It's great to be back, wearing my almost-Stanford red. I'm excited for you all to graduate, and I'm excited for the first-years to make it to the second year. It's just wonderful to be here.

I will tell you that, in the 14 years I've been teaching here, the faculty, I think, has been pretty much the same, and the students are just getting more impressive, right? So you really are the best. My guess is that 10 years from now, the students will be even stronger. That's extraordinary for the statement of what Stanford has been able to do.

What I wanted to do is talk for a few minutes about the age that we're in and the age we're going to. And then I thought, since I've been here so long and because we talk a lot about these things, I'd talk a little bit about life advice—some of the things that we talk about in our class. Maybe you can debate and disagree or agree on this or anything else, or perhaps present your latest startup idea and see how well you do in front of your peers.

I'm convinced we're actually entering an age of abundance. That age of abundance is defined by too much information, right? The first thing was to turn off your Twitter feed so that you could actually focus on something. Whenever people say, "Turn off your phone," I always say, "Don't turn off your phone. We're in the phone business." You get the idea.

We clearly have too much food, right? The world is not only capable of feeding itself, but in fact, obesity and weight gain and so forth are becoming major healthcare issues around the world. There are startups and industrial processes that will solve many of the housing problems we have worldwide, and we're going to have abundant housing of relatively high quality. It looks like we're going to have pretty ubiquitous transportation as a service, high quality also as well.

So a lot of the things that we've talked about as we arrive in this age of abundance look pretty good. But at the same time, it's pretty clear to me that while we're going to have sort of material things in great supply, we're still going to have religious, nationalistic, and personal conflicts to look forward to, right? I think that's the reality. Speaking as your local technology optimist, we haven't been able to quite fix those problems yet.

People too much in others' worlds, too much regulation, too much anxiety, a loss of egalitarianism—we've sort of forgotten the benefits of integration and tolerance, right? We seem to be backtracking on some of the core values that have gotten us to this august state that Stanford and we all are in today.

And it's funny, when you look back at the election, who would have thought that the KGB would be back through Twitter and Facebook? I mean, that's not the prediction we made a year ago here at the school, and yet it's very real.

So how are we going to react to this? Are we going to react to this in terms of regulation and suppression and so forth? Are we going to end up in the equivalent of suburban cul-de-sacs that are highly protected from this abundance? How are we going to sort this out? I don't think society has figured this out.

The principles that I've been operating under, which I would encourage you to operate under, are personal freedom and economic freedom, right? If you look at the history of the world and you look at where we've all come from, the fact that we got here is because we have those two things.

In fact, it's interesting—I looked this up—at no point in history have you all had a better likelihood of living a long and fruitful life, just on average and in specific. Very interesting, right? After all the complaining, your expected outcome is higher in almost every category.

A child born today is expected to live about 90 years. Children born by mid-century are expected to live at least 100 years. That's pretty good news. That's average, right? We all understand averages. And by the way, when you have such a long expectancy, my core advice is that you need to brush your teeth because you're going to need them. So let's establish some basic rules here: you have to think about these things.

And the world has this interesting way of changing. So when I graduated from college in the 1970s, I graduated as an engineer, an electrical engineer, and there were lengthy studies of the fact that there would never be jobs for engineers in the future. In fact, I was going into a job market where it was not obvious that engineers would have jobs at all; we might be doing other things. Of course, we wanted to be engineers, but I pressed forward.

The computer that I used in college has as its replacement, in roughly the same building, a computer that is 100 million times more powerful. I went to visit it. That's a lot of change in that period of time.

So if we're in this sort of age of abundance that I'm describing now, what happens after this? How does this transition? I'm convinced that the age of abundance leads to the age of intelligence, and I'll tell you why. It has a lot of implications—probably the defining thing for your business careers, and perhaps for your personal lives as well.

The technological boom, the parts that many of us have been involved with—people here in the room and myself—have produced the following facts: computer vision is now better than human vision, even if you're drunk, right? It's still better. That's a pretty big deal.

The analytical systems that are being built now can do better planning and analysis for any system that has variable inputs, that is variable supply and variable demand, right? Much of the business and dynamic world does that. And you say, "Well, why is he talking about this?" Because that will underpin much of what you do.

In medicine, it turns out shockingly we're all the same, right? All of us, all human beings. If you get enough of us and you do enough analysis of our healthcare outcomes, you can materially improve healthcare, whether it's precision medicine or better drug discovery.

