Fireside chat - Vinod Khosla, Founder of Khosla Ventures, and Bill Gates of Microsoft

Bill, what do you think of this since you're chomping on it?

It's quite good. I've had more chocolate since I was introduced to this brand than I've had for many years. So the taste is fantastic.

In fact, some of you got to see pieces of this, but Bill saw both the candy but also salt, beef, eggs, cheese, and sugars, I guess. So maybe you can just give us your thoughts on some of this food stuff.

Well, food is a pretty basic thing, and you know, it used to be the economy was all about food, employment was all about food, and in fact, in countries where you want to deal with the basics, getting people to have enough food is an incredible thing. And so all these companies that are taking the animal products, the milk, the eggs, the chicken, the beef, and actually coming up with a way of using largely plant-based materials—soy, peas, a variety of things—to make these things that are both cheaper, probably more healthy, You know, less cruelty involved, less greenhouse gas emission. It's quite a phenomenal thing. I mean, people can be enthused for any of about 5 or 6 different reasons. You know, I think it's a huge thing and it's one of those things that will confound the pessimists because when they look out, they see this increased intensity of animal product consumption as part of increased income, as it should be, that's what people want. But the fact that innovation will give them the equivalent without those negative effects at lower prices, I think it's an amazing example of how just sort of linear projection misses what innovators look— using science will be able to do. It's completely not part of the mainstream dialogue. 5 years from now, as these products get out there, I think the whole view of what agriculture needs to do, what the trajectory is there, will be a lot more positive because they'll see what the innovation will cause.

You know, you make a very interesting point. One of our speakers tomorrow is Dan Gardner, who wrote this book Future Babble and is writing another book with Professor Tetlock, who I met with a few years ago when I was studying this whole issue of forecasting. Food is one interesting category, but I'd love your comments on all the other things you see going on in this group of entrepreneurs. I'm biased, of course. I'm totally biased. But whether it's engines or batteries or lighting or data, or you could look— or creating new publishers. It just is. I'd love your honest assessment of all the things you've seen in this sort of portfolio of companies.

Well, my framework is a very positive framework. That is, the way that things have developed over the last several hundred years because of innovation based on new science. Most people don't realize how bad life was in terms of short lifespans, health, lack of literacy, nutrition. And, you know, it's only in the last 100 years that we've doubled the average human lifespan. That's an incredible thing. And so when I think going forward, which— there's two narratives about what's going on from now. There's the very negative narrative about limits to growth, Club of Rome, Age of Austerity. Lots of people thinking, my God, you can't continue on in some straight trajectory. And then there's a more sort of positive view, often you know, overly simplistic, perhaps, like Age of Abundance, Rational Optimist. And the biggest reason why I'm more in the latter camp, the, the positive camp, is because innovation, science-driven innovation, can solve so many problems, can improve so many things. And You know, today I saw 15 amazing examples of that, whether it's in batteries, cheap energy, IT solutions where you're taking tablet, mobile phone, and doing new things. And actually, the percentage of the world's resources, or even the US resources, any way you want to do it, that's tied up in this innovation is very small. And yet the impact on the quality of life 10 years from now, 20 years from now is going to be amazing. Cheap energy. Clearly, this work will lead to cheap energy, even a realistic use of intermittent sources because the various grid and battery problems that are often ignored. There's science here that addresses a lot of what needs to be done there. And, you know, There's nothing more fun than listening to people who learned the science, put their life, hired great people, can tell their story. This is the most interesting set of activities in the world. And so I come away very positive about what will happen. Each of these companies They may run into problems or somebody may even do the thing better, but— and in a few cases that will happen. But all— you know, that's all to the good. What it means is that this linear extrapolation, you know, is wrong— more mistaken today than it's ever been because the rate of innovation is higher now. And the U.S. share of that is still extremely high. That won't stay the same. And even characterizing something as U.S. versus non-U.S. is sort of a silly thing that— it's sort of a nationalistic impulse that people like to do. But, you know, on a global basis, innovation is accelerating.

You know, this U.S. innovation is one I'd like to spend another minute on. I was actually planning on starting in a different place, but since we started talking about candy, It's— innovation is one of my favorite topics. And you've talked about— you've talked about U.S. as the center of innovation in other forums. I'd love your views on that and what else you see evolving, if anything, and what we need to do to keep it this way. And then maybe we can end up what we were just talking about, your views on Europe. Well, the U.S.

