons this is such a hard question to answer, I think there are two. One is that maybe you have to be really brilliant. This is sort of the post-doctoral grad school view of it, that there's sort of a line of research and you have to pursue it just further than anybody else has. So it's really, really hard to get to the frontier of knowledge. And that's certainly a very common reason that people would think this is a hard question to answer. It's like, well, maybe I can answer it after I've been working on superstring theory for 20 years and studied that field.

The frontier is really far away, and it seems very, very hard to get to. But I think the other reason that it's actually a very tough question to answer is not so much metaphysical, it's more political. It's an awkward question to answer. So if you actually imagine an interview context where you're being asked that question, the correct answer is presumably one the interviewer doesn't agree with. If you tell the interviewer something the interviewer wants to hear, it's almost certainly the wrong answer. And so you can answer it maybe if you're super brilliant or somewhat courageous. But there's a lack of incredible brilliance. And there's probably a lack of nonconformity or courage or something like that.

And so it's, again, these sort of mimetic social conventions. We pick up cues. We figure out what the answers are that people want to hear. And then we give them those answers. And I think for this reason, this is extremely hard to get in any sort of context. And why, when people actually pursue something on a Zero to One context, it can often work so powerfully. It would be hard to get people started, hard to convince other people of these things initially, because a lot of people would be looking for social cues. Is this going to work? Is this not going to work? And the sort of mimesis effects are going to be very dominant. One final sort of version I have on, one final cut I have on how to get at this Zero to One question is we can have a trichotomy of how hard are truths to obtain.

And we can say, you can sort of have a trichotomy between truths that are easy, difficult, or impossible. And we don't really want to go after easy truths because we think lots of people already have those. So you're not going to get Zero to One if you get something that's too easy. And if you go after something that's impossible, you're not going to get Zero to One either. Because if it's impossible, you will not figure it out. So we want to find truths that are just at the right level of difficulty, that are hard but not—that are hard but not impossible. And I sort of labeled easy truths as conventions, impossible truths as mysteries, and then difficult truths as secrets. And so one question you can ask is, how many secrets are there for us to discover in the world?

And if you say there are very few secrets left to discover, that all truths are either easy or impossible, then you're basically in the long-globalization, short-technology zone because we've figured everything out we're going to ever figure out. We might as well just figure out ways to copy things and get them to work. Whereas if you think that there are many secrets left, then there is perhaps a lot of room for Zero to One. There's a lot of room for those singular moments in history where new sciences, technologies will be discovered, developed, created. So ideally, we're interested in secrets that are not small, not silly, not esoteric. We're not talking about gossip. We're not talking about figuring out the secrets of tarot cards or something like that.

We're talking about ones that are big and somehow going to change the world. This is sort of a little bit of a jumbled chart, but if you ask, how do you go about thinking about what are promising areas to look for secrets to find? I sort of have this—one of the frames I like to have is that there are—there's sort of metaphysical secrets and there are political secrets. And you can say metaphysical secrets are secrets about nature. So this would be questions like, is string theory the fundamental law of physics or something in that sort of a way? And then there are political secrets, which are secrets about, sort of, the psychosocial dynamics in which our society is caught up in and in which sort of a lot of the decision making is mimetic in ways that are perhaps warped in one bubble or another.

And I think there are many important truths you can get at either way. But my typically preferred shortcut in the way I like to think about these things is to try to ask these sort of psychosocial questions and to think of things like psychology, sociology, anthropology as very foundational, and that if we understand what are the kinds of questions people are uncomfortable asking in our society, what are questions we cannot go after, maybe that's the place we get to some very low-hanging fruit in identifying some places to go from zero to one. You can probably get at it from the other angle as well. But I found this to be quite a useful one. Let me just end with three applications in the nonprofit side.

The first, you know, psychosocial truth I believe that exists is that we have something like a bubble around education. This is seen as an absolute imperative in our society. People talk about this nonstop in all sorts of contexts. There is obviously something about it that can be very conformist, where people end up—a huge mass number of people studying for an exam, where it's a way to get people to do all the same things, to think in all the same ways. There's been this history of very strange bubbles. I think one bubble has been this educational bubble. And so it's been worth thinking about, are there ways to do things that are different in education? And bad education, again, looks like some sort of mass statistical product.

What I think good education—it's hard to know exactly what it looks like, but it probably looks like something where everybody gets educated in a different way that's unique to them as a person, and where you do not draw from the reassurance that you're doing the same thing as everybody else is doing. And I don't want to make it too, too Asperger-like, but maybe there's an autodidactic component to it. There's something where, you know, you're super passionate about learning about some idiosyncratic subject that, you know, and then you really, really try to understand that. And it's not a question of how many other people are studying this, is this going to help me do better on the verbal SAT test score, or things like that.

