Bill Gates: 2013 Microsoft Research Faculty Summit

BILL GATES: Well, good morning. It’s great to be here. I haven’t been here for over five years so it’s great to see the way this event has grown. And I think it’s fair to say that we’re in a golden age of computer science. The original vision of Microsoft that we had a dream about what software could do if we had infinite computing and infinite storage; that almost is our reality today. And so it’s amazing for me to see how that’s being applied, whether it’s user interface, with vision, speech, pen, with modeling in rich data areas with machine learning. And it really seems like that idea, the powerful assistant that can help us get things done, help us drive deep insights, that the progress we’ll make in the next 5 years and 10 years will be really unbelievable. Even our friends in the hardware industry are giving us better screens, better sensors, things that let us exercise the magic of software. And for Microsoft, that means taking something like Office and going from documents to deep insights, taking something like search and going from links to answers to task solutions. So software, you know, has only achieved a very small portion of what we want it to do and even you could say what we need it to do. If you look at things like climate modeling, energy innovation, so much that this field has to contribute to solving broad science problems. I thought just as a kickoff, I’d give you four examples of things that I work on at the foundation before we go on to the Q&A, and really highlight how in each of those areas it’s software and digital approaches that are letting us be very ambitious.

The Gates Foundation really focuses in two areas. One is education, that’s primarily focused in the United States. And then the other is a series of global issues focused on the poorest in the world, including disease. That’s by far the biggest thing, over 2 billion a year that we spend there; agriculture, finance, sanitation, a variety of things that relate to the poorest.

And so let’s just start with education. Last week, I was down with Sal Khan talking about the work he is doing with his Khan Academy website. The foundation has been the biggest funder of various MOOCs [massive open online courses] to see what can be done, not just for the type of people, the more elite students who are probably the first-time users of those systems, but also looking down at the students who may not make it through college, where remedial math, remedial writing classes really discourage them and they have a hard time thinking through what are their skill sets and how do they match up to a profession that might be high paying. They have a hard time organizing their time and their finances to make sure that they’re getting through, for those things to work. And what we’re seeing is that not only at the course level itself to help model what the student knows and interact with them in a way that is encouraging, but even in that idea of the career decision; what are the credentials they need, or even monitoring to see if they don’t show up for a class, immediately engaging with them ideally in an inexpensive digital way but perhaps knowing that the counselor needs to go in there and talk to them about their career goals. The United States has the highest dropout rate in higher education of any rich country. We actually are fairly middle of the road in terms of the percentage that go to higher ed, but we drop to below average in the percentage that graduate because of our uniquely high dropout rates. And of course at the same time, we have a financial conundrum that if we’re going to give more people college education, the cost is going up and the financial resources for that aren’t generally available. So taking the courses, things like math, letting you know where you are, making it more personalized, drawing in your teacher to see what you’ve been doing, a lot of excitement there, and so, you know, we’re funding that.

Over in the health area, we work on diseases like HIV, malaria, all the childhood killers, pneumonia, diarrhea and for us disease modeling is a big area. We have a gigantic Windows cluster machine where we do stochastic modeling. We built a group of about 25 people that work with other universities on disease modeling, so things like weather, transport, mosquito populations, immune system state — a very rich model because our goal with a couple diseases, first with polio and then with malaria, is to actually go for total eradication. Now, so understanding what tools do we need, drugs, bed nets, killing mosquitoes in the case of malaria. In the case of polio, it’s looking at the genetics of the virus, seeing where it’s spreading, understand how the kids move around, what we need to do to boost their immune systems. Those tools will make all the difference in terms of developing the confidence in what we do and coming up with the right tactics out in the field.

Another example is agriculture. There it’s genetics, where we’re looking at the crops, coming up with some very important improvements. We’re putting things like vitamin A and iron into the crops. Bananas in Uganda within about three years we’ll have those micronutrients, that’s very important. We’re putting in disease resistance. A big crop in Africa is cassava. There’re two diseases that are now wiping out about 30 percent of the production. And it’s not just a numeric thing that, you know, 30 percent across the continent; it’s some farmers who this crop which is sort of their backup crop they have a complete wipe out, and that means that they really are destitute — not able to feed their family. Well, by putting in some additional genes, in this case an RNAi interference technique, we’re able to block these diseases.

And so, how do you get that out, how do you make sure people feel from a regulatory point of view that’s going to be safe and available — a very tough challenge. And we also in crops put in more drought resistance. We put in — in rice, a thing called submergence tolerance. That’s actually one of our great success stories. We now have 18 million farmers in Asia who plant a variety of rice that if it gets submerged will just stop — not die off, and then when the water goes down it resumes growing. All the rice varieties before that, when they were submerged they would just die off. That’s a good one because people can see how magical it is and so it’s had a higher rate of adaption than any of the new crop work that we’ve done.

And finally, an area to mention is in finance. Of course, consumers in the rich world have account balances and credit cards and things that let them plan ahead, and borrow money when they need to. For the very poorest, those institutions don’t exist. And now there’s been an attempt to microfinance to create some of that, at least on the loan side. It hasn’t been good on the saving side. It hasn’t been able to embody programs that are unique for things like agriculture and retail. And as we move to cellphone-based digital currency though, the transition cost can be brought down a lot. One of the problems with microfinance is that the interest rates, even if you don’t have defaults, because of the processing overhead is well over 15 percent a year. And so, it’s very difficult to find economic activity that can overcome that cost structure. When we move it into digital form, then we get it down to below 5 percent. And so it makes a big difference. But also it allows us to be creative, so that when I sell my crop after the harvest, it advises me to set aside the money for the seed and the fertilizer, and even makes that deposit. So, it gives me some way to look forward to what I need to do, to make sure I’m managing my money in the right way.