The technologies such as the technology that AlphaGo represents, which is the reinforcement learning that won the game against Go (for those of you who don't know, it's a very big deal), is analogous to the kind of technology you would use to look at candidates for doing protein folding to apply to generate new drugs. So you go from winning a game to perhaps winning the real game, which is making us healthier and solving diseases.

So what does this mean for you all? Well, let's talk about cars, right? How many of you know someone who's had a tragic family loss—somebody died in a car accident? I think that in your lifetimes, and perhaps sooner, we'll get to the point where cars are as safe as flying on commercial airliners. I won't make a joke about that, in terms of accidents in the air. Basically, when you get on a commercial airliner today, it's safer than it is driving to the airport in your car.

So the engineering and the capability and the ability to do computer vision and so forth will in fact take something that is really... one crude way of saying it is that 90 people are scheduled to die today driving cars. We just don't know who yet. It's a horrific thing to say, isn't it? But those are the numbers. And by the way, they're going up, right? For various reasons, we need to fix this, and it'll get fixed very soon. In the next few years, the process will really begin.

So if you think about vision as a solved problem, then what does that do for radiology and pathology? So when you go to the dermatologist, you really want the dermatologist to use a computer-assisted way of looking at your skin to make the diagnosis, to help you get healthy and so forth. Why? Because the computer can see more skin; than the ophthalmologist, it can literally see more eyes if it's diabetic retinopathy.

So I am convinced that these technologies that I'm describing are so profoundly powerful that they will underpin the creation of some fantastic, large new corporations that are just in the process of being founded. It's always been true in my career—now this is my fifth or sixth cycle doing this—that when you get a wave like this, the kind of things that I'm talking about, the great new corporations get founded, right, by entrepreneurs who take advantage of the things that we've learned here, and the principles of open markets and the capital investment and so forth. We wrote a book about this.

So when I think about it, what is the emerging tech model of success? It's no longer slow product cycles; it's fast iteration. Learning your product, your app changes every week. Extremely deep sensing and telemetry data on everything that you care about in your business. Making the customer—or using the customer's information to educate you and learning from them.

Remember, if you've got a million customers educating you and you use machine learning, you can learn from them and become better than any single one of them. You make them better, and you get the benefit of their training of you. So any solution where there's a large training set, a large community, these technologies will transform the economics, the potential, the brands, and the impact that they have.

And to do this, of course, you'd put everything into the Google Cloud, or perhaps Amazon if not us. And you'd use TensorFlow, which is a new library that we released to do this. I think that it's possible that these next generation of companies, the ones that you and others will found, can get to 90% market share models, because you move so quickly.

That success model has not slowed down. You read the press and everybody goes, "Oh, well, you know, there's this big company and that big company." There's never been more opportunity than now to create these companies, because the barriers to entry are so low, the tools are so powerful, and the need is so great. And we can talk about this some more with questions and comments.

So I thought I would talk a little bit about being a student here, right? Being a graduate, being what you will be. And I thought about it for a while and I thought, this is a great year to be you, right? You managed to get here, and you're either about to graduate or about to graduate in a year.

And I would strongly encourage you to not succumb to the anxiety that everyone else seems to have, but rather to carefully go through the capability of the disruptions all around you and figure out how you can solve a new problem. You know, simplify your life. In this case, focus on the things that you care about: family, relationships, changing the world, making money—whatever it is, it's okay, but get focused on that. Don't be distracted.

Most people that I know, especially in your generation, spend too much time on things which are distracting rather than focusing, and often they're things that they really can't change anyway. Stop worrying about things that you can't control and try to make your own luck.

Now you say, "Well, how do I make my own luck?" Well, first, you've been pretty lucky to get here. Good answer, right? And I consider myself the same. I'll tell you the story of my very close friend, Eric Lander. He was teaching—I guess he was teaching at a business school—and he thought, "I want to do something new." And he decided to move to Boston because he figured that that was the highest number of interactions that would occur per unit of time in the areas that he cared about.

He knew nothing about genetics. He shows up, he wanders around, he meets a faculty member who turns out to be a brilliant geneticist, and working with him, becomes a brilliant geneticist himself, one of the most famous in the country. But he had to move to Boston because that's where the target was, and then he was willing to accept the risk. So sometimes life works that way, right? That you move there, but he made the luck happen.

