Bill Gates is one of those people for whom "needs no introduction" is actually true. The polymathic Microsoft founder now leads the world's largest and most important private foundation, and he's using that platform to predict that we're finally on the cusp of the clean energy breakthrough that's going to save the world.
I spoke with Gates about a week ago for nearly an hour. We talked about whether innovation is actually enough to stop climate change, why rich countries need to drive their carbon emissions down to zero, the controversial idea that technological innovation is slowing down (Gates disagrees with it, to say the least), how close we are to true artificial intelligence, and whether Americans are still good at taking intellectual risks.
A transcript of our conversation, edited for length and clarity, follows. You can listen to the full discussion on my podcast (subscribe here!).
Ezra Klein: Why don't we begin with the big prediction in your annual letter. You write that within 15 years, "I expect the world will discover a clean energy breakthrough that will save our planet and power the world."
We have a lot of technologies that in theory could power the world, but building the infrastructure necessary to scale them to that level is really difficult — and, at this point, cost-ineffective.
Your paper frames the problem as essentially an issue of technology and technological innovation. But it often seems that the problem of energy is more of commercialization, of deployment, of political will. A great electric car has to be unbelievably incredible to compete with the fact that there's a gasoline station every 50 feet. So is it really a question of innovation, or is it a question of somehow unwinding infrastructure and political path dependence that we've been on now for at least a century?
Bill Gates: When you say to India, "Hey don't use your coal, use something that is substantially more expensive," you're asking them to make a trade-off against uplifting those lives to have the things that we take for granted. If they develop with coal, they still will have emitted less per person by a factor of four than we have over the last 100 years.
So when you really look at it that way, if you don't have innovation you're not going to meet any climate goal that's out there.
Today's technologies — in terms of reliable, delivered energy — are at a huge premium, even with all the nice tax subsidization we're doing. People shouldn't lose sight of the fact that until you have a storage solution, until you get an economic [solution], asking India — on behalf of the rich world that's done so much CO2 emission — to hold back on electrification, that's not likely to happen.
EK: Let me ask you about what makes you so optimistic. Can you give me a historical analogy for a breakthrough innovation in a sector with this much infrastructure built upon it that was able to roll out globally, or even semi-globally, at the kind of speed we would need to really make a dent in our carbon emissions in time?
Because once the carbon is out there in the atmosphere, it's very hard to get it out. So if we invent the future here in 2150, it may already be too late.
BG: That's right. You can't wait till 2150 to have an invention. You really need an invention sometime in the next 10 to 20 years, and then scale that up and make it a significant part of the energy system over the next 20 years after that.
The rich countries have to get to near zero. And that's why I go through the equation, because some people think, "Well, you know a little bit of efficiency here, a little bit less consumption there," somehow makes that equation work, which it doesn't.
An example is that in the early 1900s, New York City was buried in horse manure, and people said, "Okay, should we let the horses only go out on odd and even days?" And they had barns and whips and saddles and people were trained, and there was a huge infrastructure. But that problem was solved by the automobile.
EK: You mentioned your climate change equation. I think it's an interesting way to think about it — in terms of breaking down the problem, but also how revolutionary the solution would need to be. Do you want to walk through it?
BG: Yeah, it's important for people who care about climate to not think it's easy to solve.
The equation is: How many people are there? And that's P, which today is about 7 billion, and will grow to be bigger than 9 billion.
Then you take how many energy-related services each person takes advantage of — that's heating, cooling, transport, lighting. We call that S, and that will go up quite a bit as poor people in India are getting lighting, air conditioning, refrigeration. The average number of services used by a person will increase, and it should — that's a very good thing.
Then you have E, the energy used per service. In some areas, like lighting, that number can go down a lot. In some, like transport, planes, making fertilizer — those processes are extremely optimized, and so there's not that much room to innovate on the energy-per-service front. Even if you're optimistic about that, maybe you'll get to 0.6. That is, 40 percent more efficient across all services.
And so if we take these first three factors — 7 billion going to 9 billion, double the services per person, and efficiency at about 0.6, that's increasing [emissions].
The last factor is C, the carbon per unit of energy. And so if you multiply today, you get 36 billion tons. And if you multiply in the future, you need to get zero.
