0:00
It's always an exciting time at Stanford.
Search this video
0:06
It's a particularly exciting time because I think this is by far the best time to be graduating, not the worst.
0:13
Let me tell you that change happens when things are hard.
0:20
That the economy and the situation that we're in now will create phenomenal opportunities that are before each and every one of you.
0:31
If you think about it, this is literally the best time to graduate in 50 years, assets are cheap and expectations are set, realistic.
0:42
I've been teaching here in 354 with Peter Wendell, which is a true, literally a treat of my life.
0:51
And I've watched the quality, and the commitment, and the capability of the students, and it gets better every year, which I think is a credit to the dean, a credit to the process, but most importantly a credit to you all.
1:03
So that's why I was happy to come and talk.
1:06
What I wanted to do is talk for a few minutes about the political environment, the business environment, and then really talk about ideas in technology, how innovation occurs, and then take your questions, which are probably much more interesting than my comments.
1:25
If you look at the situation politically first, the genius of the American system is that it corrects.
1:34
And whatever you think of the previous structure, it's just being corrected big time, and I mean big time.
1:43
And in one month, enormous aspects of our government are being redone.
1:50
One way to think of it is it's a reaction, sort of a take the country back from the middle class based on the excesses of the previous sort of 20 years of wealth creation.
2:02
It's obvious that the changes that we're seeing both politically and in the business world are not short term changes.
2:10
We have an opportunity, in my view, to relearn what creates real value, and I would tell you, and hopefully without offending anyone, that real value is created by building real businesses that do new things that matter to everyone at any level of society.
2:29
And one of the things the business school focuses on is a lot of that.
2:34
One of the quotes that became common in the financial circles that I spent a fair amount of time in meetings over the last few years, is that in America, financial innovation precedes financial regulation.
2:51
Okay, I would assert that this quote, which you won't find very high in rankings anymore, is now discredited, and that we're about to go through a very large change in how this works.
3:06
And in thinking about the Obama presidency, and of course, I was a strong supporter in his campaign and have been an advisor.
3:14
The best case scenario looks like the Clinton presidency, where for the next couple of years, President Obama manages to do the things that are necessary with hopefully the right leaders, and the right support from Congress, and so forth, and the right global structure to get everything going.
3:32
And then there is an ascendant, a couple of years, a reelection and so forth.
3:36
The worst case scenario is that the incentives, and the structure, and the political structure that is always in place in every political structure.
3:46
Somehow we don't overcome that, and somehow it does not turn, so that's the bet that he and others are placing.
3:55
The best idea that I've seen so far is that investment in the right areas creates the necessary conditions to create huge new businesses.
4:06
And huge new businesses are how wealth is created, how jobs are created, how export oriented businesses, and so forth are being created, and there are a number in the President's plan.
4:17
But the important point here is that, we somehow forgot that wealth is created by doing new things that are technologically exciting, that scale, and that are global.
4:30
And I hope that as a society and as a country, and hopefully with the help of people like yourself, we can really do that.
4:38
But there's another opportunity for us, and that's to learn a new idea, and the new idea has to do with transparency.
4:48
If you look at a lot of the financial stuff, the problem today is the financial folks who are trying the best they can, don't really know what's going on, even though the computers all know exactly what's going on.
5:01
This is a failure of transparency.
5:04
So how could you imagine a system where you have all of this data but nobody knows why?
5:12
Well, one is, you say, well, we'll let the regulators manage all of that.
5:16
Well, what's your defense against regulators who are not funded well enough, or maybe not competent?
5:22
What's your defense against the self interest that occurs in all governments, Republican, Democratic, whatever country and party you care about, the answer is transparency.
5:35
If we had seen, if we'd seen the details that are now coming out, there's a much greater chance that responsible people working independently, either as critics, or bloggers, or whatever structure, the system would have corrected quicker.
5:50
And we wouldn't be going through the kind of crap that everybody has to go through now, which is causing some real harm.
5:58
So the new idea is that transparency is the best defense against bad decisions.
6:04
And if I can send one sort of non-technical message, if you can talk about transparency, that we're better off when we know what everybody is doing, at least with respect to how the government works, and how our incentives work.
6:17
So in that context, what's happening is, businesses, and I mean, this in a general statement, I've said publicly for a while, the business situation is quite dire.
6:28
Warren Buffet said this yesterday, other people have said it as well, I'm certainly not the only one who is saying this.
6:33
The fact of the matter is that, we're not at a bottom yet, that we've never faced at least certainly since the depression, a coordinated, if you will, collapse, or deceleration of markets of the rate we've seen.
6:46
And at Google, of course, we see this because we talk to our customers, and the ones we don't have any particularly new insights that you haven't already seen, things like, travel, and real estate, and so forth.
6:57
So our queries shift from buying a house to refinancing a house, to hiring a mortgage broker to hiring a lawyer, right?
7:06
That's the cycle that people go, and it's all part of what you would normally expect.
7:12
But what's interesting now is that, and you have to have a plan, it's reasonable to expect that the recovery that I had hoped, and I'm an optimist in these things, would begin this year is probably not until 2010.
7:25
So that means some number more of quarters of tough business conditions which all of us are going to have to participate in, and we need to be honest about it.
7:35
But maybe we can apply the newly learned lesson of transparency to get the data out, so that people can really figure out when this thing finally turns.