We have one country, which is Kenya, that the economy is moving to digital currency. It’s called M-PESA, something that the foundation and the U.K. Aid Department were behind. And it’s amazing; once it gets to critical mass. It’s like, you know, credit cards or debit cards but it’s — you can transfer person-to-person, a small shopkeeper can set up, and we can instrument what’s going on with that economy in a way that was impossible in a currency-based approach, and lots of innovative products are coming out that serve the very poorest.

So even progress for the 2 billion most in need in health and agriculture and all these different areas really are going to be dependent on the kind of rich, computer science software-driven advances that you work on. So that I think the opportunities are quite phenomenal, particularly as we can get commercial companies like Microsoft working with you and take your great work and get it out into products that can help millions of people. So, thanks for coming and now let me ask Rick out so we can start the Q & A section. Thank you.

ANNOUNCER: Please welcome today’s moderator, Rick Rashid, Corporate Vice President at Microsoft. [applause]

RICK RASHID: See, my real job is to bring out the water. I was told that behind the scenes. So I’m going to be moderating the question-and-answer period with Bill. Let me just talk for a couple of seconds about logistics. So, instead of microphone stands, we’re having microphone people because we were told stands were not appropriate for the staircases. So, if you have questions for Bill on these topics, you go into the aisles, you go up to the person, they look like they’re in red shirts, and they have a microphone so you can ask your questions. So, I would ask you to queue up behind those people. And in the meantime I’m going to start out the questioning while we’re getting things organized here. I’m going to ask Bill a question of my own. So, Bill, you’re sort of in a unique position because of the foundation and having a chance to really get to see some of the best researchers in the world working on a wide variety of topics. There’s an incredible talent level, different approaches people are taking, different things that they’re working on. Are there things that you see that are particularly encouraging to you, and on the other side of the coin are there things that you see that may concern you?

BILL GATES: Let’s say I’m — mostly see positive things. I think, historically we didn’t do a very good job of taking the needs, the problems and getting the — that description in front of the scientist who could help solve the problem. So outside of computer science, we had a problem with vaccines, that keeping them cold was really hard. And as soon as we showed that to people — understood about multilayer insulation, great sort of super thermos-type things, they invented a way for us to take the vaccines without using any energy — electricity, propane, any of that — keep those vaccines cold. And so, I think we’re doing a better job of — you know, whether it’s things like Kickstarter in the rich world or the equivalence in the — for developing things, getting people excited.

I do think, it’s amazing to me that certain problems don’t get attention for long periods of time. So for example, take online education, you know, why is it that in the last three years this, you know, MOOC phenomena — was there some big breakthrough that took place? Not really. You know, there was a little bit of work on adding some interactivities to the video. And once that got to a critical mass then there was enough activity for people to look at the tools of collaboration, the answer analysis. And so I think we’re moving maybe 10 times faster now to solve those problems than we were just three of four years ago.

So, it says if you can get something going that’s decent enough and highlight to the field the importance of it, then you can kind of galvanize a lot of activity. For the diseases that the foundation works on, these are very small communities. I mean, you could put all the — you know, all the malaria people in the world is like a hundred total people. And that’s nice in a way; ’cause everybody knows everybody. They don’t like each other but they do know each other, and the kind of different approaches are kind of finite in nature. But, you know, more and more people are getting educated, we’re able to share information globally in a better way.

We have a deal with China where anything — any science we see in their country that may help poor people — if we put a dollar and they’ll two dollars in. We have a similar arrangement with Brazil. So, although the best research is still in Europe and the United States, it’s becoming more of a global activity and, you know, the software embodiment of a lot of these problems is bringing in people who wouldn’t been able to contribute before.

RICK RASHID: Great. Thanks. Well, I think we have a question here in the center.

QUESTION: All right. Patrick Browder, Shell Supply Institute [phonetic]. I really like what you said about massive online courses. I think there’s tremendous potential there, and spend a lot of time with friends debating this topic. At the same time it’s very interesting to see, also I think there’s going to be a very interesting side effects also on the — on this country and countries around it — in the sense that teaching itself may change quite dramatically. And I was curious to hear more about the kind of the bigger picture for like additional thoughts you have on this.

BILL GATES: Well, yeah, the question is will — there’s — teaching has many different aspects. You know, one aspect is the big lecture, and there’s definitely a view that that will become more — less like individual performances of live music and more like recorded music, where there’s a set of people who do it extremely well, get very big budgets to, you know, edit it, do the great experiments, and those things. You know, so like going back to Walter Lewin’s physics course, you know, which was the MIT open courseware initiative, you know, about a hundred times more people watched him than watched place-based activity. And so, taking that and raising the quality of it and making it available for free, that’s clearly going to happen.

But of course if we think of that as the only part that goes on in learning, you know, we’d be missing, you know, where kids are doing problems, when they’re in labs, where they’re in study groups, where they’re discussing topics, you know, many things don’t fit that big lecture format. So how we take the digital technology and take those other aspects is fascinating to me.

One thing that we — our foundation tracks a lot is this dropout phenomena. And what’s incredible is that the so-called for-profit sector, although they’ve been properly admonished for their over-marketing and over-promising, you know, in some cases suggesting people to get a high-paying career that they’re not likely to get. They’re — because they are sort of profit driven, the way they track students and see exactly what’s going on, you know, they call you within a few minutes if you don’t show up for something. They’re — that kind of handholding, some of which is embodied digitally, is a best practice. And so public universities haven’t had that idea of, “OK, what is the state of the student, what’s all our evidence about how we need to intervene or help things of that nature.”