And so the first three factors are not going [to change] — the first one is going up; the second one, hopefully, is going up; the third one is going down, but not enough to offset those other two.
You have to take transport, industry, household, electricity — and, at least in the middle income and rich countries, put it into a zero emission mode.
Economists say technological innovation is slowing down. Gates disagrees.
EK: Something your technological optimism raises is this argument that you hear now from people like venture capitalist Peter Thiel or economist Robert Gordon that the fundamental rate of technological innovation in our society has really slowed down compared to the mid-20th century.
They would say: We're seeing big innovations in IT, but in things like energy, transportation, medical care, and so on, the rate of technological innovation has slowed. And certainly Gordon thinks that's because the next set of breakthroughs are just not nearly as easy to make. Do you disagree with that argument?
BG: Totally. Yeah. I couldn't disagree more.
The advance of science, in terms of the basic understandings that we have today and what that allows us to do in materials science, medicine, energy, and continued IT advances, is even more profound than in the past.
Energy tends to come in big chunks. So you get coal, you get natural gas, you get nuclear. The big improvements come in a discontinuous fashion, because it's a large-scale system that has to operate with unbelievable reliability.
But, no, the medical advances — you just have to be blind not to see that we're making bigger advances there than ever before. I was just down seeing a robotic surgery company, which was pretty phenomenal, and our foundation is very involved in drugs and vaccines and the basic approaches. You should want to get sick 20 years from now, versus getting sick now. Honest. It will be radically improved.
In material science, for the first time, we understand alloys; we understand catalysts; we can simulate those things. A company like TerraPower, which happens to be a nuclear fission company, is able to create a fourth-generation design and actually see what happens during a Richter 10 earthquake or volcano or tidal wave.
So it really is amazing: the number of engineers in the world, the tools those engineers have, our basic understanding about solids and alloys. I'm amazed that somebody can miss that.
EK: Is your view here that we have for the past, say, 20 or 30 years been doing really fundamental, important, basic research and we're now at the point that that's going to pay off?
Because I think the skeptics here say: Well, look, we are not seeing life expectancy increases in the developed world that are anything like what we saw in the 20th century. We don't move people faster now than we did, you know, 30 or 40 or 50 years ago. In some ways, with the end of the Concord, we move them slower.
There's a lot of exciting, conceptual work you read about all the time — sequencing the genome, all this cool materials research —and yet the big leaps forward just keep not coming.
Even in terms of drugs, there's a real question in the pharmaceutical industry, about why the rate of blockbuster breakthroughs and genuinely new molecules being discovered has slowed. And we're seeing a decrease in the number of effective antibiotics we have.
So tell me a little bit more about why you're optimistic there, because I do think it is a little bit of a counterintuitive view.
BG: Well, take something like agricultural productivity. The private sector builds better hybrid maize seeds every year, and we've had 2 percent compound growth — which is a pretty phenomenal thing.
And that's one of the most basic processes you can imagine. That's just pure seed productivity. Now, with better weather prediction, better automation, understanding exactly how much fertilizer to put in each place, that's a yielded benefit on top of what the basic seed improvement is giving us. And that's why, with 2 percent of the population, we're the biggest agriculture exporter in the world, in addition to feeding our people.
Under-5 mortality has gone from 10 percent of all kids in 1990, over 12 million a year, to under 6 million a year now. Literacy has skyrocketed; abject poverty has been more than cut in half.
Meanwhile, what the very rich are experiencing in terms of how they select and share music; how they select and share photos; how they know where to drive to, or hotels to pick; or staying in touch with friends — the way the productivity figures are done isn't very good at capturing those quality-of-service-type improvements.
And, yes, in terms of the very big medical breakthroughs on cancer, I can't say if that's in the next 10 years, in the next 20 years, but it will come because of the understanding we've gained.
But ask somebody who's gay, or ask somebody who uses a really cool service to stay in touch with people far away: Would they want to go back 10 years or 20 years? I don't think so. I think the way that we spend our time has changed as much in this time period as any time period.
EK: One of the fascinating things I think about the productivity number debate, which you referenced here, is that it feels like technology has changed so much, but we don't see it in in the numbers that should reflect those changes.
