Other areas. I sat down with eric schmidt in new york at the new York Economic Club for a wideranging conversation. Here is that dialogue. Let me begin with a couple of questions. When you went to google, they said to you, we need an adult in the room. [laughter] because it was a good day for you. Eric it was a good day for me. It is an honor to be here. Thank you very much for inviting me, and charlie, thank you for doing this as well. Larry and sergey decided they needed someone to sort of run things, and they had spent 16 months interviewing people, and they make them do things. To become ceo, you have to go skiing with them, or you have to go to burning man with them. And very few people met the test, apparently. But i was fortunate enough. I told them i would not spend the weekend with them, but we made a list of things to do. Charlie 10 years as ceo. What was the most wonderful thing . For you, for the company as you see it . Eric my world is full of beautiful ideas that are not , in the sense that people have amazing ideas. In order to build a company of the success of google or facebook or uber, you need to but a technological idea also a significant change in the way revenue will come in. We invented targeted advertising which is really much better than nontargeted advertising. And we rode that really hard, and that gave us this engine. The same with microsoft had this engine of dos and windows, we had this engine that allowed us to build systems not with a particular revenue plan. We have been able to fail at big projects without two much of an issue for our shareholders. Charlie so you can risk. Eric we can take risks, invest in ideas, and do crazy ideas. Things that i dont think will work, but they work or they dont, we cancel things, its part of the culture. I learned, thats not normal. Most companies are mocked in these quarterly cycles, debt structures locked in these quarterly cycles, debt structures, and so forth. Charlie today, what is the role you have . Eric i am mostly working on science. I spend a lot of time working on policy, trying to understand how the world really works, trying to make sure, frankly, the governments would not screw this kind of amazing thing happening up, in the form of the internet. Charlie is the risk of that . Eric of course. Whenever you are affecting to medication and information, governments have a role to play. We have a significant battle with china. The democracies are generally ok, as long as you are on the side of informing people and reasonably fair. Charlie will google be back in china . Eric i hope so. We left in 2010, because they have very strict rules about censorship, and we were unable to operate, morelli, from our perspective morally, from our perspective. I spent time trying to get it reopened. Its really up to the chinese government. Charlie i wanted to talk about an issue that is strong with you. The future. The notion of the mens shot. Were of the moonshot. We are looking at it in cancer. What are the possibilities a moonshot can do for us . Eric i have come to a sort of view that we are operating under a zero sum set of assumptions in our society, and im including the western world, not just the u. S. We are not asking enough of our people, not trying to build things that are transformative. Go back to the interstate highway system, which was to movely justified people around, so the lack of interstate highway system would cause america to not grow at all. It is largely from interconnectedness and innovation. Interconnectedness, from bringing the world closer, intellectually and informationally and all the Things Companies do. And it also, new inventions come along, and every once in a while there are things which we have a consensus, right, and if we just get behind it. We call these moonshots. The Vice President did a cancer moonshot. Shaun parker has donated 250 million to some doctors that have learned to affect white and red blood cells in a major way that might be a cancer breakthrough. These things come along, but we dont talk about them. We spent our time arguing about political issues which are largely not that important compared to, how do we solve these massive problems . They can be solved. So the two biggest things going arein the world that i see, this incredible revolution in medicine, and is incredible revolution in knowledge, and the two of them become the basis, i think, for many things we do. The cancer moonshot is powered by these cancer breakthroughs, and we are able to essentially marry the analog world of cancer and biology and the digital world. Charlie there was the industrial revolution, and then the information revolution. Where we now, and what is the next revelation . Eric there are two phenomena that i think will be transformative in the next decade. The first is an health, biology, and the second is in knowledge. In health and biology, there has been a breakthrough in a way of gene editing. At the moment, it is a very tough hammer. They use a piece of gene to reassemble components. Gene editing at the basic form. The combination of that and doing databases of genes and sequencing and so forth allows us to really probe into the molecular and biological structure of life. It is certainly true within the next 10 or 20 years, you will be able to get a body part generated out of stem cells that come out of your blood. It is an extra ordinary achievement, one that i think was made in japan. Charlie and removes it from the political controversies. Eric the politics were a stupid argument anyway. The core issue here, you need a new body part, and we can be generated from euro and sells. That is lifesaving from your own cells. That is lifesaving for people who need transplants. This is real combat. People are dying, and this fixes it. Why dont we do more of it . Another one, ill get on cars in a second, but the core part here is, the combination of all of that is current. Why is it occurring . Partly because technology and science, but also because there is a great deal of money at stake, because the Health Care Industry sees new treatments as new sources of revenue and new billiondollar drugs, so you have a good alignment now of economic interests, venture money. The stuff is risky, so