So digital is in many areas. For elite students, you know, people who would have bought, you know, Learning Company, Great Course-type lectures, you know, this is a golden era, the ability to learn a lot of topics, you know, I find it just absolutely phenomenal.

And we’re seeing as we take knowledge like public health knowledge, get that up online, we’re able to raise, say, in Sub-Saharan Africa where the access to these through this material has been very poor, are even proving your confidence against this material, there’s been no ability to do it. We’re, it’s quite revolutionary in terms of driving confidence into health departments that were just isolated, not able to know what best practices were. So, you know, we’re on the beginning of something really quite profound even though the temptation to oversimplify it is quite great.

RICK RASHID: I think we have question here on the left.

QUESTION: Hi. I’m Ganesh from IT Bombay. You mentioned that there’s substantial work on online courses for the U.S. Now, what about countries like India, just to bring to your attention. In India we have more engineering colleges than the U.S. and China put together and a lot of students in colleges quite disillusioned about what they would like to do. And there’s also a kind of narrowing down of the funnel as you go to the cities, so students are left with very less opportunities. So — so my question is, what about online courses in countries like India and also importantly beyond colleges that are also data centers in villages spread across a country like India where a lot of data gathering happens. So there’s also case for capacity building at these data centers so that data analysis — I mean, using machine learning, data mining, can be performed in a decentralized manner. So what are your thoughts on this?

BILL GATES: Well, at the end of the day, a lot of education is about getting the credential that helps with your employability. And, you know, the gold standard for that today is to go to a university where you spend time physically and the idea of how you learned and what you learned and, you know, to prove that you have this ability, those are tied together. I see those things decoupling, that the way you prove that you have certain skills, you know, can be very straightforward and there can be a lot of people who are competing to help you gain that knowledge. Certainly this is going to be a global phenomenon.

The challenge of educating all the people who should be educated is much worse once you got outside the U.S., although the economics are the most grueling. There are other countries participating in these things, but having, you know, scanned the landscape, the U.S. is the place where the most interesting stuff that’s happening right now. Now that stuff is all localizable, you know. I was down at Khan we’re looking at the Spanish version, the Portuguese version, the Turkish version, the Danish version, and it’s not that hard when you do these things to make these incredibly global in nature.

Now making them appealing, making them rich, making them deep; that’s what hard. The IT lectures are available. They’ve been videotaped. But if you want to see some lectures that you might not make it to the end, you know, if it’s optional for you to watch, there’s quite a few examples in there. Just taking the camera in front of somebody who’s got a captive audience, you know, doesn’t measure up to what’s really going to be necessary in terms of being one of these lectures that contributes in a very broad basis.

So, you know, given the cultural focus on education in the Asian countries, the adoption rates and some of their own innovations are going to be really huge. And if you say it’s not just for four-year college, it’s also for vocational-type activities, it’s for the high school level. The high school and the university stuff have been a bit decoupled. The K through 12 has been mostly stuff like Khan and really kind of trying to engage the student with game-type technology, whereas the university stuff is more of the MOOC thing. But that boundary is a pretty soft — soft boundary as well.

RICK RASHID: I’m going to take a question from the right side here.

QUESTION: Thank you. Mr. Gates, my name is David Lovell. I’m from CSIRO, an Australian bioinformatics network. I’ve got a lot of things to say thank you for, I — your work and I want to say thanks in particular for the work at the foundation. I understand your daughter is about to go to university, facing some career and educational choices. And one of the things that I’ve come to understand is that when people cast themselves in narrow terms — I’m a computer scientist or I’m an electrical engineer or a bioscientist or something like that — that doesn’t quite seem to address what I think the world needs, which is people who have talents across the spectrum, have disciplines and the ability to integrate knowledge from these disciplines and actually create meaningful solutions that are more than just, you know, a hammer looking for a nail. So I probably should have thought better about the actual question I’m going to ask you. But it speaks to this sort of multifaceted nature of education and collaboration. Do you have some thoughts about whether society is well served by seeing education cast in these roles as specific disciplines? What are your thoughts on that?

BILL GATES: Well, I think, you know, fortunately society, a very small percentage of society can do the innovations that benefit society very broadly. And so the — you know, the education that we provide to make sure innovation goes at full speed and the education we provide to make sure people are kind of employable in a way that’s satisfactory for them both the work and the economics of it — those are almost two separate questions. That people are going to do the deep innovations. Yes, a lot of them will be polymathic. They will be, you know, unusually strange people who learn a lot of things and therefore are able to reason across the boundaries of two subject areas.

I mean, just look, you know, in biology how much physicists came in there. In the global health work I do, you know, a lot of the top people have both an M.P.H. and an M.D. And just think about the number of years they had to go to school, you know, and they have to have been out in the field for three or four years to see things that do and don’t work. You know, when you meet these people, you kind of go, “Wow, how did you gain all that knowledge?”

I do think this is a wonderful period of time for those things. I mean, if you want to learn — you know, I want to learn about the ocean or meteorology or game theory or, you know, just not theory, you know, there are great courses, you know. So boom, 24 hours later, I know a little bit about those topics, I can test my knowledge. So let soci — if people have that type of interest and inclination, sure, we should nuture it, but we don’t want to push everybody down that path. But that’s where a lot of the deep innovation is going to come.

You know, it’s interesting as my daughter goes to these universities they’re all saying, “Oh, you can design your major and we don’t force you to do anything and stuff like that.” It’s an interesting tradeoff, you know, how open-ended these things should be. I’m a big believer that — gotta have at least one or two subject areas that you actually know something about [laughter] in a pretty deep way. Like everybody should know machine language, of course. [ Laughter ]

RICK RASHID: Question from the center, please.