Something that economists who study this told me, and it helped change my thinking about it, is that it's always been the case that the productivity numbers miss things — it's hard for them to measure better ways to spend our time or the improvements built into new products. But they have always missed that. And so the question becomes: Is there some reason to believe they're missing more of it now than ever? Is there some reason to believe that the difference between being alive in 2015 versus 2000 is way bigger than it was from 1995 to 1975?
And I think the argument of Gordon and others is that as much as this stuff has felt very big, that if you walked into somebody's house, more is recognizable now from 30 or 40 years ago than it was 30 or 40 years before that.
BG: He's so wrong about that. You used to have an encyclopedia on the shelf; you used to have all your albums on the shelf. I mean, does he really think nothing's changing? You know, you used to have three networks that you used to watch. Binge-watching — it's a new invention.
Part of the reason that you miss it in the statistics now is that if something enters the economy at a noncontinuous price, then understanding the substitution-demand curve for that good just doesn't happen.
So let's say Wikipedia had come in at $100, then $90, then $80, $70, $60, $50, like that, and you could see: Okay, people are switching from Britannica and World Book to this thing. Now they've switched this thing — and look, they're paying this amount, and that price is going down.
The way the time series would have happened would have been utterly different, because you would have seen the substitute demand from the narrow set of buyers and gotten a sense for that substitute value.
When things come into the economy at free prices — like the way you manage your hobby, the way that markets clear, where you find restaurants and hotels in different ways — you don't see that at all.
You know, something like the sharing economy — you actually see that as a negative, because people are building less capital assets even though you're using them more effectively.
So, yes, there are some systemic things about the way economists measure things, but more importantly the big benefit of the scientific understanding of the last 20 years in terms of gene editing, biology, machine learning, antibody design, material understanding — we will see the dramatic effects of those things over the next 20 years, and I say that with incredible confidence.
What new material science means for how you build things, how long it lasts — everybody's talking about re-maintaining infrastructure. You know, we're going to be able to build infrastructure that lasts 10 times as long as the cement rebar approach that we've had today, and we'll be looking at — to some degree — the supply-side magic of machine learning, vision, and robotics.
I promise you, that book [Robert Gordon's The Rise and Fall of American Growth] will be viewed as quite ironic. It's like the "peace breaks out" book that was written in 1940. It will turn out to be that prophetic.
The three phases of the robot revolution
EK: I know you take the risk of creating artificial intelligence that ultimately turns against us pretty seriously; I'm curious where you think we are in terms of building artificial intelligence.
I know there's a lot of disagreement in the field about, are we 40 years away? Are we 500 years away? What do you think is the state of AI research right now, and when do you think it will really begin feeding back into the economy and into innovation?
BG: Well, with robotics, you have to think of three different milestones.
One is just pure labor substitution for jobs that are largely physical and visual manipulation — driving, security guard, warehouse work, waiter, maid. That threshold — I don't think you'd get much disagreement that over the next 15 years the robotic equivalents in terms of cost, in terms of reliability, will become a substitute to those activities. So that's the first stage, and you'd get less variance in that prediction.
Then there's what we think of as "intelligent activities" — things like writing contracts, or doing diagnoses, or writing software code. When will the computer start to infringe? "Infringe" is a pejorative word. When will it start to have the capacity to work in those areas? Some might say 30 years — I might be there. Some might say 60 years. Some might not even say that.
Then you have the sort of third threshold, where the intelligence involved is dramatically better than humanity as a whole, and there you're going to get a huge range of views, including those who think it won't ever happen. You have Ray Kurzweil, who says it will happen, I think, at midnight on July 13, 2045, or something like that, and it will all be good. Then you have people who say it can never happen. Then you have the group that I'm more among, that says, "Okay, we're not able to predict it, but it's something that people should start thinking about." We shouldn't restrict activities or slow things down, but the potential that that exists — even in a 50-year time frame — means it's something to be taken seriously.
But those are different thresholds, and the responses are different.
Is America still creative enough?
EK: Something you talk a lot about in your letter is the need to try genuinely wild, crazy ideas — ideas that might just fail. Do you think our culture, or our educational system, undermines that kind of creativity and risk taking?