represents about 5% of the global population, and that's a declining percentage. And yet, although it's hard to put a number on it, we probably represent somewhere in excess of 60% of the innovation. And so if we want to retain that sort of high disproportionate share, I'd say there's two strategies. One is to increase our innovation, and the other is to decrease their innovation. Now, the second element they've done quite handily to themselves until fairly recently. And recently they seem to be copying the good parts of what goes on in the US— universities that work very well and government research grants and willingness to do big engineering projects— more and more of the things that have made the US unique, other countries look at and copy. And the thing that's completely unfair is that they don't copy our medical costs, our legal costs, Amtrak, our post office, our political system. And we should have— we should have bundled these things together and said, look, if you're going to copy the good stuff, you got to copy the bad stuff too, because otherwise it's like an unfair advantage that you're getting where you have our universities, but you don't have our medical costs. You don't have our defense costs, our legal costs. And, you know, it's kind of embarrassing and it might mean that our 60% share would go down. Now, in a sense, the 60% share is kind of unnatural. These things don't erode rapidly. I mean, if you take the top 20 universities in the world, you can argue other than, say, the US and the UK, are there 0, 2 or 3, 1, 2 or 3 universities in the top 20 in the world that are non-US/UK universities. Is Tsinghua in that top 20? I say yes. Other people say no, but it's right on the boundary right now. And that's kind of amazing because again, even if you stick the British in, we're like 6% of the world's population. So they really messed up for a long time and we kind of got used to the fact that the other people have, have messed up. And so nationalism is a very scary thing to me. You know, when people say, okay, should we put tariff on our wonderful, you know, our energy market, we should have these tariffs. Well, our energy market is a very small part of the global energy market. We're not adding that much capacity in this country, whereas the Chinese and other countries are adding capacity. And so There is this sort of implicit chauvinism that historically we've always been the biggest market for advanced products. And now we're not the biggest market for advanced products in almost every market, whether it's computers, software, medicine. And so the idea that co-locating the research, the companies, the customers, that that created these clusters of innovation that that bootstrapped, some of those elements aren't there any longer. And so if we take a nationalistic view, I think we handicap ourselves. And, you know, I'm a globalist in the sense of my foundation. You know, we spend $4 billion a year to try and improve things. And about 80% of that is spent on the poorest in the world, not in the United States. It's those third who are in the toughest condition. And that's sort of an anti or opposite of a nationalistic viewpoint.

Let me carry this innovation question a little further. I actually think solving all the world's problems are— I'm an optimist, so I'll start with that. But solving all of the world's major problems within some reasonable time frame, like 20 years, is relatively simple. I love this group. But I have a model for it, and I'd love to get your views on it. I've always imagined if there were 100 portfolios like this portfolio of companies. And, and I'll be honest, our portfolio is based on our experience and frankly our set of biases. We have a set of biases. We like certain things, we don't like others. But if there was 100 such portfolios and we did 100 companies, Neet Portfolio did 100 companies, we'd had 10,000 companies. And if 1% of them were really, really successful, just 1%, and I expect a much higher success rate, you'd have 100 companies that would completely transform the planet. I don't know what you think of that model, and then if you agree with it, how do we create these 100 independently biased portfolios instead of Khosla Ventures biased portfolios?

Well, I completely agree that this portfolio is going to do amazing things and that we— it is strange that there aren't more portfolios like this. There are not. There are some other very good portfolios. But in my view, this is the strongest by a measurable amount. I think if you do other portfolios, you have to be careful though, because there's unlikely to be linear gains to the scale-up. That is, you know, if you go from 4 battery companies to 400, I think the additional societal benefit is pretty small because only, you know, 3 or 4— there are various different spaces there— of them succeed. I certainly have myself a set of problems. I'll just give one for fun that no one works on. But I doubt, given the sort of me-too nature of science and understanding of opportunity, if you did linear increase, you'd get that much increased variety. We would like to have— the Foundation would like to have what we call a drug depot, where one time somebody shows up and you inject somewhere. Say, a set of vaccines that release over time. And so instead of having to show up many times, you show up once, or you need TB drugs for 6 months. We just stick them in you and they slowly but surely diffuse in the appropriate time schedule over that time period. HIV drugs, you know, give you a year worth of birth control, you know, the ability for women to inject something and then either when she wants to change her state of being fertile, going somewhere, or, you know, just taking her mobile phone and saying, okay, yes, fertile today, you know, not tomorrow. I mean, that's not as silly as it sounds. And so there are interesting problems, particularly of the poorest. You know, we work on toilets, and I got to tell you, we dominate toilet research. I mean, you know, you think Microsoft had high market share? You ought to see my market share in toilet research. I mean, this is a field that other people just stay away from. I don't know a single venture capitalist who's focused on this. And yet the current standard, which is the flush toilet that involves all this water, and you know, you take this water and you make it very dirty, and then, you know, you have to go and do something with— I mean, it's the most You think growing cows to have steak is inefficient? Just think of sanitation and how inefficient that is to pipe all this stuff around. Anyway, I'll go on and on about it because it's— we're actually making some progress. We have 14 big grants out of a lot of different things. So I do think that we need more innovation. And but we have more than I think people recognize. You know, some of the big sectors that are still tough include education. And even though there's a tiny bit of a blip here in terms of edX, Coursera, Udacity, you know, who are slightly making the things better with a little bit of interactivity and drawing people in and thinking through the social aspects. Still the most systemic underinvestment in innovation in all of society is in its most primal element, which is education. I mean, it's almost laughable that, you know, things that help 78-year-olds not be bald, you know, gets more research money than education does. And And if I told you the average math teacher was better 100 years ago, could you prove me wrong? Would you have any piece of data to even discuss the topic? You know, are engineers better than 100 years ago? Yes. Are doctors better than 100 years ago? Yes. Even pitchers and batters, we know the numbers. I mean, this is an important topic. We actually measure this. We have feedback. We fire the ones who aren't good. We give more money to the ones who are. We don't do that with teachers. That would be, ah, it's not that important. But batters, man, we have a feedback loop and it's worked extremely well. I can promise you the batters today are way, way, way, way, way, way better than those batters from the past. And so the question is, what are your priorities?

So I now regret not investing in toilets. Honest to God, last year the winner of the MIT $100K competition was a toilet company called Senergy. And we came this close to investing in that toilet company. But I guess other voices prevailed, so we didn't invest in Senergy. But it did win the MIT $100K competition.

Well, there's still a chance to be the first venture capitalist to invest in toilets. Virgin territory.

