>>BENNET: Hello, everybody. Am I–can you
>>BENNET: My name is James Bennet. I’m the editor of The Atlantic. I have the privilege
today of interviewing Eric Schmidt. Mr. Schmidt, why don’t you come up here? These good people
need more information about your background. I think they know how to get it.
>>SCHMIDT: Yeah, something like that.>>BENNET: He is the chairman and CEO of Google.
>>SCHMIDT: Thanks, guys. Hi.>>BENNET: We’ve heard a fair amount at this
conference over the last couple of days about what Washington doesn’t get about what’s going
in the country and going on in the world. Among your roles, you sit on the president’s
counsel of advisors on Science and Technology, so you’re in Washington now, I think, every
couple of months from Silicon Valley. What do you think Washington doesn’t understand?
>>SCHMIDT: I was going to–I would invert the question by saying the average America
doesn’t realize how much the laws are written by lobbyists and that is–and I’m not trying
to offend anyone who might be of that like here–and it’s shocking now, having spent
a fair amount of time inside the system, how the system actually works. And it’s obvious
that if the system is organized around incumbencies writing the laws, the incumbencies will benefit
from the laws that are written.>>BENNET: And what is the effect that that
has on the economy?>>SCHMIDT: Well, it perpetuates the current
incumbents, whoever they are. And they are the classic power structures here in Washington.
And it’s shocking to me how hard it is to take on any incumbency in Washington or probably
at the state level as well.>>BENNET: But Google is now one of the greatest
incumbent corporations in America.>>SCHMIDT: Well, perhaps, but we don’t write
the laws.>>BENNET: But you do have your own lobbying
operation?>>SCHMIDT: Yes, one of thousands.
>>BENNET: Well, how would you restructure the systems so that it didn’t–I mean, if
Washington is, in one way, not just politically but in commercial terms, is essentially an
incumbent protection machine, how would you–what would you do to change that?
>>SCHMIDT: Well, one of the things about technology is that technology is fundamentally
disruptive. And my experience now, I’ve done this for a long time, is people are always
shocked at how real disruption occurs and how much change can occur through empowerment,
new technologies would in both good and bad, hopefully mostly good. And we’re at a point
now in technology where we really can change the entire political discourse if we want
to. A typical example would be that everybody here has a mobile phone, right? In fact, you
probably have more than one. The fact of the matter is that in Washington–and again I’ve
been part of this for years–people write lots of reports about things, but they never
actually test them. But with the mobile phone, you could actually just ask everyone. You
could measure everything. And you might be surprised as to what people actually do versus
what they say they do. It’s sort of one of the first rules of the Internet. And so, it
seems to me that you could completely change the way the government works and the way the
discussion works. Another example would be that there are large numbers of very, very
well-meaning groups that do analysis that produce lots of reports that people don’t
actually read. But we can measure whether they’re read and you could use that as a scoring
algorithm for whether you should listen to them or not. You could ultimately decide that
people don’t read anything important or you could try to figure out what is important
that they’re actually reading. I’m not trying to pass judgment; I’m just trying to say you
could organize the system very differently about what is actually happening. Another
example, I was part of the proposing, in fact lobbied there, for the ARRA, the Recovery
Act, which I think was very important at that time. It could have easily been written in
a different way. It could have been written as, “We’re going to give a whole bunch of
money and then every month, we’re going to iterate based on the best impact that it’s
had against these metrics.” It’s what business people think. That’s not how the law was written,
because the law was written for the incumbencies that help write the law, 2,000 pages long.
So, again, if you think about the way the–if you’re–if the government’s job is primarily–aside
from defense–is allocating the–allocating taxes and basically spreading money around
and for good–for good deeds, why don’t we got to a more iterative and measured approach
where we–where we take a look at what’s really happening as opposed to what everybody thinks
is happening.>>BENNET: You’ve talked a bit about what
you see as weaknesses in, in really, the American economic system relative to the Chinese system,
at least the way we practice our business. Not just the relationship between government
and business, but the way we’re organized. You said recently we have a problem taking
the ideas in world-class research and turning them to a–into a business.