7:45
Recessions and tough times force these decisions.
7:51
Change comes during the recession, not in a period of overwhelming abundance. And so the.
8:00
Good news to the degree there is one is now is when real change will occur, whether it's investment in technology, investment in culture, investment in people, and the reform of institutions.
8:13
So talking now about technology, and how all this plays in, what are some examples of things that could use some change?
8:21
What are some things that have been around for a long time?
8:25
How about computer architectures, the personal computer architecture?
8:29
All of the ways in which, at least that I grew up with, the biggest phenomena in my world is the transition to something called cloud computing.
8:38
In cloud computing, what happens is you have computers that are managed in servers by Google and others, and you pick up any device you want, whether it's your mobile phone, PC or a Mac or what have you, and all that information is stored on those centralized servers, makes sense, right?
8:55
Well, I would like to think that people are moving to cloud computing because of the brilliance of our architecture, we did a great job, people love it, it has all these new features, another possible explanation is it's cheaper.
9:07
So sometimes change occurs just because the model and technology is more scalable, sometimes it's because of a new feature.
9:15
But change is coming and it's coming very quickly, in the case of cloud computing, it's worth doing some math Moore's Law, I think people here tend to know what Moore's Law is.
9:29
It's roughly that the density of semiconductors can double every 18 months, density meaning the number of circuits, if you will, and if you follow that and you do the math, it means that roughly an improvement of a factor of 10 every five years.
9:44
By the way, that means that in 10 years it's a factor of 100, and in 15 years it's a factor of 1000 okay, how big is a factor of 1000?
9:56
Well, a three year old today, when he or she is 18, will have a factor of 1000 more of all of this, faster networks, more data storage and all of that, so what does that mean?
10:12
An example would be that this 18 year old on his or her high school graduation, you hand them an iPod or equivalent device and you can put 85 years of video on this device.
10:26
This person will die before they actually watch all the video on the device you just gave them.
10:33
These are like some serious problems, talk about never being happy.
10:38
>> [LAUGH] >> A terabyte, if you go to Fry's, a terabyte is $100, I remember when a gigabyte was $10,000, and by the way, how big is a terabyte, it's that big.
10:56
That's the size of the Hard drive, so it'll be a petabyte for $1,000, $100 in 15 years, what will you put on it, everything.
11:10
So when you think about the possibility there, there's a possibility that computer scientists talk about an inversion of architecture, that everything we grew up assumed that networks were unreliable, and that computers were reliable.
11:29
But in fact, all the new interesting activities are assuming that networks are reliable because they seem to work now really, really well, and therefore we can build systems which have thousands of computers where a few of them can break and nobody notices.
11:41
We've changed computer science to the inversion of that, what does this mean for you in a business?
11:48
It means that you will end up being able to buy computers that are collections of unreliable parts, but that by operation will become reliable.
11:58
That breakthrough is the breakthrough that Google is based on, because Google was built on PCs.
12:05
The Dean talked about founding it at Stanford, why did Larry and Sergey use PCs?
12:12
Because they were cheap, at the same time, people were using supercomputers and minicomputers to do the same things, and they figured out how to do replication now.
12:24
So cloud computing, it seems obvious if you think about it, pick up any computer, don't worry about what's on it your calendar, your personal information, all your documents are managed centrally.
12:37
What are some of the implications of that?
12:39
Well, it's a huge crisis for software companies that have been depending upon selling client software, and indeed they're all attempting now to move to the web and we'll see how successful it is, obviously including Microsoft.
12:51
And by changing their economic model, they have to come up with new ways of making money.
12:56
But most people believe that eventually most of that kind of software will be free and it will be supported in other ways, and that's another huge disruptive change which many of you are participating in when you look at web services and advertising, supporting services and so forth and so on.
13:12
But the more the most non obvious thing about cloud computing, and it's analogous, the analogy I would offer you is that when television first came out, the first shows were radio shows on television, and then they developed what we think of today as television shows.
13:27
The non obvious thing about cloud computing is once you have all that information on these centralized servers, there are a lot of things that you could do that you couldn't do before.
13:39
My favorite example is flu trends so we, there's a lot of concern about pandemics and the 1918 flu, which is a terrible flu, killed equivalent of 50 million people huge, huge, huge human suffering.
13:56
In Google, we can actually detect anonymously rises in queries because we centralize the servers and then we can alert the healthcare professionals that we think that there's an outbreak of flu in their area and get them and the cycle is about six months earlier.
14:14
This saves somewhere between 10 and 30,000 lives a year because we use that information to get people ahead of the cycle, there are many, many examples, and you all are very smart.
14:27
You can think about what would you do if you could sort of watch, and you have to do it with permission of course, what people are doing.
14:34
Think of it as data mining and you need some level of log retention so you sort of understand at least something of what people have been doing.
14:44
Again, you have to do that within the law and within permission.
14:47
But what's nice about it is if you had all of those tools, and the reason I started with such a harsh statement about the financial system is that all of those systems were architected in the old it way.
14:59
Had they been architected in this new way, we would have seen this coming so much quicker and been able to address it in my view, with so much more certainty.
15:11
So following that lead, when you have all that data, what could you do with it?
15:17
Well, my favorite one is the politician's BS detector, politician is speaking.
15:24
We're all sitting here, you all have your devices.
15:26
You could type in what he or she says and Google or Equivalent will tell you whether it's true or not.