QUESTION: My name is Bertrand Meyer from ETH in Zurich. My question is about IP. It’s clear that — so some of the issues that developing countries face are IP-related, both intellectual property related both for, say, drugs but also for software. And in fact a number of countries have enacted open source-only mandates. So I assume you must run into some contradictions, or at least some tensions, between your background at Microsoft with propriety software and your charitable — current — your current charitable work in this respect. And my question is how do you see the reconsiliate — the combination between open source, free software on the hand, and proprietary software on the other hand, and marginally are there specific IP rules that should apply to emerging countries or should they just follow exactly the same rules and use the same tools and the same conditions that we all use?

BILL GATES: Well, the — you know, thank God for commercial software. You know, it actually funds salaries, gives people jobs, you know, terrible stuff like that. And, you know, thank God for free software. Free software, you know, lets people get things out there, you can play around, build on. And so, you know, the two work very well in a nice common ecosystem.

And no, I don’t see a big difference country to country. You know, it’s very similar to what we do in — you know, the foundation at this point, we’ve saved about 10 million lives that otherwise wouldn’t have been saved. And our goal for the next decade is 50 million. But we never would have been able to do that except for our partnership with the pharmaceutical companies. You know, so thank God for patent laws that allow them to invent drugs that they get to sell that then they get to hire researchers. They are phenomenal at understanding the drug libraries, the assays, the things like that.

And in fact, none of their patents exist in any of these developing countries. So we never, you know, contrary to what you might think in the press, we never run into IP problems, not a single time on a single thing. Because in the poor countries we work in, the poorest 90 countries, nobody has got — nobody files patents, nobody enforces patents. You know, it’s great. It’s essentially a transfer of people buying drugs in the rich world who are now enabling these things to be done at marginal cost. So all the vaccines we do, we understand the marginal cost. We make sure that those — you know, that’s exactly what the pricing is to the lowest tier.

In the same way that, say, a commercial software provider, we give the software away basically free to these educational projects, you know, student-based activities. So you can do the tiering that actually, when you’re in an intellectual property where you want to fund the R&D, that works pretty well.

So, you know, I think intellectual property, a complex topic, but, you know, the value of the patent system, of allowing those pharmaceutical companies to exist, allowing the agricultural innovators, the people who really get drought, really get salt tolerance, you know, those guys doing those proprietary seeds have given us access to that knowledge to use for the poorest. And that’s why we’re so optimistic about dealing with new climate conditions and helping these poor farmers.

QUESTION: So no contradictions and no tensions?

BILL GATES: Well, the world is — it’s a complex system. Anybody who thinks getting rid of it would make the world better off, I can certainly tell you that’s crazy. Anybody who thinks we run into problems helping the poorest because of these things, you know, I’d like to know what they’re talking about. The mix of commercial with free without any coercion, you know, a country forcing it to be one model or the other, in my view, it’s — is it’s working very, very well.

RICK RASHID: Question over in the right.

QUESTION: Emma Brunskill from Carnegie Mellon University. I was wondering if you think there’s a role for a tighter coupling between some of the education sort of initiatives you’ve been doing as well as the financial services. So try and better to couple sort of the educational offerings with economic opportunity, particularly in low-resource areas.

BILL GATES: Well, the — you know, education in a broad sense, you know, there’s two key factors. One is the personnel system, that is, are the people who go into the education system well trained? Are they measured in terms of how good they are? Are they given an understanding of who — what the best practices are so the average quality rises up to that of the exemplars? And, in fact, there’s a lot of technology enablement that can be done to help that personnel system work well.

Most educational personnel systems are run very, very poorly. You know, in South Africa and in India, teachers don’t show up about 25 percent of the time. In Mexico, the teachers union wouldn’t tell the government who is employed. Now, just six months ago, they were taken out of that position. So the political willingness, particularly for K through 12 in developing countries to run a good personnel system, just getting the basics right can make a huge difference. And some Asian countries have done that very well, better than the U.S. And most poor countries have done it even worse. And we’re just talking about very basic things there.

Bringing in technology will make a difference. There’s a very exciting company called Bridge International, which is a private school in Kenya that uses a tablet-driven approach so that their teachers are all going through a very specific set of activities and it’s very well-monitored. And it’s amazing to see how quickly they’re able to branch and the comparison of them to the public system and how that can be done.

So fixing, you know, when you go into a poor country, you want to fix health, you want to fix agriculture, you want to education, you want to fix governance. And, you know, it’s the magic blend of those things, all of which re-enforce each other. If we have ill health, we tend to have all those other things worse — worse. It’s those things that get you on the path to what — you know, the good news is most of the world now lives in middle income countries, countries that to some degree, like a Brazil, Mexico, China, have gotten these things right.

There’s about a third of the world lives in countries where these things haven’t come together, and it’s clear that innovation, particularly technical innovation, new vaccines, new seeds, monitoring things to make sure government workers do what they’re supposed to, including the education, that we can make much faster progress to get these people out of these poverty traps now than ever before.

RICK RASHID: So Bill, we’ve got a question from the streaming audience. So what role do you think wearable technology is going to have in education?

[ Pause ]

BILL GATES: It will help you cheat, I guess. [ Laughter ] Wearable technology, you know, the word wearable, I mean, you know, glasses that project onto your eyes, things that can monitor whether you’re getting nervous during the tests, you know, I think of wearable technology as a very cool thing. But I don’t really couple it that much to education. You know, I think the way that you’ll be able to review what you’ve learned, have the system understand the state of your knowledge, you know, that’s a very good, good thing, but I am not sure how it connect that to wearable.

RICK RASHID: Maybe you could check to see how your pupils dilate when they ask you the questions, you know —

BILL GATES: Whether you’re telling the truth.