Yeah. Speaking of that, we haven't talked much about it, but one of the panels today was by this Impact Group we have. And the Impact Fund we have is set up to fund entrepreneurs, for-profit entrepreneurs, who who are going to go produce services for the bottom 3 billion people on the planet, as opposed to the 3 billion people with the most money. Give me your thoughts of that model versus the foundation model, because the more I've looked at foundations in general— Gates Foundation accepted— I've been disenchanted with foundation work generally. And so capitalism for poverty seems like one avenue. Now, it can't do everything, but it could do a lot.

Give me your thoughts on that. Yeah, absolutely. The— there are a lot of problems that are very hard to scale up unless you invoke capitalism, unless you're using market forces. And so a lot of the things we do, for example, we fund the cost of new seed research. And it's one of the mind-blowing underfunded things on, on the planet. You know, the whole Green Revolution was based on about $50 million of investment that avoided mass starvation, raised nutrition levels in Asia, which raised IQ levels in Asia have been very key to the positive feedback loop that took place in large parts of Asia. Even the Chinese miracle starting in 1979, the fact that other parts of Asia had adopted the Green Revolution varieties and they finally put in place an incentive system, capitalistic, that encouraged farmers to pick those and use those properly. That's part of why they, off of such a low base, were able to do dramatically well. There are some things that I think are best done purely philanthropically. For example, inventing a malaria vaccine. It only needs to be invented once. I don't need a whole bunch of franchisees trying to do that. And I do think central sort of R&D management works very well for that problem. And I can't think of how you'd create a market model. But when it comes to distributing drugs, when it comes to doing safe birth deliveries, when it comes to education, those things have to be done in every location. And you're not going to find somebody who's philanthropically minded to execute on those activities. I'll take a simple example. We have a way of doing cataract surgery that's extremely inexpensive. And if you train the people, they can go and make money on doing it in the local community. Those people will be capitalists. They will not be philanthropists. It'll be a, a certified group of people who are allowed to make as much money as they can giving cataract surgery the right way and completely outside of the public health system. System because that has failed to work in this area. So there's a ton of things like energy, education, certain healthcare delivery that whoever can create capitalistic models is making a huge contribution. But when it comes to the fixed cost of R&D and the pilot programs that prove out these things. If you rely on capitalism alone, you will grossly underinvest in these things. You know, the people who want a malaria vaccine just don't have any money. And so the marketplace, they, you know, you're— the size of your megaphone is based on the size of your wallet. And so I have this, you know, continent-sized megaphone. And, you know, people who get malaria are just completely silent in this system. So if I wanted to build pyramids or have 1,000 blondes fanning me at all times, it will happen. Or brunettes, sorry, I didn't mean to offend anyone. And it— that's the way capitalism works. I have a very loud voice, and the people in this room have above average voices in it, and so it doesn't necessarily treat the human needs of all people in any sort of equitable way. The foundation, we have a threshold. If we're not saving a life for $1,000, we should spend the money on something else. And the only exception we have to that is that we've decided that about a quarter of what we spend should be spent in the US. And the marginal impact of that 25% is absolutely lower than the 75% we spend in foreign countries. But because the fortune was made— both mine and the Berkshire fortune were made here— okay, so, uh, maybe some payback is appropriate. And that's why scholarships, education, those things there's that 25%. It's by far the riskiest thing we do. You know, I know we'll eventually invent a malaria vaccine. Is it 7 years, 11 years, 17 years? We're going to do it. In terms of education reform, where there's a personnel system and appropriate use of technology, and we spend about 800 million a year on that, we could utterly fail. The money could be completely wasted. And yet, you know, I think it's sort of just in a way that it, that at least some meaningful portion go back to this country.

Uh, let me switch topics a little bit, um, uh, and talk about time and how you manage your time. Um, you have your investment group, the foundation, family, almost everybody in the world wants to meet with you. There's probably time slicing in a way that I probably can't even imagine. Talk a little bit about how you manage your time, what you use to manage your time, and your perspectives on it.

Well, the key thing is I don't watch TV, and I gain a lot of time. Back from doing that. No, you know, picking how you spend your time is very important, and you want to kind of track it and be serious about it. And I gained— when I was CEO of Microsoft, that was my dominant activity. I mean, it was CEO of Microsoft, some family stuff, and maintain a little bit of breadth of knowledge. Just for fun, to relax. And then when that changed in 2008, basically 4 years ago, I got some extra time. And you want to be careful about how you allocate that time because there's not many periods in your life where all of a sudden time frees up. You can develop habits or obligations that tie up that time without ever having to step back to say, say, okay, is this what I really want to do? And definitely I picked spending time with science-based innovation, both through the foundation philanthropically, through a set of activities that Nathan Myhrvold and Intellectual Ventures does, and through other activities including the great relationship we've developed I get to spend time on that because I find it fun. You know, I say to myself, "Boy, I don't understand quantum dots." You know, "I don't understand intercalation energies very well." You know, so next time I meet with people, I'll know those things slightly better. Maybe not as well as I should, but that's cool to, you know, kind of say, "Geez, I'm confused. What does that mean?" And— Anyway, so I intentionally chose to put a lot of time in that. I have some things where I multiplex, like if I'm on the treadmill, I'm watching an educational internet video or DVD. So I get 90 minutes of— it feels like exercise because I'm running, but I'm actually with this beautiful wireless headset watching an educational internet video or DVD of some type. And that You know, that's great. When I go on vacation, I get to read at least a book every day, which is a luxury, unbelievable luxury. And so I take a ridiculous number of books with me. I have books that have been around the world several times without being opened. Eventually I get to them. So everybody picks what's important to them to to them. For me, it's if I'm confused about something, I want to take the time to understand it. I mean, I'm— education, I'm trying to learn about healthcare costs, I'm trying to learn about. 3 or 4 years ago, I spent a lot of time on vaccines and immunology because that's important to the foundation. And, you know, then family time is, is fun and relaxing, and I get more of it now that I'm not not a CEO.