15:31
I mean this is like a real crisis for these guys, you know, I mean at least it's going to improve sort of the, sort of the spin to be a lot more fact based, which in my view would be an improvement. Okay, well that's fine.
15:46
So why don't we have a little fun with that since every one of these devices also knows where it is, can talk to each other.
15:53
Why don't we just have one of you type in what the politician is saying and then everyone can see the ranking of what the person is saying.
16:03
Or maybe you can do it over here, and you can do it over there for a little while, and maybe we can do it together.
16:08
And maybe at the same time we can have a truly intimate social experience in these group settings while we all have the shared experience of a real time experience with a politician.
16:23
So one way to think about that is that for every experience you have, there's a query that you should do.
16:30
And eventually we would like to be able to do that query for you so that you don't have to type it.
16:36
But that's the kind of stuff that you can do when you move to this new model.
16:42
So we're talking about phones.
16:44
Everybody here has a phone.
16:45
I suspect you all have them with you right now.
16:47
Many of you have many phones, BlackBerrys, iPhones, Android phones, and so forth and so on.
16:53
What's interesting about these phones?
16:56
Well, there are more pictures taken on phones than there are in cameras, by the way, every one of which is trapped in the phone forever because you can never get them off. It's a separate issue.
17:05
Every phone now has a GPS in it for 911 services. GPS or GPS equivalent.
17:12
All the high-end handhelds of one kind or another have GPS and so forth.
17:17
And by the way, these cameras are becoming movie cameras.
17:21
Hmm, that's pretty interesting.
17:23
So on an Android phone a couple of months ago, somebody built a product where you open up the phone, and you have it scan where it is, it knows its GPS coordinate, it takes a picture of all the buildings, and then using Google, it tells you what's going on in the buildings.
17:39
Hmm, well, that's pretty interesting.
17:44
So all of a sudden you've got a GPS and a camera and a supercomputer in your hand, backed up by all the world's data.
17:55
So one way to think about that is, okay, well that's pretty interesting.
17:58
Well, why don't we just walk down the street and since my phone is my phone, it's like the one thing that I always have with me.
18:05
It knows a lot about my behavior.
18:07
So for purposes of argument, I love history.
18:12
So as I'm walking around, since it knows where it is, why doesn't it just do a search to tell me the history of every building that I pass, and I'll let you think of the more exciting applications than the one I just gave you, right?
18:26
But the principle, and that's a pretty big deal because not only do you have the device and the access to the information that you have today, but you also have all of the information from everything around you.
18:43
I could do a lot with that information.
18:47
Somebody else built a product where when you walk into a store, you hold up the barcode, and it tells you if it's cheaper online.
18:55
So you can imagine what the shoppers think about that.
18:57
You can imagine the sign, no cell phones, now they really mean it in the store [LAUGH].
19:01
>> [LAUGH] >> So there are about a billion phones coming out in new people's hands in the next three, three and a half years who've never had a phone before.
19:18
We've never heard their voices.
19:21
I don't speak any of their languages.
19:26
We don't know what they're going to say to us, but they're going to to be heard.
19:29
And I think that is fantastic.
19:32
We're going to hear those voices for the first time in our society and in our culture.
19:39
What's another example of something we can do?
19:41
So here we are, and Stanford is founded more than 100 years ago and the model has been forever that we sit here in pretty comfortable seats, and we have textbooks.
19:53
What's another way you could teach?
19:57
You could teach with Wikipedia, couldn't you?
20:01
So I'll give you my example of my friend Bill Joy, who's a venture capitalist and a brilliant computer scientist.
20:05
And I said, well, what's it like being a venture capitalist?
20:08
He says, well, I just use Google. I said, give it a break.
20:11
And he goes, no, no, no, it's what I do.
20:13
I get interested in something like, I said, like what, Bill?
20:16
And he goes, hydrodynamics, okay?
20:18
It's typical Bill interest.
20:20
So he starts, and he does a hydrodynamics query, and he starts reading the research papers.
20:26
And he reads enough of them that he gets interested in a piece of hydrodynamics.
20:30
And then he finds all the papers, and he looks at all the authors that have referenced each other.
20:36
He figures out who the two or three best authors are and he calls them.
20:42
And by the way, these are people who no one ever calls.
20:44
So they call him back, right?
20:45
>> [LAUGH] >> So, and he talks to them, and he says, well, like what's the most interesting thing in your field?
20:54
And every single idea that he's had has come by finding the new idea.
21:04
How could we use that here at Stanford, for example?
21:07
Well, rather than having a textbook, we'll give you a set of queries.
21:11
This is obviously self-serving, but you get the idea.
21:14
We'll give you a set of queries and your job is to do your research before class so you don't have a textbook.
21:21
So now the one third of the class that sits in the back, they're not going to do anything, right?
21:25
Sorry, maybe not at Stanford, but in a normal place you would.
21:29
So they're going to wait for everyone else.
21:30
So what about the two thirds?
21:32
Well, out of that two thirds, some of them will get interested in something, hydrodynamics or whatever it is you care about, and you'll become the world's expert in preparing for a single class.
21:44
And the other people will get it wrong.
21:46
And then you have a classroom conversation and eventually people get to the right answer.
21:51
Think about the modality of teaching is very, very different because you have all the information in one place, and you have all of this access, and it's all tied together in this very, very new way.