RICK RASHID: — see when you really know the answer at all. I always used that when I was a professor. Question on the left.

QUESTION: I’m Gordon Wyeth from Queensland University of Technology in Australia. Mr. Gates, you wrote, I think it was six or seven years ago in Scientific American, about the personal robot revolution. I just wanted to get your feeling on how’s the revolution going.

BILL GATES: Well, the word robot can be interpreted very broadly, you know, anything that we can use, that we can automate. I mean, you know, the history of farming is the history of automation. And, you know, so you can think of all those great farm things as, you know, very specialized robots.

I saw one recently where they actually use a set of cameras. And so as they go through the field, they actually see what’s a weed, what’s a plant, and they imply herbicide to get rid of those weeds. Now the cost of that today, of actually having the thing that you drag around the field that does this recognition and is smart about doing that targeting, that’s rich world farming today. For poor world farming, it’s much more about taking satellite data and looking at things in a broad area; soil health, when you should plant the seeds, things of that nature.

You know, I do think robots will raise some very interesting issues as they replace certain types of labor. And, you know, we’ve always wanted to have a robot that can go out in the rural areas and help out in certain health care-type things. That’s a very high bar in terms of what type of robotic capability you have to have, say, to help do a C-section in a rural area where, you know, that absolutely needs to be done. You know, so I don’t think that’s in the next 10 years, but maybe in the next 20 or 30, that kind of physical expertise can be made available very, very broadly.

Robotics continues to, you know, in terms of the numbers in deployment look pretty small, but, you know, we’re always on the verge of when is the breakthrough going to happen. And certainly, the work you are all doing on vision, planning, modeling, those things, you know, it feels like, “Boy, are we awfully close on that.” I think it’s going to be in developed countries in changing the job markets in developed countries where we’re first going to face the generally pluses but also the disruption that that kind of innovation will cause.

RICK RASHID: I’ll take a question from the center.

QUESTION: Hi. I’m Henry Fuchs from the University of North Carolina. I want to ask about how come — let’s see, how do I start this? I, and I think most of us around here, as well as most people in the country and the world were — are so impressed at how much you have done, personally with the foundation that you as a person, who was a revolutionary leader in hi-tech, devoted much of your life to helping the poor. And I for one thought that this would not just change the world, you know, in the direct way with your foundation, but that would be an inspiration for other billionaires to do similar things and Washington and people to get basically on this sort of bandwagon. But maybe it’s because I haven’t been following closely enough but I don’t see it. I don’t see Washington getting on the bandwagon of helping the poorest two billion people. I don’t see other billionaires changing their lives to help the poorest two billion people. Am I missing it? It’s not in the news or…? When you go to these places, when you go to Washington, when you see other billionaires — [ Laughter] They’re still doing other things like building biggest yachts in the world.

BILL GATES: But there’s —

QUESTION: Am I missing something here?

BILL GATES: Yeah, I think to a small degree you are. And it’s a work in progress. You know, I was certainly inspired by the first generation of big philanthropists, Rockefeller, Ford, Carnegie, they did an amazing job, and you know, they funded research, they funded schools for blacks, they went around the world and did some good things. It was, you know, really surprising how smart they were about the topics they picked.

If somebody has a lot of success in the world, there’s two ways you can give back. One is to give your money, and so for example, my friend Warren Buffett, who is brilliant, high integrity, you know, the person I’ve learned the most from, a very close friend, he’s choosing to stay doing his job. But half of the four billion a year that our foundation gives away actually come from him. So he’s delegated to us, you know, that giving piece.

Then, if you’re successful, you can take your time. You know, if you’re good, if people listen to you, if you have colleagues who are good or if you’re good at building institutions. And the Gates Foundation in an area like global disease is an institution that’s hiring scientists, researchers, deciding how to give the grants and it’s taken us 10 years to get that institution to a level, you know, sort of the level of excellence that say Microsoft had in 1995 where you really feel people are very analytical, on top of things, and you know, that’s hard work. It’s fun work. It’s something that my Microsoft experience taught me how to do, although I had to learn some biology and different things along the way.

We’ve created the thing, myself and Warren and my wife, Melinda, a thing called The Giving Pledge, where now a hundred billionaires in the United States have all committed during their life and through their will to give away the majority of their wealth, and we get together. It’s been going for three years now. We get together on a regular basis. We get more people joining.

We’re doing the same internationally. Whenever I’m in China, I go and sit down with people who are wealthy and. The nice thing about China and to some degree India, it’s mostly first-generation wealth. And first-generation wealth is where you get the greatest generosity because you don’t think of yourself as a dynasty or a bunch of aristocrats. You actually think, “Hey, this is strange. Why does somebody have this much of society’s resources” and, you know, it’s probably distortionary to give it to my children. Therefore, what can I do? How can I give it back to help those most in need?

So I do think we’re in a second golden age of philanthropy. A lot of that philanthropy expressed itself in very diverse ways, not necessarily for global health.

One other point on one thing you said, the government foreign aid budgets which are, if you take the Europeans that are 60 percent, the U.S .that’s about 20 percent, foreign aid budgets are about a 130 billion a year. And so in terms of the bulk of money that helps the poor. Now some of it it’s not well spent but even, you know, factor — have them by factor of two, is government aid money. And so our four billion a year, although it’s very big in the upstream, malaria vaccine, AIDS vaccine, diarrheal vaccines, when it comes to downstream delivery, we have to part with these governments. And their tight budgets are making, you know, making us, so we have to go off and really make the case for this money that goes to other countries.