It's something that you're just really interested in learning. And so it is unique and personalized, as opposed to psychosocially conformist. A second one that I think is an area where we have some really powerful psychosocial things at work is the aging and longevity set of things. But we don't really want to think about aging. I'll go to the happier slide. But there is probably, when we talk about death and dying, this is probably one of the things about which there's more psychological denial than any other topic that one can think of. I got involved in helping Aubrey de Grey with the SENS Foundation in 2006. It was a modest donation. It was about $350,000. Managed to get on the front page of the San Francisco Chronicle.

My parents got the newspaper. They were really upset at me. They called me up. It was like, "What will the neighbors think of this?" And it—and there is something very, very strange about the longevity, life extension debate, where you go through all the different... I think there's probably almost this general logical principle that when people—when people, when you have a lot of different, very bad arguments, there's reason to think that something psychosocial is at work. So, it's not good for people to live longer because they'll get bored. It's not good for people to live longer because it's not natural. It's not good for people to live longer because they—because, yes, every myth on this planet teaches us that the meaning of life is death.

And so on and on and on down the line. And so I do think that it's—it's very, it strikes me as grotesquely underfunded. I mean, I could be wrong about this, but when you look at every single—every major disease is sort of linked to aging in one form or another. You have a one-in-a-thousand chance per year of getting cancer at age 30. You have a one-in-10 chance per year of getting it at age 80. Something like this is true of, you know, all major diseases. And so trying to understand the phenomenon, trying to do some basic research on it, trying to figure out how to reverse it seems like a very foundational sort of question. And I think, again, this is one where there are these very powerful psychosocial forces that are stopping us from looking.

And that's why I think that this may be a place where there's some relatively low-hanging fruit rather than where it's impossibly hard to figure out in one form or another. And the last one is, again, the sort of hostile AI versus friendly AI. We'll again end with a happier slide. And again, we sort of debate that our imagination with friendly AI is somewhat limited as well. But I think that if we go back to the 0 to 1, 1 to n paradigm, I can give some general reasons why I think AI seems at least plausible in the next few decades. There's been this very fast improvement in computer power. There's sort of many areas where it's gained ground over time. I used to play scholastic chess in 1980 in junior high school.

It was at a tournament here in Northern California. A former world chess champion gave this exhibit. A former number two world player, Max Orbe, gave this exhibit. One of the people in the audience asked him, it was a Q&A, one of the people asked him, will a computer ever beat a grandmaster? Answer, definitely no, because chess requires such a degree of creativity that it's something only human beings can do, only human grandmasters are capable of. The chess computers got better at the rate of about 100 rating points a year. They were at 1,300 on the USCF scale in 1980. They sort of got to roughly my strength by 1990. They beat Kasparov with a deep blue in 97. But then what happened after, and they've gotten still somewhat better.

They're probably at about a 3,100, 3,200 scale today, and would probably beat the top player in the world about maybe eight wins, maybe two draws out of 10 games, something like roughly in that ballpark. And when people ask today, but the conclusion is not, "Wow, AI is getting really close." The conclusion is, "No, we didn't really understand it. We thought chess actually required intelligence. And actually, chess is just this stupid sort of artisanal activity. It's like a hipster making coffee in San Francisco or something like that." And so it really has nothing to do with intelligence at all. And so I'm inclined to—since I'm still somewhat pro-chess, I'm inclined to say that this is a small version of this sort of psychosocial denial, I think, sort of a—

Another way of getting at it from my presentation is if we're in a society that's very overweight 1 to n and very underweight 0 to 1, probably the biggest 0 to 1 thing would be to get something like generalized artificial intelligence. It would change the world in ways that are more radical than we can imagine. It would be basically analogous to extraterrestrials landing on this planet, which would be probably the most important event in the history of our world to date, if that were to happen. I think it's unlikely extraterrestrials will land on this planet in the next 50 years, because they haven't done so yet. It's sort of the Fermi Paradox, a version of the Fermi Paradox. On the other hand, the case for AI seems much stronger than the case for extraterrestrials.

If you ask me to measure it, to put probabilities on it, to tell you when this will happen, I actually don't think the math does a very good job and so therefore when people insist on measurement, you end up with a problem that does not get tackled and gets very badly underestimated even if it is potentially the most important zero to one problem there is. So I think there are sort of all these different, you know, I want Maybe just, I can go on and on all these different things, but maybe just open up some questions. I think, just in conclusion, I do think that we're not at a point where technology's over. We've been in this sort of 40-year lull. It's time to start re-accelerating it. There are a lot of secrets out there for us to find, and then in finding them, to find application in creating great businesses, great non-profits, great opportunities for altruism in many forms, preferably definite forms.

Thank you very much.