22:02
What's interesting now is that user information is becoming authoritative.
22:07
If I gave you a choice between a friend of yours sending you an email saying the following thing occurred on news, or you're waiting for the evening and turning on the evening news, which do you think is the primary way in which you get news now? It's from your friends.
22:21
It's a huge issue for the news industry.
22:23
I'm going to come back to that.
22:25
But the explosion of what is generally known as user to user content or user generated content is not something that's going away.
22:33
And whether it's Facebook, YouTube, all the Digg and Twitter services and so forth, all of those things are really about all of us participating in a broader conversation, user to user, and without the intermediaries. We love it.
22:49
The intermediaries are being disintermediated, and it's a very real crisis for them.
22:56
So here's another example of what you could do. Well, we have Wikipedia.
23:00
And Wikipedia is a phenomenal achievement of mankind.
23:04
It is one of the greatest things I have ever seen.
23:06
It's amazing to me that Jimmy and all those people have managed to keep it so accurate and in the face of so many of the worst sort of human impulses as well as the best.
23:17
Imagine the same thing for doctors.
23:20
We have 16% of American GDP goes to the medical industry.
23:24
Doctors typically are not very current on information.
23:27
They typically aside, some of the best places don't have a lot of information, and they're very, very busy.
23:32
So why is there not the equivalent of a Wikimed where the doctors essentially collaborate to not only talk about the sum of all the treatments, but treatment outcomes and so forth.
23:42
And why is that database not both transparent to themselves, but also to APIs where programs can then suck that data and then do outcomes measurements and things like that?
23:52
It's obvious that such a system should be built.
23:54
The same thing should be true for any technologically sensitive scientific or analytical based field.
24:00
Can you imagine if you had the same thing about the financial world and we had been able to use that to monitor the things that happened over the last five years in the broad-based global financial industry.
24:11
Think about how different we would be in terms of knowing where we were and what we'd have to do.
24:16
So once again, it's the union of all the world's data, the mobile phones and the ability to have users that really I think will drive this thing forward.
24:26
So search, which of course is what Google does first and most, changes, it becomes much more personal because now we know much more about who you are and what you want.
24:39
And this all has to be said with your permission because most people are happy to give you that information, but some people are not.
24:46
So an example would be why is it, my classic example is when I turn on the television, why does the television show me shows that I saw yesterday?
24:57
What a stupid television, it's my television.
25:02
Doesn't it know to delete that show and show me a different show?
25:05
Well, in search it's easy for us to do that, we know what you read, we know who you are, we can show you new stuff, show me what changed, show me what is new today.
25:18
Remember what I looked for yesterday and then suggest something I should look for today, tell me what my friends are looking at again, if they are willing to tell me.
25:32
I suspect most would, but some of my friends probably wouldn't, right?
25:36
But that's okay, we can do that now because of the interconnectedness of all of all of this.
25:42
But maybe in that model, search can then morph at least somewhat into entertain me.
25:48
because ultimately entertainment drives a lot of learning and a lot of outcomes and a lot of activity, just as well as the analytical nature of search today.
25:58
And at the same time, I think we need to take back the web, I don't know about you, but I'm tired of the spam problem, I am tired of these, and I've called it publicly the sewer problem of the internet.
26:10
When we worked, and I was privileged to work on the very beginnings of the internet way back as a very young programmer.
26:17
It never occurred to us that people would come in who were both evil or crazy or both, showing you how naive we all were, and again, that's a separate subject.
26:28
But the fact of the matter is that the web now has this very large component of idiots, right?
26:35
Who basically they really, there's a reason why they're at home in the basement.
26:39
>> [LAUGH] >> And then a term we could invent called disinformation, basically, information that looks legitimate but is in fact manipulative, it's being paid for.
26:54
And it should be possible for us to detect both those categories now with better ranking and better understanding and better user involvement, to really show you the very best parts of the web.
27:05
So in this model, advertising is likely to remain the primary source of revenue for much of the monetization here, of course, Google is heavily in that business, as you all know.
27:18
You can imagine, remember, ads are valuable if they're useful, why does my television?
27:24
It's easy to yell at my television because it's passive, it doesn't talk back, sorry.
27:31
The ads on television repeat and furthermore, they're the same ad, so you show a baby products ad in a home that has no baby in it, it's a complete waste to the advertiser.
27:43
So a targeted ad makes sense, but in this new generation of ads, we can also do dynamic and interactive ads, so as a result, we can do things that are much more narrative.
27:54
And ultimately advertising is about narrative and about compelling people to get excited about something, the technology is now getting strong enough that we can actually do that.
28:04
We're building an ad exchange for display advertisers so that in fact, all those ads that you see you can put in one place and then you can suck the appropriate ad to make that happen.
28:14
And so think of it as the ad for espresso machine, not only should it show you a picture, but but it should also talk about the ad and have a video and have a user content and be able to buy it.
28:25
We think we can make a lot of money, and of course, everybody can make money with this sort of approach.
28:30
So if you imagine that, then think about the integration, the historic separation between enterprise and consumer in computing doesn't make a lot of sense.
28:41
People are the same whether they're at home or at work, they want the same powerful applications at home and at work.
28:49
The separation existed because of security and because of this sort of notion of how the chief information officer operated 20 years ago in client server computing, which is a very different world.
29:00
So with cloud computing, it's now possible to have the same apps at home and at work and still have the same level of security that the corporation depends.