BG: Well, certainly the US is probably the best at having risk capital and risk-taking activities. The mixture of our universities, our national labs, our venture capitalists; the fact that smart people from all over the world, to the degree we allow them, want to come in and join in these activities. This is the place where, on all the fronts we've been talking about — energy, biology, IT, robotics — you name it, the US has an amazing share; not 100 percent share, but over 50 percent share in almost all of these areas.
And I've been impressed that that's something that's been maintained. Though the US overall education system has severe failings for low-income people, for the people who get a good education in the US — who get to go to one of the, I don't know, top 50 to 100 universities — the amount of creativity is the envy of the world.
And, yes, China is imitating some of those things, but it takes 30 to 40 years, and we'll see how far they get. Of course, their innovation improvement just adds to what I just discussed about my view that there's a lot of innovation net. Even with 90 to 95 percent being dead ends.
EK: One thing that you talk about in the letter is that government has a really important role in funding some of these basic innovations. And that seems true to me, if you look at the history of these big innovations.
But one version of how you'd want to see that happen, potentially, is that government would be funding things that are big ideas that are likely to fail. Because it's those kinds of big, weird ideas that might not pan out that are where the market is likely to fall short in terms of directing capital.
But then you have something like Solyndra, where the government funds something that doesn't work out and it becomes a huge public relations scandal and people feel like the taxpayer money is being misspent. Every year there are members of Congress who will put out these rewards for the most ridiculous uses of government money where they make fun of research projects that sound funny when you write them down.
Do you think there's a tension between the role a government should be playing and the role taxpayers are comfortable with it playing in terms of its funding of basic research and of science?
BG: Well, basic research fortunately has got quite broad support, and it has success stories like the internet that are pretty fundamental.
People can argue about shale gas, what the relative role of the private sector versus the government was, but certainly in terms of some of those advances there was government research money. Nuclear — the government played such a strong role it was amazing.
What's being discussed today is an increase at the basic research level. So you may never get a consensus about cheap loans once you get into that deployment phase, which is why I really do believe the private sector, including the group I'm pulling together and the Breakthrough Energy Coalition, should step up for the risky part of that, because I don't think you can count on government to get into that.
Gates's book recommendations: The Better Angels of Our Nature, Sapiens, and Seveneves
EK: I know we've only got three or four more minutes, so let me ask you a more random question to close: You're a legendarily big reader, so what is the book you would recommend to someone who is 15 years old, somebody who is 25 years old, and then to someone who is 45 years old? And why?
BG: Well, the book I'd recommend the most ever since I read it is [Steven Pinker's] The Better Angels of Our Nature. It's a tough one because it's longer than your average book. It's about 700 pages, and so there's quite a bit you have to read there. But I don't think that would vary by age. I guess it would be a pretty intense 15-year-old who would go and read that; then again, a lot of people read less as they get older, which is too bad.
EK: What is Better Angels of Our Nature about?
BG: It's about how humans have treated each other over time. It educates you about instruments of torture and how demand for those things seems to have fallen off. It looks at the falling rate of violence over history. Why have we been able to achieve that? Why, when you're dealing with lots of strangers, has the violent death rate come down so dramatically? What is it about morals, taboos, systems of laws, and shifts in attitudes toward violence over the last 50 years?
It's a pretty amazing piece of work. But I don't know how many of the people I recommend it to end up reading it.
EK: And then if it's not going to be by age, what are two other books that you've read over the last couple of years that you would recommend to an interested audience?
BG: I had a while where I hadn't read science fiction, so a friend said, "Now you've got to read some of the latest Neal Stephenson stuff," so I read Seveneves, which also is not short.
It's very clever — gets you to think about all sorts of situations, including what will happen with technology. Now, I have to read his other ones. I don't know how they compare — people say they're all really good, though this was the one people said the most I should come back to and read.
EK: And the final book?
BG: I just finished reading Sapiens.
EK: I've been reading Sapiens. It's really fascinating.
BG: Yeah, I don't agree with all of it, but it's a really powerful framework, so I think I'd end up recommending that one pretty high up. My wife just gave me one, Getting Younger Every Year, which there's actually a version for men and a version for women — so I committed to her that that's the next one I'd read, which is about diet and exercise and taking care of yourself type stuff.