>>SCHMIDT: Yeah.>>BENNET: Do you think there’s some sign
that the–[INDISTINCT] American model is actually slipping behind the Chinese model?
>>SCHMIDT: Well, two separate data points, each of which, I think, is very important.
The first is that America’s research universities are the envy of the world. We have 18 or 19
out of the top 20 research universities in the United States. There are many reasons
for this. Let’s give some credit for government expenditures. The enormous explosion in spending
in the last 50 years in the Defense Industry post World War II helped create the modern
research university, all of those research funding. So the Internet was created essentially
out of Defense Department funding. So all of the money, all of the investments payoff
now where we have literally 90% or 95% of the top researchers in the world in the United
States. There are errors in our strategy. For example, we have this bizarre policy where
we’ll train people in these universities and then we kick them out because we won’t give
them visas, which makes no sense at all. But those are small details compared to the enormous
achievement of America and top flight universities. Now, let’s talk about China. China I think
can best be understood as a well-run large business. And China has roughly the following
objectives: it wants to maximize its cash flow, becoming the creditor–sorry, the–if
you will, the bank of the world; and the second, it wants to maximize its, both internal demand
as well as export demand. And the entire country seems to be organized around that principle.
So if you look at all of the interesting new businesses that China–again, think of them
as a business and I appreciate that’s not what they really are, so I’m just using this
to make the point–can be understood as a focus of business expansion. So China decided
that renewable energy was important. They now control 80% of the solar panels. Realistically,
we had that choice five or ten years ago, as a country, realistically that would never
have gotten through our political and social climate. Nobody in America would get away
with saying, “What we should do is we should organize a fundamental industrial policy to
go and dominate this section of the–of this part of an important new industry.” And yet
that’s exactly what China did and now those jobs are in China, not here. By the way, all
that technology was invented in America. Similar stories occur in South Korea and in Singapore.
So, my point here is that there is a different industrial model which is creating high-tech,
high and advanced manufacturing jobs, and we are losing those jobs. One more thing to
give you some more data, we’ve been looking at this question of manufacturing because
I strongly believe that America needs a manufacturing base, a [INDISTINCT] advanced manufacturing
base, to [INDISTINCT] high quality jobs, drive innovation, all the things that we all care
about. The vast majority of global companies have their manufacturing now outside of the
United States because they can invent the technology in America and then they can get
the same quality manufacturing for stuff that is invented in American universities in other
countries as well. So it’s a very, very tough competitive environment.
>>BENNET: But is that purely a policy-making problem? Connect those two ideas for me. I
mean, the Chinese are not exactly embracing disruption. That’s not the key to their success;
it’s scale. And actually in some ways the opposite is a long term, consistent, industrial
policy. Is the problem here too much change and unpredictability in the business environment,
you can’t tell what the tax system is going to be year to year? What is it from a policy
making perspective that needs to change to achieve the kind of industrial policy you’re
describing?>>SCHMIDT: Well, I–first place…
>>BENNET: effectiveness rather it’s [INDISTINCT]?>>SCHMIDT: The American business community
disagrees on a lot of the answers to these questions. One thing everybody will agree
on is that they would prefer consistency rather than constant change. One of the, sort of,
jokes is that the best–it sounds very obnoxious, sorry–that one of the best things to do is
to sort of have the government be distracted by other things while business gets its thing
done. Because business takes a longer term–a longer term view than much of the political
process. And there’s some evidence that in the current recession, the lack of long term
capital investment–basically, companies are hoarding cash for all sorts of reasons–is
related to business uncertainty. That’s one of the criticisms of the Obama administration
and probably to some degree, true. It’s probably true in all administrations as well. The Chinese
model can be understood as classic industrial policy around creating new industries. Now,
that is disruptive, but it’s very top down. Again, very unfamiliar to the American model,
the American model is much more entrepreneurial, has created [INDISTINCT] great industries.