29:08
That becomes the future of enterprise applications.
29:11
That's why software as a service has in fact survived after a lot of initial attacks by a lot of incumbents, and looks like it's going to become the primary way in which enterprises consume applications.
29:24
Partly because of cost, and cost is always a benefit, but also because the benefits of scale and the benefits of sharing.
29:30
When you have in your company, you have everyone in the same place, you get the same benefits, so sharing of calendars, simple things like that, would you say? No, big deal.
29:40
Huge issue in terms of productivity in a corporation, so maybe we should set the goal higher than we already have, and maybe we should set the experience of the Internet as easy as reading a magazine.
30:01
That seems like an odd, I'm not talking about magazines and the Internet are the same thing, but there's a reason why magazines work really well.
30:09
They're beautiful, they're the right size, you can scan them real quick, you can really understand them very quickly, and you can deeply read them at the same time.
30:20
That shift is the next technological shift in speed and structure of the internet.
30:27
And if you think about the speed, feel and beauty you'll get there, Kindle is a good Kindle, and Kindle too are a good first step along that way.
30:36
And although the Kindle lacks many of the features of a magazine, it shows you that there's a path, and again, do the math, do the thousand times math, and you'll see that that will be very much part of our future in some form.
30:50
So what happens to radio, internet and broadcast television, newspapers?
30:57
Well, radio, of course, is now Internet and broadcast at the same time, television is internet, cable and broadcast at the same time.
31:05
And newspapers look like they're going to be pretty much all online over a long enough period of time because of costs, primarily.
31:14
So what does this mean, what's the new thing you can do?
31:16
Well, the most obvious one is you can have live concerts streaming fan videos onto the stage while the performer is playing, with some censorship, I assume.
31:26
Now, why would you do that as a performer?
31:29
Because it would get your fan base going wild, right?
31:33
But every one of those people in that audience has a phone with a GPS that's identified with them and a video camera, and you can transmit it and you can put it on stage. What's another example?
31:47
You could have movie theaters with instantaneous group Twitters where everyone is watching the movie and talking about whether it sucks or not, right?
31:55
Do you think that would drive more attendance?
31:57
You betcha, because we're social creatures.
31:59
This, This whole concept of bowling alone miss the fact that the Internet is a fundamentally social medium and that we finally now, we as society are finally taking it over to drive it in this sort of new social way.
32:14
So what happens then to content?
32:16
Well, if you have something where your audience is going to be 2 billion, you're going to use advertising.
32:21
If you have something where your audience is going to be 2 million, you're probably going to want to use some form of micropayments.
32:27
And the next generation of payment systems are, in fact, will allow micropayments, $0. 01, $0.
32:32
03, as opposed to the current payment systems, which sort of don't really work at that level.
32:38
And then, of course, if your audience of content is 10,000 or 20,000 very highly skilled professionals, you're going to charge them a subscription fee and it's going to be a high one and they're going to pay it, because your content is so good.
32:50
All of those models will evolve.
32:55
What happens to philanthropy with this?
32:58
Well, all of a sudden now you can get all of your donors in one place.
33:03
People here are doing that.
33:04
But you can also imagine systems that optimally combine a web experience and a field operation.
33:13
And the most interesting philanthropy, in my view, will be the building of a platform using the web, using the data, using all the things that I'm talking about.
33:23
Building that platform to do something that then is delivered by people in the field, whether this is medical related, health related, science related, or political freedom related, as examples.
33:38
You can imagine many others.
33:40
Why do I mention this in this context?
33:44
Because we've not really applied the benefits of technology to the NGO and the other worlds at the scale that we should have, but now we can, because the prices are low enough and the hunger is there, and it's all user-to-user.
34:02
Why is all of this so very, very hard? I'll finish up in a sec.
34:07
Why is this so very hard?
34:09
Well, almost everything I've talked about, there's an incumbent who's really struggling.
34:17
Now, as a young scientist, I used to think, well, it's really new technology, everyone will just sort of welcome it.
34:23
But there are many, many powerful forces that oppose the adoption of new technology that solves a new problem in a way that's cheaper, starting usually with the incumbent.
34:33
Often they're very nice people, by the way.
34:36
It has to do with the fact that they're incentivized around an existing business model.
34:41
And under the thesis that change occurs in a crisis, which is sort of what we're going through now, now is when you're going to see these models change literally in the next year.
34:52
That's how quickly, how long does it take for some of these?
34:57
How long will it take before television, in one form or another, will become available as available on your mobile device as it is on your traditional television, five years, ten years?
35:08
I don't know, but it's coming and you can feel it.
35:12
There are startups now that will, for example, watch cable and also watch the Internet at the same time.
35:18
And then give you a selection of shows and then combine a DVR so that you have control over when you view them all perfectly legal and working and licensed in the appropriate ways.
35:30
As people learn this model, the user wins and they get control, and those things will be on those devices.
35:38
I'd like to end by saying that I think the right tone here is one of optimism, not pessimism.
35:48
Everywhere I turn, it gets to the point where you just have to turn the news off every day, because it's a race to the bottom in terms of morale, this crisis and that crisis, and that, and so forth.
36:00
Every leader that I know globally is trying very hard to address this.
36:04
They're working very hard, I'm not criticizing them at all.