venture capitalists, some of whom are in the room, some will make money, someone lose money. Thats how it works. Information google is working hard on the concept of assistance. The way it works, and this is all opted in, with your permission and so forth, it uses all of the knowledge google has generated, the knowledge graph, we understand how language is spoken, we have 17 years of theories on those things, to help you out. Versions, werst built a messaging app which can reply for you, and it sort of learns what your patois is and knows what to say. An first foray in that was email product that would automatically reply to your email. So we launch this thing, and the most common reply is i love you, which turned out to not be the correct answer in a corporate setting. [laughter] so we have bugs. But charlie let me stop you there for one second. The idea of a Virtual Assistant coming out of Artificial Intelligence is, everybody is trying to do that. We have amazon already on the market with echo, and people could up and say, alexa, what time is it, what is the news . All the Major Companies are there. I assume the competition will be good for the end product, but is google behind the curve on that . You dont have a product on the market. Eric we have, we Just Announced a product with different technology. We will see how well it does. This is how our industry works. We are far more collaborative than competitive. Everybody won to focus wants to focus on apple versus google, but the whole ecosystem moves forward, and it is building those platforms, building that knowledge, its reasonable to expect that in a decade the vast majority of York Computer interaction your computer interaction will be by voice. 10 years ago. , i protected this would never happen. Shows you how right i am. But you look at technology, and the gains in voicerecognition, and alexa, which is now common, and you see it. Particular to the voicerecognition part. Im interested in voicerecognition with knowledge understanding. Charlie and attaching to the data google has. Eric or the underlying algorithms. We have a product now or you can get on your phone and speak in your language, and it comes out on another phone in another language. Does this really work . Yes. Is it as good as a human translator . Not yet. Is it good enough to have a casual conversation . Yes. How does it work . Turns out it takes your voice, it digitizes it and puts it into text, which is done using a neural network, an ai concept, and it uses a different translation neural network, which learns how to translate by looking at language pairs, and then translates it back into voice. So you have three different translations to go from voice to voice. Charlie is there a name from industrials to information, for the age we are looking at now, the transformative age of all of the technology and what it is doing for us . Is there a name . Eric there is not a consensus name, but i can define it a little bit. Another example. Theres a game called go which americans dont typically know about. It is infinitely harder than chess. And a group, a subsidiary of havee, called deep mind been working for a long time to develop cognitive intuition. They developed an ability, the ability to take a game in the form of bits, they can watch the bits on the screen, and with enough work, they can, playing enough games, they can figure out what the objects of the game are, how to win it, and then beat all the humans. Thats pretty interesting. Whats interesting technologically is that you dont have to tell it what the game is. So how does it do this . Well, that sort of watches for a while, sees common patterns, and begins to develop a base of knowledge, and then turns against that base of knowledge. This is not a human intelligence we are nots not, making that charlie you said yet. Eric we dont know how far it will get, but we do know this has never been done before. It applies this to the game of go, which is thought to be noncomputable, and it learned how to only look at certain positions by the same rough training mechanisms, and we decided to have a game against the best player in the world, in korea, who is a perfectly nice human being. Which wasm 41, historic. Charlie historic and huge. Help us understand. All the things you just talked about, Artificial Intelligence, a friend of your said to me recently, eric is thinking a lot about Artificial Intelligence. On the one hand, you have deep mind, which is able to beat go. A huge thing. On the other hand, you have lots and which watson which won jeopardy. Explain Artificial Intelligence. In this audience and elsewhere, everybody hears about it, some people make investments in it, hedge funds, others eric remember what i said about the biological role. Its happening because there is a confluence of a platform, a set of ideas, and a large amount of money and a lot of investing, a lot of people coming out of school, working on it, and the sense it is transformative to everyones lives. The same thing is going through. With respect to ai, the current ai uses, a couple that are similar. If you are familiar with a disease called diabetic retinopathy, because diabetes, tragically, is taking over the world, it causes your eyes to go bad and you become blind. It can be detected by an ophthalmologist, but we did a test where we took pictures of your retina, and we could do it better than an ophthalmologist. How is that . We see more eyes. We sell one million eyes. It is hard for an ophthalmologist, as hard as they work, to see one million eyes. Theres a large number of cases where if you just let the computer see more examples, they can come up with better decisions. Another example. Theres some, we believe, you can apply this to oil and gas is to beat networks. Theres a great deal of leakage and sort of, decisions that are made about flow and storage, and so on and so on, and by using that data, we think we can reduce the emissions from that. So thats by the way, the emissions are also hurting the industry. How far can this go . We think we can develop enough intuition that a physicist or a chemist could say you know, they wake up in the morning and say, i want to combine this chemical without chemical, and i want to have the following crazy reaction. Its not going to blow up, and it will produce a new substance that will get me a nobel prize and a promotion at work. So they do that, right . Thats not a clock in the morning. 