So I, you know, I have to say one thing I admire is how deep you can go in anything. The story I want to tell, last year I think when we met one time, you had just finished doing a college chemistry course all over again. And it's this idea, if I don't understand something, I'll just go learn it again. I really admired that. I was so impressed. I've remembered this and sort of been a model. But you brought up this issue of books. I feel like I try and read a lot. I generally take 5 or 6— I get through 5 or 6 books. I take 15 on every vacation I take. But tell us about sort of maybe the 5 most important books you've read recently, or 3 or 2. Some of my favorite authors are here from last year's reading. Talk about the books, why they're important to you, why you think they're important and—

Well, books, I love sitting and reading books. There's one book that I'm absolutely nuts about. I'm this evangelistic person. Steven Pinker wrote this book called Better Angels of Our Nature. It's 718 pages. It's— takes a long time to read. It is brilliant. It's about how mankind treats mankind over the ages and how that's changed— violence, murder, duels, slavery, genocide. And it not only explains those trends, but then it constructs a notion of what are the systems, systems of morality that we've used that have led to this what turns out to be mind-blowing decline in violence. And it's completely against sort of the common sense. If you say to people, "Hey, what's going on with violence? Is this a violent time?" People won't say, "Yeah, it's about 1,000 times less violent than it used to be." They won't say that. They'll say, "Oh my God, did you hear? There was a child kidnapped in Florida. I saw it on TV." I mean, there's this sense of, I mean, it's awful. You know, somebody spanked their child. If you had asked in 1955, parents, should your child be physically punished by the teacher? 95% said yes. Now it's less than 10%. There's this mind-blowing change. Anyway, that's an unbelievable book. There's a Daniel Yergin book called The Quest about energy and how things have changed. That's super good. There's a biography of Deng Xiaoping. You know, if you wanted to pick the person who did the most positive thing, that is, without that one person, the world would be worse off, I think you'd have to pick for that century Deng Xiaoping. You know, what he did after Mao's death about redirecting the government systems in the right direction still are kind of mind-blowing. And it's an amazing explanation of it. There's another book that's sort of complementary called The Man Who Stayed Behind, which is about an American serviceman who stayed and became a communist and worked for Mao. And Mao imprisoned him, and he stayed loyal to that. System. And so you get this incredible contrast between what one leader did and then with the sort of same basic material what, what that next leader did. Then there's a lot of books. There's a lot of books. You know, Peter Diamandis has this book called Abundance, which is a little Pollyannaish for me. But, you know, it's good to see that. Then there's one called Age of Austerity. There's all this breathless stuff about new medical technology. Topol's got a book out. The Atul Gawande checklist book is good. You know, you have to read fiction from time to time. There's one about living in the Mumbai slums called All the Beautiful Forevers that's really a kind of a wonderful reminder of what it's like to be a slum dweller in India or any poor country. One thing I do is I have a website where I, when I read books, I put reviews of them up of about, I don't know, half the books that I read. So my gushing review of Pinker's book I think is posted there by now. And that's kind of fun to explain why certain books are good. And a lot of people suggest books to me. I mean, that's one thing the internet's been amazing about is that Nathan suggests books. There's a lot of people, and then you kind of repay the favor by suggesting books to them. It's awkward if the number of books suggested to you is more than you can read. So the next time you see that person, it's like, "No, I didn't read that book." But I try and read whatever I can. And it's possible to increase your reading speed quite substantially if you choose to.

Yeah. I actually think one of the more important technologies we need is summarization technologies for text, given the amount of material that's become available, at least Feels like an essential tool for me.

Microsoft Word actually had that feature about 10 years ago, and people took like the Bible and said, okay, give us 10 sentences, or the US Constitution. It's very funny what that— technology is much better now in doing these things. It'd be a good Kaggle thing to do a summarization of long, ponderous things like, you take the Obamacare healthcare bill and boil it down to 10 sentences to try and understand it.

Time's going by faster than I imagined. Let's spend a few minutes on Microsoft. What are the things, whether it's in the '80s, we used to hang around at PC Forum, '90s, or the last 10 years, you do differently if you knew then what you know now? The reason I ask this question is most CEOs are so much smarter after they finish doing their jobs and would do things differently. Would love to hear your perspective on Microsoft and what you would do differently if you'd known then what you know now, or maybe you wouldn't do anything differently.