36:07
But the fact of the matter is, I think we should view this not as a terrible situation, but rather in an optimistic context, by understanding that it is the nature of innovation that has gotten us these amazing things that I'm talking about.
36:22
And that innovation, which is centered really here at Stanford, right, as the Dean talked about how phenomenal Stanford's contribution has been, it's not stopping.
36:35
The quality of innovation is, if anything, getting better.
36:39
There's a quote about America from Mr.
36:42
De Tocqueville in 1831, and he is a Frenchman who traveled around.
36:51
I can't do it in French, unfortunately, who traveled around America writing about the American culture and spirit.
36:56
And he said that America will do well, because of its abundance of land, the sunny optimism of its people, and its absence of a king.
37:10
Everything I've talked about today is a reflection of that.
37:14
The absence of a king, the change of government, the change of priorities, the optimism of the creative people and the tremendous resources that we have.
37:23
Whether it's in renewable energy, high technology or innovation, or what I consider to be the most important resource, which is the great universities, of which Stanford is one of by far the greatest.
37:35
So with that, thank you very much.
37:36
>> [APPLAUSE] >> Speaker 1: Do you have a question?
37:48
>> [APPLAUSE] >> Speaker 2: Thank you.
37:54
You spoke about your naivete at the start of the Internet, the World Wide Web, and many of the physicists who created nuclear science were dismayed at Hiroshima, Nagasaki.
38:05
So your description of the cell phone as the supercomputer, camera, and GPS system almost makes me want to just destroy my cell phone right now.
38:14
>> Speaker 1: They're cheap, you can get a new one real quick.
38:16
>> Speaker 2: Yeah, get a new one.
38:17
But I'm just wondering, so how do you feel about the fact that you could be creating a super weapon of control and information?
38:24
And how does that dialogue go in the upper echelons of Google and also across in DC?
38:29
>> We talk a lot about this question.
38:31
And the good news about networks and the technology is that it's highly unlikely that the sort of the centralization model of my supercomputer being used for evil purposes, it's highly unlikely that that will occur because the technology and the networks are so decentralized.
38:49
So the issues of misuse, which I think is what you're really getting at, have much more to do with the kind of evil that can be done on the net, whether it's password, whether it's identity theft, which really does have very significant victims, or virus things, or worse.
39:07
We worry at Google that people will publish information that can be used to harm others.
39:12
It is, for example, in the United States legal to publish information that we don't really want you to publish, but it's still legal.
39:22
So those are things that we worry about a lot, and society will adjust if anything really bad happens, I think.
39:27
I'm not as worried about it as others, because I understand the NAT is decentralized. More questions. Yes, sir.
39:35
>> Speaker 1: You talked a lot today about the power of using a person's past behavior or past expressed interest in providing more precise content results, even television entertainment.
39:46
I'm wondering how you think about the ability to still maintain the discovery process some would call the serendipity factor when those things become more precise.
39:56
And I think some would probably argue that serendipity is actually enabled.
40:00
By what we might consider error or fuzziness of results.
40:06
The easiest example for that is newspapers, okay, half of newspapers are stories that you desperately want to read.
40:12
And the other good part of newspapers are the stories that you should have been knowing about, but you didn't and yet you read them and now you love them.
40:20
That's the serendipity part.
40:22
It's relatively easy to bias the systems for one or the other.
40:27
You can have the highly targeted part and then the serendipitous part.
40:32
We have a historic thing called I'm feeling lucky where we do the best we can for you and then see if you like it.
40:38
So there are all sorts of variants of that.
40:39
I don't worry as much about the over targeting because whenever we over target something, there's such an explosion of other choices that our accuracy declines.
40:50
So the good news is there's so much new information coming in that we won't end up over specifying the solution, which is, I think, what you're really getting at.
40:58
Ultimately, the products that I'm talking about will have a serendipity section and a highly targeted section to allow you to make the choice of, do I want to be surprised?
41:06
Or do I want to see what I want to know?
41:09
It's not for us, and I don't think it's for you all, to judge whether people should be serendipitously surprised or see the things that makes them most comfortable.
41:17
The great thing about humanity is they get to make that choice and we'll let them choose it.
41:23
Some more questions, a couple more, go ahead.
41:26
>> Speaker 2: Google is an undisputed leader in search that's known to everyone.
41:31
But can you talk a little bit about competition and particularly Microsoft and their new search engine Kumo? Thank you.
41:40
>> Speaker 1: Well, Microsoft has had a series of announcements of new search engines in the time that I've been at Google.
41:45
And so far they've, what can I say that would be polite?
41:51
I won't comment on their lack of previous success.
41:54
>> [LAUGH] >> Speaker 3: Hi, I'm curious what you view as the significance of a search query in itself.
42:06
What it can tell you about the world around us?
42:08
Is it predictive of behavior?
42:10
Is there some great social meaning?
42:11
Or does it just really depend on whatever search term you're looking at?
42:16
>> Speaker 1: There's a lot of evidence that we can use the aggregate of search terms to predict things, both individual and so forth.
42:23
There's a real issue if we start to use that to try to profile people or things like that.
42:30
So we're real careful not to get too far down the path that you're describing.
42:34
The obvious scenario is you've got somebody searching for breast cancer, there's some probability that either they or a loved one has it, that kind of thing.
42:43
And so we worry a lot about the misuse of that information.
42:46
That's why I was very careful whenever I talked about personalization to say with your permission.