And in fact, when I say this, remember that almost everything that happens to you happens because of the people that you're with, right? Every single success that I've had has been because I was surrounded by these incredibly interesting and successful people. As a technical person, as a PhD in my career and so forth and so on, but force yourself to take advantage of all of this. Don't just sit at home; live your life to the fullest. The most successful people work hard and play hard, which means you should not watch television; you should watch YouTube, right? It's very clear.

You know, this sort of "change we can believe in" is in fact you—your ability to adapt to this environment around you. I would also suggest that you develop at least a fake charm if you don't have any. Sorry. Take a picture of someone's child and send it to the parent. Explain to someone's parent how valuable their employee is—by the way, it works every time, trust me.

Don't say, "Nice hair," which is what I would say. Say, "Love your new hair, and the combination with that jewelry is sublime." Right? That's fake charm. It works very well. When you're asked what your favorite number is, always say 17. Shows you understand prime numbers, right? I mean, you need a leg up; you went to Stanford.

The best prefix in a sentence is the word "imagine". If I say "imagine" to you, right, that means I'm talking to you. I'm getting you to think, I'm getting you to be curious, that you're hopeful.

Seek meaning versus novelty. It's fun to have novelty, but meaning is really how you make your life happen, right? As you create accomplishments, it's the meaning and the impact that you have and the people that you worked with which will provide this incredible, incredible satisfaction.

One of my pet peeves with strategy is that it's always taught in a short timeframe, but in fact, the best strategies have longer timeframe horizons. There are always these traditional sayings which I like, which are: we tend to overestimate what can be done in 18 months and underestimate what can be done in 10 years. Watch a movie from 10 years ago or an interview in the technology space from 10 years ago and see how much has changed, starting with your mobile phone to make the point.

But almost all of the strategies, if you don't think about them, right, fail within a five-year period. So have a five-year plan, and of course, do it not just for your business, but also for yourself personally.

Be extraordinarily curious, and don't assume the status quo is unchangeable. I learned this from Larry and Sergey, who basically just eschewed tradition because I was sort of a traditionalist. And what do we have today? Publicly we have Trump and Brexit, all of these things. Almost all of the emerging tech startups are essentially disruptive in the economic sense and the political sense, and they were unexpected. You should expect the unexpected as a result.

Success then is determined by persistence, that is grit, as well as curiosity. And you have to combine both; you have to be willing to do both. So the way to be happy then—you put all this together—is to have meaning to your life. It is literally to serve others in what you do, right? Ultimately, that's the thing that will matter the most. It's why I teach here. It's why many of you will come back and honor the university that you're a part of.

But if I make a list, I think: stay open to new things and unlock more happiness. Be happy. Surround yourself with new things, new ideas. Avoid preconception, and know more truths, right? Literally know things. But it seems to me that in order to be disciplined, in order to get this—that is to have your own freedom—you have to have enough self-discipline and enough sort of focus, persistence, if you will, to get what you want. So it's almost an oxymoron, but it works.

Now you sit there and you go, "At the end of the day, what's your proof, Eric?" Well, poverty: $1 a day changed 80% in my lifetime. Plunged. So if you think that... I was sitting in a meeting a couple days ago with somebody who said, "Oh, you know, things are terrible in the developing world." And I said, "They've always been terrible, and they're less terrible." Right? Those are the facts.

My father took pictures when he was roughly your age. He was traveling around Africa, and you look at the pictures of Africa then versus the Africa now. It's still poor, but it's a lot less poor, to the credit of the people who built that.

I think to finish with this and go to your questions, I would say, I think you should recognize that inequality is going to increase in the world in our current economic system, and you need to be grateful that you're on this side of the inequality dispersion. I don't think we know how to solve that problem as a society, and it's an important one to solve. It's a problem, and you're on the good side of that. Be grateful about it.

And I also think that, as I said at the beginning, we seem to have lost this egalitarian notion, at least in the media and in the politics of many countries now, not just in America. I think you should recognize that diversity is a strength, tolerance is a virtue, that most people are doing what they think the world asks of them, and that the essential human virtue is goodness, right? Start there and then look for the exceptions, as opposed to the opposite.

And I'll finish by saying what I say to everybody that I meet who asks me sort of how to organize themselves: what I do is I tell them to say yes. Say yes to things. Just say yes. Just say yes, figure out a way to do it. Find a way to say yes to things. Say yes to invitations to a new country. Say yes to meet new friends. Say yes to learn something new. Yes is how you get your first job, your next job, your spouse, even your kids.