To give you an example, we just raised five and a half billion for polio eradication to get that done in the next five years. Our foundation gave 1.8 billion. Other philanthropists including Michael Bloomberg, Carlos Slim; those two gave a hundred million each. There was 600 million from philanthropists. And the balance about 3.1 billion came from governments. So the governments are very important. We’re trying to make sure they don’t back off particularly the Europeans are by far the most generous in these areas.

RICK RASHID: I’m going to take another question from the streaming audience. So the question is, how is machine learning helping computers understand how diseases and climates work?

BILL GATES: Well, in terms of disease, the state-of-the-art disease modeling over the last decade has improved very, very dramatically. I mentioned I have a group that I’m funding that uses a stochastic modeling approach and does a lot of work. But, you know, Imperial College, Harvard, Swiss Tropical, there’s about 20 places around the world that do serious disease modeling and being able to have more sensors, to have the genetics so you can look at the course of the infection. Disease modeling has gotten very, very good.

There are a lot of deep mysteries like you know, who spreads HIV, why is the AIDS structure like it is? Why does TB — how does that spread? How does it go from latent to active? You’d be amazed. These models are very necessary to help us gather data, see the patterns and then go back and refine the model.

In terms of climate, they’re — they have the 20 different climate models. They are, you know, fairly rough. To some degree they are curve fitting type things where they, you know, try to match the past to some degree. And, you know, they disagree on a lot of things, so that’s a helpful thing. And when you want to look at certain interventions they run them against those models, things like, you know, geo engineering or different energy system, changes that take place. It’s still not nearly sophisticated enough compared to the phenomenon they’re trying to model. If you look and say how the ocean is modeled in these things, it’s not very strong. If you look at how agricultural sector or just even forests are modeled, it’s not very rich. It’s a very tough problem and they’re getting better all the time. But even, you know, the basic parameter if you raise CO2 by X to, you know, what’s the multiplier in terms of the temperature effect, you have huge variation. So, they haven’t been able to resolve the — some of the basic uncertainties.

RICK RASHID: How about a question from the left over there.

QUESTION: David Anderson here. Fine, OK. Dave Anderson, Carnegie Mellon. So, 30 years ago, if the quote was correct, your motto for Microsoft was a computer in every home and running Microsoft software. And it seems like you won. But when I watch my friends who are not the kind of people you’d see in this room, and I watch them use their computers and I watch when I use my computer, I think one of these is very not like the other. And admitting that we’ve all got big geek hats, I think there’s still this enormous gap between the way that a trained computer scientist or uber geek uses computing, kind of pervasively throughout their life to automate and simplify and avoid all the boring stuff that we all hate doing. But our history of computing is littered with the carcasses of attempts at automators and visual programming languages and things like that. So, what is the next step that we ought to take in order to make it so that more ordinary mortals without computer science degrees can benefit even more from the computers that you’ve managed to get into their houses?

BILL GATES: Yeah, that slogan “a computer on every desk in every home,” that comes from 1975. And it did have, at the end, internally, we said “running Microsoft software.” But outside, we always said a computer on every desk in every home. And I think we actually got rid of that as the mantra like, you know, 10 or 12 years ago because, you know, at least in the rich world, we were on the path to have that take place and the idea of, you know, tools of empowerment, tools of understanding, tools of creativity and communication. I think we took something on. It was a bit more broader.

You know, after all now, it’s not just one computer. It’s a computer in your pocket. It’s the deep sensors. It’s the way these systems are going to work together with profoundly different UI. You know, it’s certainly true that the way people use the Web, the way people use Office, you know, they’re way short, there’s all sorts of latent capability in there that if you had time to explain it to them, they could get more out of it. And it’s surprising how little we’ve had a breakthrough in terms of how you fully take advantage of software.

Even if you look at what people type into search engines and, you know, that’s one of the most studied thing in terms of you know, that’s a real time digital stream that we’re [inaudible] again and again helping them do that in a way that would be most effective. You know, we’re not — we’re not all that great in it. So, we’ll always have this, you know, frontier of what you can do in software and what we’ve made easy for lots and lots of people to do.

You know, software is going to be more ambitious as we’re actually representing things in deep ways. You know, as everyone gets essentially what we’ll call the personal agent, you know, it’s been talked about for decades and now really is possible, where we see where you’re going. We see your calendar. We see your various communications. Some of those communications, we can actually, you know, look at the text, look at the speech, try to be helpful to you in your activities.

I think we will be more connected and therefore, you know, if somebody wants to do a task like, you know, find a gift of a certain type, organize a trip of a certain way, that there will be a closer match; that is, the gap between what the software could do for them and what most people end up doing, that could be reduced.

But, you know, they always make mistakes on these things. When the machine, you know, tries to do the table with numbers or the dog, you know, comes up and says, “You didn’t do this right,” Microsoft Bob-style, which is a long time ago we tried a little personality that was definitely premature, I think it will reemerge [laughter] but perhaps the deep bit more sophistication, you know, we were just ahead of our time like most of our mistakes. [Laughter] So, you know, it’s a research issue to make software better and, you know, the price when you do these things is very large in a commercial sense.

RICK RASHID: Maybe we just need a Border Collie. They’re a lot smarter. I’ll take a question from the right.

QUESTION: My name is Alex Wolf, Imperial College London. So, first, thanks for giving a shout out to Imperial College, I appreciate it. My question has to do with unintended consequences. I think that, you know, if you talk about killing billions of mosquitoes, it can have a wonderful impact on malaria but it has, despite all of the power of Azure, incalculable consequences perhaps for the food chain. If you build an electronic-based economy in Kenya that could be monitored? Again, wonderful ability to manage an economy but you can easily imagine that there’s unintended consequence of the centralization of that information and control. And even with MOOCs, again, wonderful opportunities but also the potential for homogenizing education and information and controlling what education — what shape education takes. So, the question is how do you think about those unintended and incalculable consequences?