11 00 in the morning they do that, and they have the subsidiary chemist do it and it fails. So they have dinner, and the next morning they come up with another one. Thats how it really works. And these are incredibly intelligent people. To we can ask the computer go through all the combinations and give you probability. Does that matter . These are trillion dollar industries, globally, around chemicals, synthetics, drugs. Charlie with respect to Artificial Intelligence, on the one hand deep mind has a different operating idea than ibms watson. Explain how they are different, and what they are trying to accomplish on their own . Mind tson and date deep mind are doing completely different providences. Watson, they use a particular kind of inference model, thats the technical term, which works particularly well for jeopardy and complicated problem solving, and they are having a lot of success applying this to complex systems and explaining them. Think of it as, theres a complicated system, and they can read it and tell you what is in it. Complexity, and you take away, however it works. So you can read all the contracts, and get an answer. Facebook announced this week, they have a breakthrough in language understanding around communications, sort of what they do, and they say they have made progress in detecting hate speech. Thats clearly a good thing. In our case, we took a position we wanted to build an underlying platform that allows you to do all this, so we built a network, sorry to describe it technically, that is called tensorflow. A multidimensional matrix, and these are multidimensional matrix algorithms of one kind or another, and we have built this framework to all our competitors. It is so strategic to us to build, that we took this amazing intellectual property and gave it out. Charlie you bought deep mind. Some will say to you what they are trying to do at deep mind, they try to understand how humans think and bill from there. The ibm people will say it is man plus machine. Eric the deep mind people are interesting, because the founders came out of neuroscience. Buildy imagine you would Computer Systems that use the same way we do learning, and i will give you an example. I wont do a good job of this, because im not a neuroscientist. But think about it, when you came into this room, how much cognitive time did you spend figuring out that the floor was where your feet went, the lights were up there, the table was there, and you had a knife and a fork, and there were roughly eight people, and everyone is dressed properly . Zero time. You had already learned that. You had chunked that, if you will. Theres evidence that our brains will study a scene, and then we will subset the things that are really important,. And the things that are really important, we then put into the brain. This is called reinforcement learning in their bernanke letter, and we believe free their vernacular, and we believe this will be a core part of ai. Charlie interesting, in terms of business and economics and the Global Economy take these five companies, apple, amazon, google, facebook, microsoft is their race to do any one thing, are they all in the same business . We know apple has made a fortune on smartphones. A fortunes made a smar on search. Microsoft made a fortune on software. Fortune on all kind of different things, but are these companies in pursuit of a holy grail . Eric Tech Companies as a group are highly competitive, seeking new, i want to say the right thing think of each of these as a platform driven by innovation that solves a problem that is global, in some cases a problem you did not know you had, but you discovered they solve very well. So if you go back, microsoft being the eldest of them, saw a problem of interoperable workstation platforms, and we know where it went from there. Amazon started off as a virtual bookstore. I did not know i wanted a virtual bookstore, but now that i have one, i know its fantastic. Apple s transformation was legendary. We saw the resuscitation of the company, and when steve took over the second time, he clearly wanted to do phones, and led the category. Each of these companies we have never in our industry had, and i would argue the first four are leading the platform fights, we have never had that Many Companies fighting so brutally against each other, yet also collaborating. In our industry, we always had ibm and microsoft and a few others. Charlie where is the collaboration . Because you are also suing each other. [laughter] eric thats normal. Theres lots of collaboration. A typical example is our apps all run on microsoft, or run on apple. Charlie last year, they announced a partnership with microsoft. Eric the traditional model would say, we will not put anything on that platform. It turns out apple is a customer, in our case, and we are also a customer of apple for things. On and on and on. In fact, everyone sort of argues with each other, but we benefit, we as an industry have benefited from Global Markets open markets, globalization, a sense that markets are transformative. Charlie should we insist on open sources, in terms of the future, in terms of all that is being discovered, or are we looking at a world in which each organization is going to be jealously guarding not only what it knows and is learning, but also trying to hide away and protect itself from losing their most talented human beings . Eric this is the genius of competition. We fight brutally to hire the top people, to get our products out. Think about the Android Phone or the iphone you use today. Think of how powerful it is. I figured out, it is 100 million times more powerful than the computer i used when i was in college. By the way, it cost a lot less. There was only one of them in college, and i stayed up all night to use it, because it was the only way you could get access to it. So it is real consumer benefit. I think the competitive structure, the architecture you described of those four or five companies, will continue for a while until another