Even if it's a bit edgy, a bit out of your comfort zone, saying yes means that you'll do something new, meet someone new, and make a difference. Yes lets you stand out in a crowd. Be the optimist, see the glass full, be the one that everyone comes to. Yes is what keeps us all young. So with that, thank you very much.

Let's see. How are we going to do... do you have some mics? Comments or questions? Any topic is good. As we talked about before, if you have a question, just please make yourself known to Julie or Ali, and then stand up, introduce yourself, and ask Eric directly.

Hi, Steve R, MSX. I was interested in your concept of abundance, and that obviously feels right. One of the interesting things is that the societies where this abundance is occurring do appear to be becoming more unequal. I wonder how you think about that from a policy and technology point of view, and whether you think technology has any role to play, I guess, in the future. And is this a risk to the world?

So Jared Cohen and I wrote a book on this, and we looked a lot at the basics of connectivity. My initial assumption was that connectivity would be uniformly improving for everyone in the world. So one of the stories that we talk a lot about is the story of third-world connectivity, that is developing connectivity, and the power of very, very inexpensive smartphones.

And one of the things that I am proudest of, and Google and Android play a major, major role in this, is that that phone, when it arrives in the village, is far more important than the arrival of television a while earlier, because it's a source of entertainment, of information, of communication, of safety, of photography, and all of that. It really has changed for the much better. We're going to go from roughly three billion people online to roughly five because of those phones.

But then we went back and we said, "Well, if that's such great news, what's the math?" And the math is that while connectivity for those people is getting much better, the bandwidth to the developed world is increasing far greater. So if you take that simple dialogue, which is: capitalism and global trade improve life at the bottom, but the people at the upper crust are vastly more empowered with something as powerful as bandwidth—that's a simple example of this. It's obviously true of many other things.

So in industries where you have declining costs, which is communications, computer chips, silicon, and that kind of thing, I think the story is at least you can get it everywhere. And that's also going to be true of food. So remember, a lot of the food in the future will be essentially processed, manufactured in new ways.

Eventually, certainly in your lifetimes, some if not all of the meat you will eat will have been developed from real meat. It'll be grown in the equivalent of a culture, not in the equivalent of a cow. And you sit there and say, "Well, that sounds like a terrible idea." Well, it's going to have a lot fewer diseases, it's going to be a lot healthier for you, and people will work that out.

So the economic answer is that you're better off letting trade happen and letting this inequality occur, and then use redistributionist policies or other mechanisms to support the less fortunate. I'm roughly in that camp. I'm skeptical that... if you look at an aggregate basis, we're probably better off doing the trade, doing the investment, creating the companies, allowing it to become, and then making the difference with policies of one kind or another. You can do this with minimum wages, you can do this with various other forms of subsidies—lots and lots of benefits that are needed to make up for the gap.

And I should say, by the way, that what I just said is completely politically unpalatable because of the hollowing out that's going on in rural and manufacturing areas around the world. For example, if you go to France, Paris is very dynamic as it always has been, but there's a job problem outside of Paris, even in a socialist country, because of all these changes. Go ahead.

Hello, my name is Patrick Daniel. Thanks so much for sharing insights; I really appreciate it. So I was wondering to hear your thoughts on, you know, sort of the revolution that's happening with blockchain technology and, you know, decentralized networks. I'm really curious to understand how, you know, 200 years from now, when people look back at our time and the big turning points that will redefine society, you know, our civilization...

...if we think about it, like all these things are happening, you highlighted some of them, and I'm just wondering how you think about this whole clash between hierarchies, centralization, and decentralization, and all this stuff that's going on. Because as Edmund Burke, the historian, said, we have to have a contract between generations, and a lot of people don't understand blockchain technology—it's not very intuitive. So I'm just curious to understand how you see all these different decentralized trends and distributed networks, and how this affects society, because it's happening right now.

So your question is exactly at the center of what I mentioned was, I thought, the dichotomy on this issue. Because on the one hand, we're putting enormous power in the hands of individuals, and at the same time, we're terrified at the result of that in some cases, by overreacting.

The specifics about blockchain are pretty straightforward, which is that the general ledger that's part of blockchain is likely to be a fundamental computer science platform. The fact that you can have distributed ledgers that accurately keep information that's not copyable or fungible, if you will, is a major computer science advance.