Well, certainly the— Microsoft got itself into a very strong position through DOS and through Windows. You can say 1995, somewhere in that time frame, was probably the peak of Microsoft's relative position in the industry. I mean, to be clear, there's many metrics one can look at. Microsoft's profitability today is about 8 times what it was back in 1995. So that's a figure of merit. But in 1995, you know, the percentage of ISV activity, the sense of, oh my God, whatever Microsoft does, we better pay attention to it, it was like, it was so high that it scared people. And, you know, that has some negative effects which we felt soon thereafter in the form of our own government suing us. That wasn't much fun, and if there was a way to avoid that, I think I would seek to do it. It's not clear— that's complex. It's not clear what could have been done to avoid it, but it was unpleasant and It's certainly damaging in a pretty fundamental way to be in a company where if you say, "Hey, we could gain share if we do such and such," and any email of that nature is viewed as a negative thing. And then people even get confused about, "Well, are we supposed to do good products that gain share, or are we not supposed to do that?" But I hope— I don't think many of you will run into that anytime soon. It seems to only come with sort of extreme situations. There's always the question about delegation. And there's a few scenarios that I wish I'd been more tops-down in driving the scenarios. I got very— particularly during the Department of Justice thing, there were a variety of people who convinced me in I was convinced that I should be more bottoms-up, you know, that, "Oh, well, I won't tell you what the schedule is. You tell me what the schedule is, 'cause I'm not gonna write the code." And, "Oh, you tell me what you think the feature should be." Well, okay, there's some of that that's okay, but when you want to really drive a feature set, you know, you want to advance interactive reading, you want to advance tablet with a pen, you want to advance voice recognition, digital payments, a little bit you have to be tops down. Now you don't have to be quite as tough as maybe Steve Jobs was. I think it's a little dangerous now that people sort of worship that model and they think that sort of unpleasantness is an inherent element of it. I don't think it is necessarily, although that's unproven since we're dealing with a sample set largely of one in those particulars. But I do— there are definitely some things between '98 and 2005, including tablets, which we did a bunch of, but we didn't do them in a tops-down unified fashion, or high-end phones. It's easy to look back. What about search? What happened there? Now, sometimes we had things like in networking where we got behind in networking and I don't know if anybody here could name who our primary networking competitor was, but— Noelle. Yeah, very good. In a museum somewhere. But you can catch up. And who knows? I think Microsoft's actually— I don't think we'll get back to '95 because Google's doing great work, Apple's doing great work. I do think that it's going through a little bit of a renaissance right now. Bing, if you blind the results, search results, Bing results come out, dominate Google results. You know, we've pulled ahead of them. Now, it's a dynamic game, but they have a basic logical structure where their thing is very much if statements and ours is machine learning that I don't think they'll ever catch us ever again. I don't think so. Now, our market share means that they make infinite money and we make no money. But when you have a better product, you're supposed to be able to do something about that. So we'll see. You know, that— what do you want out of a tablet and a PC? You want the thinness of a tablet, you want the input capability of a PC. Can that be done with hardware and software? Well, this year you'll get to see. Anyway, I'm very biased about this renaissance thing.

You know, I want to get a comment on one thing before I go on to my next question. It's surprising to me that Microsoft and Google have focused on search in a traditional way. Facebook, as dominant as it is, has focused on the web in a traditional way. Nobody's really looked at the mobile opportunity as a way to change the rules enough. I don't know if you have a— it appears to me that search in mobile is different. Facebook or social networks in mobile are different.— and I would think it's an opportunity to wedge in there. It was interesting when Dan Rose spoke today from Facebook, he said every product has mobile design in it. He didn't say mobile was the only thing we were gonna do, which is what I would have wanted to do if I really wanted to intercept Facebook's weakness. Just comment on this sort of opportunity in transitions and how they create opportunity.

Whatever your views are on that. Well, over time, you know, you'll have flexible screens that roll out to any size you want. The boundaries between phone, tablet, and PC and large screen on the wall that exist today, those are going to largely go away. That is, the ability when you walk up to your whiteboard that it's an intelligent device, that whatever you're scribbling up there, a piece of software will help you create that org chart, create that chart, create that map. You know, that software will just be watching, and every cool thing you've ever seen on a whiteboard, you'll be able to create. You can walk to your tablet and do that. You can go to your phone and do that. The— these boundaries— people are kind of worshiping these funny boundaries now in a way that doesn't make sense. Natural user interface, sort of cloud-based state roaming, a very interactive way of dealing with things that are across multiple devices. So No, I don't think ignoring intelligent whiteboards, ignoring desktops that will be intelligent screens, ignoring tablets and worshiping little screens— no, I think that's sort of taking the last 5 years and locking yourself in. There clearly was a change that took place there. No one knows how to make money on search on that that small screen. It's one of those infinite-size markets that everybody bids for. And perhaps that can be solved. But when you look at what is really important in terms of commercial intent— travel, gifts, mortgage— this is what search really gets paid to do. It's still done off the large screen. And as we get these kind of flexible screens and the device boundaries disappear, you know, I think you have to think more about that and how natural interface changes that. The whole thing about people and groups of people and automatically organizing groups of people, nobody's done that well to date. It's, you know, communications is more fragmented today between mail, Skype, Facebook, Facebook, Twitter, all these things. The notion, when I look at this device, what do I want to do next? Is software helping me or do I just have this time-ordered sets of mail, time-ordered sets of tweets, time-ordered sets of Facebook timeline things and I have to go back and forth between things? Anyway, so I do think there's many synthesis across silos still exist today, but for me You know, it's not like in 1995 where you said, okay, just base yourself on the internet. I don't think mobile is the rallying cry for the next 8 years. I think it's much more about the rich interaction across, across devices.

One other question on Microsoft, and it's more about what you think is really important in company decision-making. Vinod Khosla. I could ask you what are the 2 or 3 hardest decisions you made at Microsoft, and I'd love to hear your view of that. But there's a broader question I have that I struggle with. Is it really 2 or 3 hard questions, or is it more like a baseball batting average of 1,000 at-bats, and you bias the company in one direction or another? I don't know. If you believe in the first model or the second?