42:51
And I think otherwise, if you don't do it that way, people will move away.
42:55
They won't give you that kind of information, they won't give you their most intimate, if you will, search queries.
43:02
>> Speaker 4: A lot of the future applications of the web that you were talking about depend on- >> Everyone is enjoying your outfit.
43:09
>> [LAUGH] >> You're welcome.
43:12
Depend on the benevolence of people.
43:14
So for example, WikiMed actually requires doctors to take time out of their day, huge opportunity cost to actually donate to this knowledge base.
43:22
So without paying for content, how can you encourage that continued benevolence when presumptively people would look at Google and say they're making profit while I'm not, but I'm contributing.
43:33
>> Speaker 1: I understand the question and I actually just disagree with the reasoning, so if I may.
43:38
Experience is the inverse, experience is that when people see a legitimate activity that will make the world better, the high principle people will take the extra time to make it better.
43:50
I think that's what Wikipedia proves.
43:53
We've seen example after example of completely selfless devotion of information publishing, monitoring, all of that, both in Wikipedia but in other activities as well.
44:05
I have every reason to believe that if you have a legitimate place, in other words, one which is well policed and not full of idiots and that kind of thing, that people will take the time to share their knowledge and experience.
44:17
It's one of the greatest things about humanity is people really do care about making the world a better place.
44:22
And I think you see it, you really do, when you look for the high quality sites.
44:28
>> Speaker 2: Here, I was wondering, you talked about a lot of, sorry, no outfit today, but you talked about how, if there was more transparency, if there was sort of cloud IT in the financial markets, we would have alleviated this problem.
44:46
That politics, education, the voices of billions are going to be heard with better technology.
44:51
And I know that when you have a hammer like Google, you see a lot of nails.
44:55
But to me these problems are a lot more social and the technology can't really fix them.
45:02
Just because some guy in the middle of some rural area has a cell phone, that doesn't mean that the world is hearing his voice.
45:08
It just means he can talk to his friends who are equally marginalized.
45:12
So how do you see technology as just one piece of a puzzle where the other factors are actually much more important?
45:22
>> Speaker 1: Again, with respect, I just disagree with the thesis.
45:25
Let me argue that Ted Turner should be a hero of everyone in this room.
45:29
Because Ted Turner pioneered the use of commercial satellites which created CNN, which make it much less likely for dictators to randomly go kill 100,000 people because they were being watched.
45:43
So there's a lot of evidence that now with cell phones and cameras and so forth, some of the greatest abuses against humanity are less likely to occur because they're more likely to be recorded.
45:55
So that would be a counterexample to your first point.
46:00
The second point is that let's use the example of Iceland, which is a beautiful place with a sorry banking system.
46:08
>> [LAUGH] >> At the point when Iceland's stock market was three-quarters value with three banks, that was too late for them to make the regulatory changes.
46:22
It was just too late for them to make the regulatory changes.
46:25
Had there been much more aggressive transparency with what was going on, had there been much more active social conversation, had there been people in the country and outside who kept saying, hey, hold on, what's going on, and so forth.
46:38
Had there been much more active citizen involvement, the regulators who have since lost their jobs, I should add, would have been under greater pressure to make the changes to slow that phenomena down.
46:50
So I'm not arguing that technology solves these problems because I view these problems as fundamentally political.
46:56
What I am arguing is that technology allows the citizens who are ultimately the victims to put more pressure on the powers that be.
47:06
What's going on, what are you doing about it?
47:08
And I think that transparency, broadly defined is a big deal. Some more questions.
47:15
>> Speaker 2: My question is somewhat related to what you were just saying.
47:18
But in the near future, do you see the Internet apply to government in a way that will allow a more direct democracy by effectively allowing the citizens to be somehow involved in day-to-day decisions of the government and affairs?
47:38
>> Speaker 1: I think it's unlikely the Internet will be used for direct elections, although there are people who are proposing that.
47:47
Excuse me, there have been a couple proposals in the new administration that I think makes a lot of sense.
47:53
The first is everyone knows how the Congress works, is that there's no smoke filled rooms, there's just dark rooms now where everybody gets to.
48:00
Together and they hash out and they make a change in a paragraph, and that change deletes basically an aircraft carrier or adds one, these sorts of things in the negotiation.
48:13
Wouldn't it be better if the markups were on a public board where you could see the changes from the point at which the legislation was first written to its final outcome?
48:26
The only people who would be opposed of this would of course be the politicians, because that's where all the bad stuff happens.
48:33
It's where all the lobbying activities and the earmarks and so forth.
48:36
So that's an example of something that's being talked about right now.
48:39
Another example is that much of the government operates publicly, but without a lot of access because people don't know that they have to go to the conference room.
48:50
So another thing that you could do is you could have a rule that every public room had some form of HD camera and it was streamed onto the Internet.
49:00
So that literally, if you cared about something and you're sitting in whatever state you're in, you could watch the hearing.
49:07
And then you can get excited or get upset about what's really going on in your government.
49:12
I don't think that the tools of communication have been used very effectively by government.
49:17
Andy Grove, who taught here for many, many years and is a hero of mine, said that he's very analytical.
49:24
He said, there's a formula, and I said, what's the formula?
49:26
And then he goes, well, high tech is three times faster than normal business and government is three times worse than normal business in speed.