But if you look, all the great optimists about blockchain and Bitcoin have been stymied by the historic structures of banking, which were erected after enormous financial failures, and they're not likely to be very liberal, right? It's just not who they... they didn't hire liberals to do this, by definition. The liberals were the ones who got them in trouble, you know, 50 years ago.

If you ask a question about 200 years, I think here is a build on how 200 years might look like. For the next five to ten years, the AI stuff that I'm talking about is incredibly, obviously beneficial, right? So I mentioned healthcare, I mentioned drug discovery, I mentioned basic science, transportation, these kinds of things. And it's happening so quickly.

I don't think people understand how quickly our lives will change in the same sense that 10 years ago—remember the iPhone came out in roughly 2007, right, so that was roughly 10 years ago—I don't think people understood the impact of having such a powerful supercomputer in their hand would have. I think it would be a similar thing.

What happens after that? A lot of people in my community believe that we're going to be able to get to the point where computers can have aspects of human intelligence that look like intuition, right? That they have memory and then they have the ability to do the equivalent of dreaming. I don't mean soul-like dreaming; I mean dreaming in the sense of imagining scenarios and putting interesting things together.

If that's true, then at some point, perhaps in my lifetime, it should be possible to feed, for example, all of math into this computer and all of physics into this computer, and behold some new discoveries and some real effect on timing. It gets harder to see what happens after that, right? How does this emergent technology evolve in a 50-year period?

If you'd said... I mentioned that the computer that I used, which was 40 years ago, is a 100 million times slower than the one that is in its place today. It's very difficult to imagine 50 years from now what that could be. Eventually, you have to assume that the brain is a biological moment, and eventually, the brain will be much, much better understood and that there'll be greater connection between the brain and these digital systems. Exactly how that will be achieved is a matter of science fiction. Over there, yes.

Hi, Marki, MB1. Somehow related to the previous question, what do you believe are the biggest megatrends right now? And especially, what do you believe we should focus on, like on which trends we should focus on, especially if we want to start our own startup and get on the tailwind of those megatrends? Thank you.

If I were a computer scientist at your age, I would work on systems AI because that's, I think, where all the real power is going to be, which is what I've been talking about. I'll give you an example in AI: one of the problems with the models that have been generated is they can't figure out exactly how they work. They work, but they don't exactly understand. And so there's research now under what is called attributive models, where they can actually attribute a decision that was made to how the model learned.

Because the systems are about training, one of the bizarre things about these systems is that you don't program them. So if they come up with the wrong answer, you can't just go in and change a little bit of code; you have to retrain it, and that's an important example.

My favorite example there is: you're in the self-driving car, and the self-driving car has been taught by the firm, and it learned what humans do, which is it did sort of a rolling stop at the stop sign. And the policeman sees you and pulls you over. And the policeman says, "You didn't stop at the stop sign." And you say, correctly, "Officer, I was not driving. It's not my fault." And the officer says, "Who was driving the car?" And the car says, "I was driving." And the officer says, "Why did you run the stop sign?" And the car says, "I don't know." And the officer says, "Who do I arrest now?" And every player is truthful in that scenario. These problems will get solved.

If I were not a computer scientist, I'd be a biologist, because taking the analog world of biology to the digital world means answers to things that have beeviled humanity for a thousand years. And so that says you can make money in those two areas, literally in the next five years, and the rate of investment is at a scale I've not seen in a decade, maybe 20 years. Literally, we're taking mathematicians and physicists who are brilliant and otherwise unemployed and retraining them on this stuff. That's how powerful it really is. More questions? Yes, sir.

I'm Paul Needum, MBA 2. To pick up on what you talked about in your comments earlier with the question of really big companies, obviously we've all read in the papers recently about monopoly questions and antitrust questions. And obviously, you don't think Google is a monopoly, but when would you start to wonder about any company, if antitrust intervention is necessary? And like, what are the signs you would be looking for where you would worry about monopoly power as a general matter?

I obviously know a lot about this because I fought Microsoft, and I've also testified on this and spent quite a bit of time with the European government with respect to the Google activities. The US answer, which came out of the Rockefeller age 100 years ago, the standard is not set, but basically, whenever you're doing things which are anti-competitive, which do not allow entrants to enter the market.

Certainly, when I dealt with Microsoft, the problem was Microsoft owned the channel, owned the distribution, and they were the only place where the apps could be distributed, whereas the internet has essentially eliminated that control over distribution. So the good news about the internet—and again, we just published a version of a book, our latest book, *How Google Works*, which talks a lot about this—is that the internet is so competitive that your product really does have to be good. And if it's really good, you can raise the money, you can get the distribution. I just don't see the barriers to entry, at least on the internet side, at least for a while.