BILL GATES: Well, I’m always glad there’s people around to dampen my enthusiasm. [ Laughter ] ‘Cause I have a tendency to overly see the good side of these things. I mean, you know, is it — you know, we used to have all these individual performers who would sing before we had pre-recorded music and it’s very sad now that you just listen to this performance that’s a hundred times better.

I think MOOCs are going to be like that and I feel sad about that. In terms of unintended consequences it’s very hard to match what is potentially scary in terms of bioterrorism. The fact that we understand and a small group of scientists could create something that can literally kill billions and we actually need to allow there to be surveillance systems that would be out there and try in advance in a trusted way using the information for the right things, to interdict that type of unintended consequence. You know, I think, you know, it’s hard to beat that as something that you hope technology doesn’t take us down that path.

Actually, Nathan Myhvold who started Microsoft Research with Rick, recently wrote just a few weeks ago, came out with a memo talking about this area. So, yes, digital currencies are attributable and, you know, there’s a ton of laws about cash transfers and know your customer, things like that that actually are quite painful for these systems to implement. There are going to be political decision policies that go around these things.

In the education space I frankly don’t see that much of a downside. The general issue of online, should you be able to tell when, in which context should you be able to reliably tell who something is, what sort of trust do you want to have in those situations? That’s going to be debated endlessly. And, you know, some people have a model that maybe the government should be able see people planning bioterrorism attacks and that they won’t use it in maligned ways and some people have the view of hey, you know, ultimate privacy even for terrorist is absolutely a good thing. So, you know, we’ll have debates about how these digital systems get used.

If you go out in the developing world and you see kids dying of malaria then, you know, when you see the malaria vaccine come along, when you see a woman who can send their kids to school because she has better crops, it’s hard to feel too bad about the general arc. You know, we don’t want to be like 300 years ago, where over a third of kids died before the age of five. We really don’t. We have a few places left that because we have terrible health, we have terrible population growth, Yemen, Pakistan and, you know, we really understand how innovations can come in and help there. So, you know, because I see that, I mostly see the positive side, but we should be, you know, aware of these negatives and see how we can mitigate them.

RICK RASHID: Let’s take another question from the streaming audience. So, Bill, what role do you see in social media in terms of improving society?

BILL GATES: Well, communication is a good thing, and social media is a form of communication where you have your set of friends, you know. Eventually, the understanding of what kind of relationship you have with people, which things you want to share with them, like they’re trying to get in touch with you and you’re busy, do they get to see your calendar what you’re up to, do they get to see when you have free times in the future and automatically set things up based on that? You know, do they get to — if you really love an article, is it very easy for you then to benefit and see that as well?

Social media is, you know, people can make fun of it, that because a lot of it is used for, you know, what like my cat had for breakfast type things. But, you know, that — so I think of it in a more broader context, which is the enrichment of communications to people that have common interest. In there I see it will be a strong foundation.

If you take, say, on Khan Academy, that they generalized, it’s not just a student-teacher relationship, it’s a coach-student relationship. So, parents can come in and be your coach. If you’re willing to let them, they can see what you’ve done well, what you’ve not done well. If you have a relative who’s good at this subject, they can come on and see and give you hints and things like that. And so, the social nature of education, where you’re kind of competing with each other, where you’re learning from peers, this is one of these great breakthroughs where it doesn’t cost, you know, great sums of money to set up systems where peers are competing and learning from each other. And so, I think social media will be profound in terms of people getting advice about their life and learning different things. It’s not at that level yet.

RICK RASHID: Question from the center here.

QUESTION: Hi. I’m Rafael Calvo from the University of Sydney. So, when we talk about education and educational technologies, most of your forecast being on improving employability for the students. There has been very little work on improving skills like resilience, emotional intelligence, empathy, skills that are very, very important for happiness, for psychological well-being. So, I wonder what are your ideas about these from the point of view of what the foundation can do and what researchers like us can do on this area.

BILL GATES: Yeah, that’s exactly the topic we had a full day gathering a year ago and a full day gathering earlier this year of the foundation education group because there are — to some degree, going to a four-year university and getting a degree isn’t as much about what you learn in that experience, as it is — it’s a badge that you’re a certain type of person who’s organized and, you know, you like to achieve things and you like to be measured in terms of what you’re doing and so employers look at that and say, “OK, that person is very employable.”

So, the question is, well, why do some people have those skills and some people not? You know, yes, there seems to be some income correlation, the quality of your education on measurable things like your math skills — it tends to correlate to those things. But the question is always, how come we can’t get to these capabilities more directly?

You know, maybe they’re more primary in a certain sense so we should invest in that and then teach you math. And, you know, there’s — there are people like Carol Dweck who has the learning mindset which is, I think is a super important way of looking at things, that you teach kids to feel good about their struggle and how they made their effort, as opposed to what they achieve, because if you praise them for what they achieve, then they think, “Oh, I better not try too much because I’m going to, you know, start to look bad.” And if they got lucky at first in an area then, you know, they won’t go back and really cement in those concepts in exactly the right way.

But actually, putting and embodying this in software or finding the teacher who does this particularly well other than just a general sort of mentoring encouragement-type support, there really hasn’t been that much in terms of a breakthrough. And you even have people who look at, you know, there’s like this thing called the marshmallow test where you look at, are people able to delay their satisfaction? And you can see some of these non-measurable skill-based things general approach really are very predictive of how people — they’re not totally overwhelming, that doesn’t, you know, take away from the rest of the staff are doing.