49:35
So he said it's a factor of nine slower, and that may indeed be part of the problem here. More questions? >> Speaker 1: Yes, sir.
49:45
>> Speaker 2: I think many of us are interested in the clean tech and energy space, and I'm just wondering if you could comment on Google's ambitions.
49:52
And a vision for Power Meter and then anything else in the energy sector that Google's involved with.
49:58
>> Speaker 1: A lot of people here, I think, are aware of the renewable opportunity, the issues of climate change.
50:04
My own view is that that area, aside from nuclear proliferation, those are the two.
50:07
Climate change and nuclear proliferation are the two ways in which we can harm the most number of people unless we act and act responsibly.
50:15
In Google's case, we're spending money to help create the market, so we're investing in startups that do various forms of renewable energy, I think quite successfully.
50:24
We ourselves are large users of energy, and so it's important that we behave responsibly.
50:29
We've been criticized for this, but I would argue, and I think the data will show, that green energy will ultimately be very good business.
50:37
Today it carries this sort of stigma from 20 years of subsidies to the oil and gas industry.
50:42
But in fact, done right and with the kind of technology that's being invented here at Stanford and others, it really will be a great business.
50:49
In Google's case, that's the setup for what Google is doing, which is around the smart grid.
50:53
And simple answer there is that if you look at power distribution, the most expensive power is the peak power, which is in the middle of the afternoon on a summer day.
51:04
So imagine a grid which understood how to do all the necessary shift in power and conservation to marginally change that.
51:12
Literally to use the battery that's in your car to power your house so you use less from the grid and then charge it at night.
51:19
The dynamic power controls that can be done within the home.
51:23
The sum of all of that could very, very significantly reduce the need for power and particular electricity.
51:30
Because the marginal power is the one that's the most polluting, and it's a very big opportunity.
51:35
Another example would be that with respect to power meter, most studies indicate that if you give somebody a power meter, their electricity consumption just goes down ten percent.
51:43
because they have something which tells them what the number is.
51:45
The analogy is that today in your house, you use power.
51:48
It's like going to the supermarket and not knowing the prices of anything you buy until you get to the checkout stand.
51:52
You know, it's time for us to fix this, it's obvious, that's what we're doing. More questions?
51:58
>> Speaker 2: Yeah, I want to allude back to the point that you mentioned earlier about, essentially that the intermediaries and the dissemination of information are going away.
52:05
If you have the cloud, and I have my cell phone and everybody else has their mobile device, they can take the information, capture it, publish it, and receive it.
52:13
This makes news agencies and cable news networks and other intermediate publishers obsolete.
52:20
So my question is, in the kind of medium and near term, what are some of the specific obstacles that you think Google is going to face with regards to that, and how are you going to address those?
52:30
>> Speaker 1: Well, first of all, I don't agree that the intermediaries become obsolete.
52:35
I think the intermediaries have to change their model.
52:38
Much of the media industry is organized around both the creation of content as well as the distribution of content.
52:46
And the distribution of content was built based on the principle of scarcity, in other words, windows, and you can't have it now, you have to wait, we're all familiar with this.
52:56
DVD, movies and so forth, and it's true of all of them.
52:59
And the problem with the Internet is it does not fundamentally believe in scarcity.
53:04
The Internet believes in abundance, it's built around that principle.
53:08
So you have to change your business model so that you have what I call max strategies.
53:12
Where you maximize your distribution, you maximize your premise, you maximize how you make money, and then you use either advertising and so forth.
53:21
So the problem that we're all going through now, and I live this every day, is you have an old model which is based on scarcity.
53:28
And you have the partner or the person suing you, in our case, trying to move to an abundancy model.
53:38
>> Speaker 3: Thank you very much for your talk. >> Sure.
53:40
>> Speaker 3: Much of what you said strikes me as talking about kind of the services that Google provides and I think a lot of the questions are around that.
53:46
But I think the truth of the matter is Google is an advertising company.
53:50
And I've heard a lot said about kind of the expense side of the equation, but not the revenue side.
53:54
And I'm kind of curious about how do you feel about that and if you have more to add about that.
53:59
>> Speaker 1: By revenue, what do you actually mean? You mean.
54:02
>> Speaker 3: Well, I mean all these services that Google provides require some kind of funding so that you can.
54:09
>> Speaker 1: Right, well, first place, we do subsidize free products by our advertising model and search, product search is, it's a gold mine in that sense because it's so highly targeted.
54:21
So the way we like to think about it is that we make the most amount of money with the most targeted advertisement.
54:26
Ultimately, what happens to prices?
54:28
Well, our prices are set by a second price auction.
54:31
So they're set by, and this has all been well studied.
54:35
Those prices hit the substitution point, where eventually those prices are essentially the cost per acquisition that the advertisers are willing to give.
54:47
In our case, we use tools to actually publish that number to the advertiser.
54:52
So from our perspective, it's a closed loop system.
54:55
You know how much you want to spend for a customer, we can give you the math to solve that.
55:00
That number grows and it grows as a combination of both the skill with which we do targeting and also the growth rate at which the internet grows.
55:10
And also it grows based on the shift from offline, which is not so measurable, to online.
55:15
So we live in a complex equation that roughly solves to that.
55:19
The good news is it's a very, very lucrative model when run well.
55:23
So with that, thank you for hosting me and thank you all.
55:27
>> [APPLAUSE] >> Speaker 1: Thank you very much.