And I think what I would say is, if you look today, there are, depending on how you count, four or five companies that are defining most of this. They are Apple, Google, Amazon, Microsoft, and Facebook. And that's roughly the order of the top five corporations in the American stock market in valuation.

So we're up to something, right, in the sense that this is working. They're global businesses. The competition is brutal, and you get the benefit of that. So the reason that your iPhone or Android phone is so incredibly powerful is not because we're so wonderful; it's because the competition is the most brutal competition I've seen since the PC industry, maybe even more so. So that'd be sort of the way I would answer it. Yes, sir.

Hi, N. Gardenswartz, MBA 1. Thanks so much for coming. Would you be willing to comment on the relationship, as you see it, between the science and technology communities on the one hand, and government and politics on the other? And what are some ways you might envision the two different parties, perhaps, engaging more constructively with each other going forward?

Well, it's not getting better, it's getting worse. A little bit of the history is that the NIH funding, which funds the research for medical stuff in the United States, peaked around 2008–2009. It fell roughly 30% in real dollars between then and the end of the Obama administration. Non-medical research has been roughly flat as a percentage.

And now the new Trump proposal is an evisceration of those numbers and a significant reduction. Furthermore, at NIH, why you would want to—speaking as a person with some self-interest in my own health—why you would want to decrease the amount of investment in health research when you just have invented the technologies like CRISPR that will cause the breakthroughs that will solve these problems that have existed for thousands of years in humanity, is beyond me, right? The only policy stupider is the H-1B policy, which is another discussion, right? I don't need to tell you about that one.

So there's some weird mix of nativism, lack of understanding, scientific bias. So I'm not sure how to solve this problem, but I would go through some things like this. First, science is true, right? And while there are debates within scientists, the fact of the matter is the fact that you have to prove stuff and that you're constantly being falsified keeps it roughly right.

Climate change, by the way, is also true, right? It's actually happening, right? It's a fact. You can choose to ignore it—a reasonable strategy for the Luddites, I got it, okay—but don't deny it, and don't be a denier of basic facts.

I mean, I love these people who say, "Well, you know, we're operating in a matrix." Okay, well, good. Okay, take a knife, put your hand out, and put the knife into your hand and tell me if it hurts. Okay? Does it matter whether you're in a matrix or not? It still hurts, right? So let's deal with some facts.

So, there's a community of people, including myself, who are working hard to try to get these messages through to people who spend their time denying science for political or material reasons. But a simple rule is that often special interests in our system are paying them to believe falsehoods, right? That's a bad way to run a country; it's a bad way to run it forever. It's not a new problem, and it's gotten worse, so we need to fight it, and we need to fight it hard.

There are proposals to cut—I can go on, I'll stop—there are proposals to cut basic research funding that goes to universities, including to this university, right? Think about the value of the product that is produced by this university. You shouldn't cut it, you should increase it. More questions? Yes.

Hi, Lam, MSX, originally from Israel. I wanted to get your take on: how do we close the gender gap in the business and tech world?

Silicon Valley has worked hard on this, and there were, if you go through this, like three phases. So the first phase is the traditional prejudices, sort of male-only world, and again, watch a movie from 50 years ago, you'll see all those patterns. I think those have largely been addressed.

The second one was trying to get employment systems that would work for men as well as women, and that especially involves focusing on childcare, equality of promotion, that sort of stuff. And I think recently we figured out even to deal with some of the pay gap issues.

One of the secrets of what we do, a story from 10 years ago, is that when we interviewed women—I'll give you an example: pay gap. We have studied this exhaustively, and the way you solve the pay gap, since women come in with historical salaries lower than the men, is don't base your salaries based on the historical salary. And if you follow that as a basic principle, you will solve the compensation issue, which we indeed believe we have done, and we've issued lots of reports on that.

The unconscious bias thing, which everybody's focusing on now, is very real. If you talk to women, they'll say, "I feel like I'm not heard, I feel like I'm interrupted, I feel like I'm talked over." And corporations are now beginning to include training. There's no evidence that the training is working yet, but maybe people have to go through multiple such trainings.