And it’s not necessarily that one will dominate the other, but we have to embrace the one
that we’re good at, which is the entrepreneurship one.
>>BENNET: Speaking of the President, you’ve been a supporter of Barack Obama and you served
in his council. Mayor Bloomberg was here yesterday and took a little bit of a shot at Barack
Obama over his relationship with business. He said, “Obama is a Liberal guy, pro-union,
not particularly interested in business.” How do you respond to that?
>>SCHMIDT: Well, my experience with the President has been that he’s very interested in these
things. The criticisms that the business community have, have a lot to do with tone. If you actually
look at the criticisms that the Obama administration did in fact, extraordinarily quickly move
to essentially–not nationalize, but fundamentally take over and control and regulate the financial
system in such a way that provided the capital that business needed to continue operation.
And there’s no question that [INDISTINCT] most people’s minds, that it would have been
much worse. So, my reading of the criticism, speaking as a supporter, is it’s mostly about
tonality and so forth. The fact that people are now saying, “Well, they need more business
executives in the Cabinet.” Would it be better just to have the best people in the Cabinet
and not worry about where they came from? Can this a way–what I think about. It’s about
symbolism and focus, I suspect.>>BENNET: Actually, I’d like to ask you one
specific intervention the–that this administration made in to the economy to assist the auto-makers.
Did you–do you see that as a kind of classic instance of the government protecting an incumbent
rather than allowing a major sector to be disrupted as it was naturally being disrupted
by competitors? Or do think that was a justified move?
>>SCHMIDT: Well, if you review the arguments, a classic capitalist which is–my personal
view is it you would have entered–intervened in none of these things. And that in fact,
you would have allowed the system to work it out and that through properly functioning
bankruptcy laws, the firms would have been able to continue to operate, that there would
have been enough capital and so forth and so on. There were a set of reasons why it
didn’t occur; one is that there was a feeling that the bankruptcy structures–which again,
have since been changed in the law–would not fundamentally have survived, you know,
the firms–there would have been too much damage. And I think there’s a lot of truth
to that. And so, the intervention in the financial markets, AIG for example, which is essentially
an insurance agency, where they go minimum to 80%, General Motors where they’re doing
the majority thing–I think the judge should be now whether the government makes money.
And then we should have a long debate in the future as to whether we ever want to have
that as an incentive. My own objective–my personal view would be that we try to avoid
this at pretty much all cost. And that we go into these ‘too big to fail’ kinds of industries
and we figure out a way to make them not be ‘too big to fail’ because you can’t have capitalism
without failure. So, don’t confuse yourself. If you think it’s a private company and yet
it’s ‘too big to fail’, why is it not a regulated government entity? It makes no sense. You
have inconsistent incentives. In any government, incentives drive everything as well as in
the business world. And I think you can understand how we now spend some time here. You can understand
that the product of Washington is largely driven by the incentives. The politicians
behave based on the incentives that they have to get re-elected. The way television works,
the way politicians, you know, politicking works, we all know that details here. It’s
true with business as well.>>BENNET: Let me ask you a little bit about
your own business. We had–on theatlantic–theatlantic.com’s tech channel, a really interesting interview
yesterday with your chief economist Hal Varian and he said, “Now, you Google things on your
computer of course, and you Google things on your phone, that’s the next stage. And
I believe, people may laugh, but I think there will be an implant.” Which he said would be
created by Google or somebody else within 10 years. And so…
>>SCHMIDT: And did you ask–did you ask him at the time, are you the first volunteer,
Hal?>>BENNET: He seemed eager as to the interview,
where I have to state–to actually acquire one of these.