Every once in a while there are industry transitions. For example, the move from— this will really date me— from character interface to graphics interface. And, you know, that's one that Microsoft played properly, and we got ahead of the curve. And, you know, then there was the Are mobile phones basically internet browsing devices, or do they have unique characteristics? And that one we played it wrong. We listened to the carriers, we followed what they want, and Apple did a browser-based device with a decent screen and touch and all those funny phone-specific things that all those carriers were telling us that if you were customer-oriented you must do this and you must do that. That was a bunch of garbage. It was a complete waste of time. The tablet, I do think you can— a right bet could have been made in terms of a device that had good input as well as those things. But that'll happen. It'll happen 5 years later than it should have. So I do think there are some tops-down scenario things. What is the future of reading? What is the future of interactive documents? What is the future of commerce? What is the future of programming languages? How you express corporate-specific logic? I think there are things that somebody from the top just has to— not the very top person, but some small group has to say, hey, we're going to make this bet on this big thing. And there's very few of those. And then if you're a large company, the bottoms-up thing is an important dynamic of, OK, should we do something special in Japan? What is going on with Chinese piracy? But the big changes in technology, there has to be intelligence at the top that's picking those transitions. Most technology companies don't do that well after their founder's gone. If you put aside IBM, you consider John Chambers sort of a founder of Cisco, you know, there aren't— I can't think of examples where technology companies of size have done well without their founders, and that's kind of embarrassing for the industry that nobody's created this enduring thing. And even the IBM case, in my biased views, is by becoming a non-technology company in the technology space, kind of a weird thing. Yeah. That they've managed to maintain from the stockholders' point of view quite a valuable proposition.

So what are the 2 or 3 really critical decisions, say, in the last 15 years, post the internet, 1995, post your internet memo, if you remember that? What are the most hardest decision— the 2 hardest decisions you had to make?

Well, I had to decide whether I would retire and go work at the foundation full-time. That was a tricky decision. That's a little self-centered, I guess.

I suspect it seems obvious in retrospect.

There became a time when this lack of top-down direction clearly was an error, about 2003, 2004. And I knew I— by then I had decided that I was going to go to the foundation. And so deciding that I want to create a system that had more top-down elements when I wasn't sure who those top-down people would be or not, that was an interesting question. And picking who your visionaries are going to be, which, you know, Microsoft has a set of people that we bet on, that we feel good about. Those are the hardest decisions. It's the people you pick. You know, the fact that natural interface is going to be important. Okay, that's clear. But which thing when, you know, people— we've had a lot of people tell us we should get out of search, even board-level people. And I don't have much patience for that. Maybe I should listen to people more about things like that. You know, you always have this question, if you're not full-time, should you give people your input or not? That's tricky. There's sometimes where you do. You have to try to decide if you've stayed up to date enough to make that, that relevant or not. Now, sticking with things is hard. I mean, breakthroughs don't come around that long. Xbox has required a lot of perseverance. Bing, a lot of perseverance. This sort of tablet PC uniformity. And there's some revolutionary stuff in Office that needs to be done that really picking who could do that and would it survive in the— or should it be done inside the Office organization, outside the Office organization? That was a very big deal. I think the best thing I ever did, although it's got to be ended with other good things to really show through in the future, is how we created Microsoft Research. And, you know, right now I'd say today in terms of IT companies that do research, really good research, Microsoft is distinguished. Now that alone isn't enough, but it is something and it's very important the way we built these research centers and developed them even versus Google, Apple, whoever you want to name. They haven't done that and that should create a better understanding of options in the future and more things to pick from to change these digital experiences. It hasn't all— we haven't taken advantage of that very well, but I'll still stand by that as, as probably the best set of decisions that I made as to how to structure that. Okay.

Vinod Khosla, we are running out of time, but I want to cover two quick final topics. One's energy and the other's education. I know both topics you love. Given the physics, even with all the innovation and efficiency improvements and other things, do you think— you know, the question I always pose is 500 million people live an energy-rich lifestyle today. Could 5 billion people live an energy-rich lifestyle?

No problem. No, I mean, that's our job. And, you know, you can argue about whether it should be US levels or European levels. You know, there's a little factor, slightly more than 2, that sort of Americans, Canadians, Australians, we live in this overuse of energy thing. One guy who's an energy pessimist that— the best writer on energy ever by far is Václav Smil. All of his books about technology, about fertilizer enriching the earth, Why the U.S. Is Not the New Rome. He's written, I think, 14 or 15 books. Two of them are about energy and about how One's called Energy Transitions that reminds us that energy infrastructure doesn't change very quickly. And that's why I'm slightly less optimistic than you that 20 years from now things will have changed that dramatically. But I'm basically in your camp. I mean, you know, we can argue 30 years, 40 years or something like that. And I don't think it's that important in terms of the size of the opportunity for the people here and how you invest in things. I just— I think there's more hysteresis in the system in terms of the installed base and willingness to try new things. But I absolutely— if we can't provide essentially European levels of energy to everybody on the planet, we will have failed. And science is mind-blowing. It gives us all these paths to get there. I mean, there's a ton of renewable paths. There's There's solar chemical, there's biofuels, there's nuclear, there's carbon capture. And no, the path forward of which one, and you can divide it to electricity and transport, but the further out you get, the more those really look like one thing. It's not clear which path we're gonna take. So it's not like the Manhattan Project where we'll just say, okay, let's all go to Los Alamos Tonight, and we'll stay there for a couple of years and solve this thing. We need a bunch of people trying different things, but I don't have any doubt that we can achieve that kind of energy level for lots of people. Whether or not we mistakenly allow a dangerous level of warming to take place somewhat related to that, that I'm a little worried about. And that's why I fund geoengineering that could buy us the 50 years, which is what I think we're kind of off on in terms of the speed of invention and deployment to get the CO2 emission levels down to avoid temperature increases that are at least dangerous. No, it's not like Hansen says, we're definitely, you know, is, is a disaster, but it's courting disaster, particularly for tropical agricultural activity. And for species die off that, that can't be restored. So that one is a little bit of a footnote on my optimism is I don't— that type of long-term problem, our governance structure doesn't work very well to take slight pain now to avoid potential but not proven. I mean, yes, warming is there, but the dynamics of the negative impacts are very complex. Potential large pain out in the future.