You know, cognitive science, learning science, it’s surprising how much — how little really good researchers there is on it. Just like the research on, Why Some Teachers Were Getting Such Unbelievable Results was very weak. Now, the foundation put a lot into that videotape. The great teachers looked in, saw how they interacted, how they made the subject growing. So, we know a little bit more about that but this thing of teaching approach mindset, it’s pretty vague and therefore soft but it is important if we can, you know, make — have some deeper understandings there.

RICK RASHID: Question over in the left.

QUESTION: Hi. Sid Suri from Microsoft Research New York. What is the most important problem facing humanity right now and then as a follow up question, what is the most important problem facing humanity that people in this room might be well equipped to handle?

BILL GATES: Well, the most important problem that we want to avoid is, you know, biological and nuclear terrorism or war. But nation-state war right now, the risk of that on any large scale, fortunately, is lower than it’s ever been. So, you want to avoid war as much as possible, bioterrorism, use of nuclear weapons. So, you know, some people should focus on that.

In terms of an ongoing disaster, the thing that people just don’t have this awareness of, they look at, oh my God, you know, some people are dying in Syria and that’s very, very tragic, but seven million children die every year. Now, you know, when the foundation got started back in 2000, that was 12 million a year and so now it’s down to seven million. The majority of that reduction is new vaccines, higher distribution rate, new vaccines for a diarrheal respiratory-type diseases, getting those out to people.

Certain interventions that have helped with malaria brought that down from 1.2 million to 0.7 million, 10 percent of the 7 million a year who died. Over the next 15 years, we will be able — the world working together if we do the right thing. We should be able to get that number below 3 million. Now, if you go back to the year I was born, it was 20 million and so, you know, that’s serious progress, 20 million a year to 3 million. Actually, the birth cord is twice as big so in terms of percentage reduction now, it will be the birth cord peaked about 8 years ago at 136 million kids born in a year and that’s coming down now. So, all population growth is just filling in the different age ranges.

I tend to think of either the 7 million kids a year who die or the fact that in poor countries, about 40 percent of the kids because of bad nutrition or disease insults, their brain never develops, and so they’re not able to become literate and contribute and give back. Average IQs in Subterranean Africa is about 82, and that’s nothing to do with genetics or race or anything like that. That’s disease, that’s what disease does to you and that’s why these things are such an extreme poverty trap. If you have kids with ill health, their brains don’t develop.

Also parents, because they want to have a very high chance of two kids growing up to support them in adulthood, they have — they just have more children, and so you get high population growth where there is high childhood death rate, which seems as the opposite of what you might think.

So, those to me are big problems. Somebody else might say that employability of youth, you know, how we educate kids so that they can have meaningful jobs and not just feel like, you know, that things aren’t very worthwhile. Some people would focus on old people, some people would focus on healthcare cost, you know, everybody gets to pick their political dysfunction. A lot of things to choose from but, you know, childhood death gets pretty high up for me.

RICK RASHID: One last question then we’ll be done, here in the center.

QUESTION: Well, thank you very much. Kevin Sullivan, University of Virginia. Bill, there’s a whole class of systems that it seems we’re still unable to build today. Some people call these ultra large-scale systems; examples would be a national-scale health information system for bio-surveillance or, you know, in electric power, a generation distribution transmission system that, you know, optimizes across the nation. These aren’t just software computing problems, although software and computing are the fundamental enablers, really they’re very large-scale systems engineering systems design problems. Do you have any thoughts on how our field goes forward perhaps with others to get us to the point where we have, you know, the science, the methods, the tools, the concepts that are needed to enable us to build the ultra large-scale systems of the future which seem like they’re within reach by just keep kind of receding beyond our immediate capabilities?

BILL GATES: Well, there are areas where the progress, you know, take programming, is programming today much different than, you know, 10 years ago or 20 years ago? Not really. The level that we express things at, the ability to take domain knowledge and encode that in a way that’s kind of transparent, understandable, easy to update. You know, I know there’s work along those lines but you could have said that 20 years ago, 10 years ago and, you know, just take entry-level programming, is it, you know, dramatically better now than it was back then?

So, I think we still have a lot to do on knowledge representation, logic representation, despite that the difficulty of doing things like a medical record system is surprising, you know, that, you know, you’d think just — they — the data representations we have are pretty good at doing those things and yet doing them at scale, testing them, the fragility, people have had still a very tough time with those.

And I do think the move to the cloud where you can share more patterns and that the richness of the primitives particularly for storage would be a lot better. I do think that provides some promise.

Some of the things you mentioned like the grid, that’s more of a physical, you know, right of way-type problem. I mean we are just — we don’t even have 1 percent of what you would need as a grid if you’re going to move power on a national basis. And so, it’s not — it’s not even in the realm of software at this point but, you know, it is very funny what’s gone really well in writing large complex software systems; which things have not gone that well. And, you know, it’s not like we prove these things correct or we specify them at some very high level. So, I hope, you know, some of you can surprise us with faster events in those areas.

RICK RASHID: Although we have — you know, there has been quite a bit of progress, I mean I can remember when I was first learning here the old saw was you could expect 600 good lines of code from someone in a year. And we get a couple more than that now, so… [laughs]

BILL GATES: Yeah, you really — by what objective measure? I mean I remember when, you know, we first got symbolic debugging or tracing or things like that it was like, wow, now we really know what we’re doing. I don’t know. What objective measure would you use to say that software development is better today than 10 years ago?

RICK RASHID: I think we build much larger, more complex systems than we thought we were capable of doing 30 years ago.

BILL GATES: Well, we write a lot more code.

RICK RASHID: And we get more people working on it. [Laughter].

BILL GATES: That is true.

RICK RASHID: Yeah. All right, well, I want to thank Bill for the — giving us this opportunity to speak to you and answer your questions.

[ Applause ]

End