But I would say to you that it's smart to try to have every conceivably talented person in your company, including all the men, all the women, all the gay people, all the lesbians, all the Russians, all the crazy people, all the people who are slightly insufferable, right, that are kind of annoying to you. If they really are the best in the world, you need them in your company, which is why the H-1B policy is so stupid, and why the tech industry is so unified on this question about both equality but also the egalitarian and openness that we aspire to.

What I've discovered, which I've been shocked by—I'm always shocked by this, I've been doing this for so long—is that American firms are mostly not like the tech firms. I can't speak for Israel, although my wife and I did a fund to promote female scientists in Israel to try to get them the funding for that crucial period when they graduate and before they get tenure, so we're trying experiments like that. I think this is a solvable problem. More questions? Over here, yes. Yes, ma'am, excuse me.

Hi, I'm Kelsey Page. I'm actually an undergrad. You talked a little bit about falsehoods earlier, and my question is, especially considering the recent election and Google having so much influence over the information people see on the internet and how they get that information, what do you see as Google's role in maybe social responsibility in this realm? And could you elaborate more on how you see Google fighting falsehoods?

So, there's sort of two areas that we're focused on. The first is improving our ranking. So I want to say to you right up front that Google is full of a number of false statements that are not on the first page, right? Our job is to rank, and our algorithms are pretty good at getting the truly idiotic, crazy people who have no life and ranking them below, right below the sort of more legitimate debates. And I'm very proud of that, and that only gets better in my view.

The second is that we have a team called Jigsaw, which is working very hard to use advances in language and understanding to look for threatening speech and hate speech. Which is, if you're familiar with Gresham's law, that bad speech drives out good speech. A lot of these communities are full of people who just pile on in very distressing ways.

And it's gotten worse because of automation. So, an example is—I'll give you a stereotype—you have a woman who is very thoughtful, and she issues an impassioned statement on social media, and she is viciously attacked in personal, sexist, threatened in ways that are just terrifying. And so she has lunch with her girlfriend after this, and she says, "Well, frankly, I just didn't know there were that many evil men in the world."

And her girlfriend correctly says, "Actually, they weren't men; they were bots." Right? That those attacks, that sort of relentless attack that you got, the suppression of your speech, was organized by evil people who were using automation to do so. I think we can detect that with some work.

So I don't think we're ever going to be able to detect a legitimate group over here versus a legitimate group over here that has a disagreement, right? That's a matter for God and judges and politics and so forth to sort out. But if it's a bad person over here who writes something, and this group takes it, automates it, makes it look like a huge phenomenon, and a press person over here writes it as a huge new phenomenon, which serves the interest of no one—I think we can detect that. And that would be a big improvement, if we manage to pull it off, which I hope we will as an industry.

Now, it will arrive just in time, as the evil side will become more automated. So if you look at the Russian attack on the election, which appears to have been done by humans, not by computers (although we can debate that), imagine four years from now, with the power of the tools that I'm describing, weaponizing that, right? And making your voting participation, your confidence in the process, less so, so you're less likely to vote, less likely to express your opinions as a citizen. That's not a good thing. Yes, sir.

Hey, my name is Ben. Thanks for coming to speak today. You had mentioned earlier some comments around focusing in your career, and related to that, I'd like to ask how you think about your role now and your role in the future in ensuring that technology is human-centered, instead of trying to create sort of behavior change around humans that's technology-centered?

I can talk about my personal thoughts on this. I think that the businesses that tech represents are all working very hard to earn your trust and earn your attention and keep you involved, and to monetize that in some way, right? That's why everybody's addicted to all this stuff. So I think that will continue in one form or another.

If you imagine the following scenario: imagine we're so connected that we're all using these knowledge networks that are so smart that they help us every day, which is part of our goal. There's also a possibility that those knowledge networks can enforce a certain uniformity of view, or a common view, right? In the same role that the three networks had when I was a young boy growing up with television.

And our principles need to be independent views, independent access, independent voices, and not suppression of it. And I worry that this adoption of these platforms could lead to a commonality of view rather than looking for divergent and listening for the curious new voices. I worry a lot about that because I think the logic around the platforms means we want everyone in one place. By definition, any such platform has its own language and its own structure, and humans are more creative than that.

And the answer, of course, is not regulation, but more competition, more startups, more ways of expressing yourself. So to me, this next generation of AI and knowledge systems will take advantage of this and produce new, amazing insights, and will provide some competition for the incumbents, including for us. Yes, sir.