I completely agree with you. Having talked to Vinod Khosla, I can tell you he hasn't— my— we mostly agree. Where we disagree is he isn't exposed to the entrepreneurs in this. Absolutely.

I'm so much more optimistic than he is. I mean, he's brilliant, brilliant, and you should listen to everything he says, but he is wrong.

Yeah, no, because every time I've talked to him, he doesn't allow for the possibility of black swans, that he has a static view of innovation.

Now, he would say that, okay, natural gas, you know, was available 50 years ago and it took 50 years before it was 20% of the world energy mix. And so he'll show you that nothing happened in less than 50 years. Nuclear is actually the fastest. Although it's fading now, but, you know, it was invented in the war and then actually got to meaningful deployment in the '70s. That's the quickest any new energy source got big. But the past does not predict the future. The amount of capital and IQ we have today is not, oh well, we couldn't do it in the '20s, we can't do it now. And he grew up in Eastern Europe. He doesn't believe in capitalism. He doesn't have a natural sense of He's up in Manitoba. It's very cold up there.

But yeah, no, I completely agree with you. I'm gonna ask you one last question from me, and then I'm gonna take two questions from the audience. And what I request the audience is raise your hand if you absolutely must ask a question. I know 100 hands would go up, but I'm only gonna take two if you have a question that's relevant to most of the people here, you get to ask it. While you think about that, and I'll pick two at random, one from the left and one from the right. Bill, I'd love your comments, since you talked about education, on the most interesting, innovative things you've seen in education, things that might give you hope.

Well, I think two exemplars of education innovation are Sal Khan, KhanAcademy.org, And if anybody hasn't used it, if you want to refresh your understanding of various scientific things, if you want to understand the financial crisis the US went through and the various numeric elements of that, what is the Fed balance sheet, why did they have to do what they did, how did we get confused about housing price appreciation and correlation in these various securities. Anyway, it's quite a brilliant thing and it's a, strong phenomena for individual learners that we're trying to map into a tool to be used in the classroom in this flipped classroom way. So that's super interesting. Coursera is an exemplar, the leading exemplar of actually trying to make the interactivity— the best video lectures are still The Learning Company and Charge4lectures. There's some good ones on the internet, but clearly for most learners If you interact and practice, it draws you in, it engages you in a richer way. And Coursera, although it's very early days, is creating toolkits to create better interactivity. And there's some improvement in answer analysis. You know, Kaggle had a text analysis contest. So over time, machine scoring of things will move from multiple choice to a broader set of things. Maybe not matching humans, but coming awfully close. So the idea that for a motivated learner, if we put the right social elements around it, we can change education— I do think you can see that coming. There's— Manga High is good. Grokut is a $10 a month thing that has this social phenomenon of people trying to learn things together. There's Teachers Paying Teachers. There's a lot of very cool stuff that the digital revolution is starting to point us in the right direction. And so I think certainly for motivated learners, education is getting a lot better. Now, how we take the unmotivated learners and draw them in to be motivated is, is a non-trivial thing. But I think that will be challenged as well. And we have to, because otherwise these costs of education are essentially making it available to a narrower and narrower set of people, both in the US and globally.

So we'll go with the last two questions. Ronnie, you get to ask one.

And then we'll— OK, so I've been to quite a few of these conferences over the last few years.

And every year, there seems to be a change in topics.

Bill, maybe I'll ask you. 10 years from now and you're sitting on the stage again, what do you think Vinod will be asking you about? He'll be asking about how robots have made humans feel kind of impotent and unimportant because if he thinks doctors are replaceable, believe me, people who cut hair, who guard buildings, who— I mean, someday, of course, software is very powerful stuff. But my god, doctors are not at the bottom of the list of things that robots, software in general, will be able to do. And the milestones we'll have in the next 10 years in terms of mechanical agility, picking things up, moving a patient from a bed to a bed, moving around. It's not that hard. Okay, I'm not a mechanical engineer, but we are going to have some wild robots, and that's gonna screw up people's minds a little bit. So he'll ask me to explain that. Thank you.

I have to say, we were at TED recently. You weren't at TED this year. This guy from Denmark— I think it was Denmark— who walked in a guy called DK, you can Google him, with an exact replicate robot. It was almost like a 3D print of his image. And when the robot talked, you couldn't tell which one was the human being. And so the robot asked this guy DK, "How did you feel last month when those people who visited us confused you for me and me for you?" I— honest to God, you couldn't tell. It was the same hair, they were dressed identically. It was amazing. You should Google that TED Talk. On this side, who really wants to ask a question? Bindu, I'll go with you.

Okay, good.

So— oh, thank you. So I believe like Microsoft is kind of the epitome of capitalism and has also exploited capitalism in some ways? You know? Yes. Yeah. Bill Gates says yes, so we know that's true. Better than Goldman Sachs, you think? I'm sorry? Was that better than Goldman Sachs? That was like a— did you exploit capitalism better than Goldman Sachs? But that's not the question. The real question is, I think you understand capitalism really, really well. And I think you understand kind of the deficiencies in capitalism really well. Right? And I think what you can do right now— and this is my own perspective— is write a book about how capitalism fails and how it can be improved. Because I think you understand that well. And I think that's partly what makes you want to give half of your wealth to charity and work on malaria and blah, blah, blah, blah. Right? So I want to see a system which actually solves things like malaria, AIDS, and whatnot, but works within the incentives of capitalism.