Hi, James P. Nice to see you again. How are you? So, another article I read this weekend was about yourself and how Larry and Sergey, when they would bring you on, really cared about someone who could foster group flow in an organization, so they took you to Burning Man to see whether you had it—

That is an excellent article. I suggest you not believe everything you read online. I'm pleased and disappointed here; it's fake news.

But my question was really: when we're making important people decisions in our life, whether it's a senior hire, backing a founder, choosing the romantic partner to spend your life with, what capabilities do you look for, and where would you take them to see if they've got it?

Well, I strongly suggest you go to Burning Man, then. So I had been going before, and they had also gone, and so we found a kindred spirit there, and that's how it really happened. But I enjoy the false narrative as much as anyone else.

There are people who have extraordinary skill in picking other people. And I would suggest that that's something the Dean and I were talking about, things that could be added to the curriculum and so forth. And the judgment of, "This is the person I'm going to invest with," is not a simple one, right? So you need to be thorough, but you need to be excited. You need to sort of fall in love as well as be practical—just all of those things. I don't think that's a solved problem.

I'll tell you what I do, and what we did at Google. First, we wanted people who did more than one thing, so we wanted them to be curious and we wanted them to be accomplished in something. So Sergey's position was we should hire every person who was a rocket scientist we could find because they were clearly oddballs. And any and all of our salespeople should be, I don't know, Olympic winners.

And I said, "Okay, well, that's an interesting approach to sales." And he argued correctly that somebody who can win the Olympics has enormous discipline; think about what it takes to do what they did. He was very impressed with them.

So what I learned was, when you interviewed people, find out what they do that's not work-related, and see if it's compelling, if it's insightful. "Tell me something I don't already know" usually produces a very good candidate, and it's a pretty hard test. If you develop that skill, you'll do great in personal as well as in life. More questions? Oh, there was one over here, this gentleman. There's one here, here, and there's a lady in the back.

Good afternoon, Sorab Singh, MSX. You yourself just mentioned that Android has had a big impact on smart cell phones penetrating into the world, and the big aspect of that was Android being distributed almost for free—

Free is always a good starting price.

Great. Going to autonomous technology on four-wheelers, is that your intention going there as well, or how do you see that percolating into the world? And how do you envision that different industrial structure, different strategy?

We haven't really determined exactly how to do it. We're just trying to get everything to work. So in phones, the story is pretty straightforward: there were a number of initiatives underway. Apple's initiative, very strong player, clearly not licensable by its competitors, right? When you call Apple, they wouldn't return the phone to license the technology—literally the phone, sorry for the pun.

And so we were there with a solution that had the right price, free, and worked. And it didn't have the encumbrances that the other choices did, and that's produced a platform that is well more than a billion users. Most of the studies indicate that a vast majority of all the smartphones in the world are using Android, and we've now taken Android into other markets, including television and cars and things like that. I don't know how far we'll go into the other markets.

Eric has to leave us now, so if you have any more questions—well, we have, I think, the lady in the back has a question.

Even better, right there. Why don't you have the last question?

Thank you. My name is Tatiana Nunes, and I'm from Mexico. We are very excited about the Calico research that you are developing to help enlarge the lifespan of humans. And I'm more interested than you are, because I'm older! So that's very exciting, and our question is: is there already a business strategy around that, and how are you planning to generate revenue from this?

The answer is not yet. Calico—for those of you who haven't been following it—Google has two large initiatives in the health area. One is called Verily, which is not Calico. Verily is about out today, anyway, about medical devices.

We did a fantastic contact lens that could help you with blood glucose, which is a problem that affects diabetics and others. They've since announced many new products in the sort of... think of it as medical device, medical instrumentation, and sort of everybody wants to partner with Verily. They're doing very well.

Calico is more today of a research project trying to understand the underlying processes of aging. I've been briefed on it, but I could not accurately describe the science that they're doing even if I wanted to. But they have the labs, they're looking at the various chemical and biological processes—can they be slowed or affected by treatments, new drugs, that kind of thing?

If it works, which we'll know some years from now, it's a classic biotech. The economics would be incredible, right? Because this would be like a brand new drug that everyone would want to take. Who doesn't want to live longer? And if it doesn't work, well, we'll have tried valiantly, and we'll have advanced the science and the research in that area in a material way.

We recommend catching Eric at Burning Man at the end of August! But otherwise, thank you so much, Eric, for coming.

Thank you all. Please have a great spring, okay? To you.