Great. Right? So Bill, your perspectives on capitalism and where it fails. And I have a follow-on question because I have a particular bias on capitalism.

Well, to isolate any single element of the system we have I think is always a bit strange. We have a mix of government, capitalism, a small slice of philanthropy, academia. And, you know, for the US, you know, say from, you know, 1962 to 2012, 50 years, that's worked extremely well. I mean, real GDP is up a factor of 3. It's been a good 50 years on any type of absolute or relative measure. And it's because these elements have worked in concert. It's not because— you know, we are less capitalistic than many other countries. You know, we have Amtrak, we have a post office that— I mean, we have a lot of non-capitalistic things going on in our society. Society, which those particular examples are just waste, basically. We— the greatest concern I have is that as you get rich, the two areas of the economy that grow most rapidly are education and health. And so the degree that you can't bring the feedback loops that you have in other elements of the economy— call that capitalism, but they're really just feedback loops, which money is, is a key element of. To the degree that health and education are not measured by excellence and shutting down inefficient and, and rewarding efficient, then your overall economy, because it's putting more and more resources into that thing, is improving life more slowly than it, than it should. And so I think bringing appropriate mechanisms— I'll use the word market mechanisms, although that'll sound a little tooth and blood to people. You know, good personnel system would be an example of that. We better in health and education bring what we have in other endeavors, like what computer should I buy or which restaurant should I go to? Where it is a very well-functioning market and nobody says, you know, nobody's going to Italian restaurants, shouldn't the government bail out the Italian restaurants, and the Italian restaurants union, you know, thinks they ought to have pensions for life, and, you know, they're nice people and they vote for a particular party. It doesn't work that way. It's a system that's very dynamic. And so it is concerning The world looks to the US to lead, and neither in terms of managing the deficits or these two major sectors are we an exemplar. And it's possible some middle-income country, you know, will show the world how to solve these things, but not that likely. And, you know, to get back to the basic theme of innovation, I do think that both of these areas can get on a very, very positive track if we figure out how to reward innovation, because there are innovations in healthcare that do save costs and improve lives. And so if we over-reward those things, you know, make the indemnification strong, make the patient trials available, make the speed speed that regulatory things happen very different for things like preventing Alzheimer's, preventing Parkinson's, that we can get medical costs down even though we're on a path to hell with our current incentive system. I do think that's true in education as well. And so capitalism isn't some generic thing that automatically works. It really needs to be shaped by government driven mechanisms that were falling short within those two sectors today. Philanthropy, unfortunately, is at best a Band-Aid— not a Band-Aid. That's too simplistic. It's a little more than that. It's like an ankle wrap. It's, it's more— it can take things that are really screaming out that are small scale, like, you know, $600 billion a year spent on K-12 education. Our foundation of the $800 million we spend on education, about half of that's about this personnel system thing. And we have a better than 50% chance of that catching on and making that $600 billion at least 1.5 times more effective than it is. So even a venture capitalist would find that an exciting return numerically. And so yes, philanthropy can play a role.— but you can't count on it. It's there for new and unusual things to try something out. So government and capitalism are the dominant things you have to get right, and then philanthropy and academia kind of come in to try out new things. So yes, you know, I have thoughts about how we tap into capitalism, but as part of an overall system not in isolation. Well, someday I will.

So just a narrow follow-up question on capitalism, my particular beef. You know this well. I've seen it firsthand. When the system we live in, the capitalist system, I call incumbency capitalism. By that I mean the following. Piece of legislation that impacts anybody is influenced by swarms of lobbyists walking the halls of Washington. Most words are picked after midnight in various committees among— you know the process. I think within capitalism you can have that system where the incumbents engineer the laws to their advantage. Or you can have innovation capitalism, where the system is biased towards innovation. Since we've talked about innovation a lot, let me give you a simple example. When you give accelerated depreciation on capital investments, it's mostly for the incumbents, because the incumbents do that. If you get an R&D tax credit, it favors innovation. That's just one little example of— But very few large companies are interested in R&D credits if they don't— in return for trading off depreciation. So within the definition of capitalism as we understand it, there's various flavors. I call it incumbency versus innovation capitalism. I don't know if you agree or not. And then we'll call it a night.

Yeah, the— well, if I say yes, I guess that's easy. But I'll qualify that. Yet capitalism, the biggest flaw is that it, it underrepresents the needs of the poorest. The second biggest systemic flaw is that innovation, it grossly underrewards. That is, if you invent something, the percentage of the benefit that comes to you is such a small percentage that on a societal basis We invest way less in R&D than we should. You know, education just happens to be an extreme example of that. But that is true very, very broadly. And we offset that a tiny bit by having government R&D in areas like National Institute of Health, which is over half of global spending on health research coming from one country. The R&D tax credit is another example where in a small way we offset that. But it's still— it's, it's amazing how little we put into R&D. And it's wonderful that venture capitalism and the enjoyment that the people here get coming up with these new ideas offsets that a little bit. So the work that's done here has societal benefits that are so large compared to the amount that will accrue to the capitalists involved, you know, that's, that's a great thing. And if we can tilt capitalism to take more clever ideas in that realm, which, you know, right now in the medical and education space, I have specific ideas about how we could improve the incentives for innovation. That does remedy a truly dramatic flaw in this wonderful best of all systems. No, alternatives, so we better just tune this damn thing, uh, system that we, we live under.

Well, thank you. You've been really generous with your time. I know it's late.

Uh, thanks very much. Please, thanks, Bill. Great. Good job. That was fun. Thank you.