>>SCHMIDT: Okay. So, they’re ahead of me.>>BENNET: Yeah. Continuously–he did say
he wasn’t sure if anybody inside the company was actually working on this yet. Maybe you
could tell us.>>SCHMIDT: At Google, one is never sure.
We flip–completely bottoms up company [INDISTINCT].>>BENNET: Uhh-uh, Uhh-uh. Is this is a realistic
prospect in 10 years? And do you understand why it kind of creeps people out…?
>>SCHMIDT: Oh, sure.>>BENNET: …a little bit?
>>SCHMIDT: There’s what I call the ‘creepy ‘line’, and the Google policy about a lot
of these things is to get right up to the ‘creepy ‘line’ but not cross it. I would argue
that implanting things in your brain is beyond the creepy line.
>>BENNET: Mine in particular.>>SCHMIDT: Yes. Yes. At least for the moment
until the technology gets better. Let’s talk a little bit about information and search
and how people consume it. I think we’ve–Google, certainly, have been successful in traditional
text search. You know, you type in a query, you get the answer, off you go. That is the
majority of our business, usage, brand, and so forth and so on. And we invest, you know,
gazillions of dollars into making that happen. We just introduced a product called Google
Instant where it literally does the searches as you type. And the calculation of the savings
of a couple of seconds there translates into enormous human productivity when you compound
it over the number of couple billion of searches that we get everyday. So, we’re very proud
of that. The next thing of course, is the same kinds of things in a mobile device. Again,
going back to the principle that everyone here has a mobile device, you have phones,
soon you’ll have tablets. And so far, you carry these things around. And if you think
about those devices, the phones and the tablets and so forth, they are–they know a lot more
about you. They know for example where you are. So, that context could inform a better
question, you know, if you said, “I want a hotdog,” it could say, “Well, you’re in Washington
D.C.,” you know, “Near the museum,” and off it could go. A lot of languages about disambiguation
and the more information we have about you, again with your permission–I need to say
that right up front, you have to choose to do this, otherwise you can just do anonymous
search, they’re just not going to be as accurate–but with your permission, you give us more information–if
you give us information about who some of your friends are, we can probably use some
of that information. Again, with your permission, to improve the quality of your searches. Okay?
So the next thing we can do is we can take pictures, and do the same thing. So, here
are your camera, take a picture, take a picture of the Capitol right there, we can tell you
it is the Capitol, we could also infer that you’re near the Capitol. So again, all of
that is about improving accuracy. So one of the things that eventually happens in that
perceived line of reasoning is we don’t need you to type at all, because we know where
you are, with your permission. We know where you’ve been, with your permission. We can
more or less guess what you’re thinking about. Now is that over the line? That is at–right
over the line?>>BENNET: We should take a vote right here.
>>SCHMIDT: Is that right over the line?>>BENNET: It’s kind of what we [INDISTINCT].
>>SCHMIDT: Okay. Well we were–so maybe our guess is won’t be very good. Maybe we’ll limit
those; you see what I’m saying? So we’ll try to find that, that line to try to help you
understand more about the world around you. But the fact of the matter is because of the
development of phones, and the world’s knowledge based, this enormous explosion of user-generated
and other information, we can provide a lot of services that can help you get through
the day.>>BENNET: And I’ve–I’ve heard you call this
‘augmented humanity’ and describe what’s happening is kind of letting computers do what…
>>SCHMIDT: I should say, by the way, as far as I know, we do not have a medical lab working
on implants.>>BENNET: Thank you. Just for the record.
>>SCHMIDT: Just for the record. As far as I know.
>>BENNET: Okay.>>SCHMIDT: I’ll give you [INDISTINCT]…
>>BENNET: Could you check and see if you actually do?
>>SCHMIDT: I will check after this [INDISTINCT].>>BENNET: They are working on projecting
on to glasses though.>>SCHMIDT: Yeah, there are actually are–there’s
a quite bit of research, funded in America’s universities by the way, of various forms
of near projection–near field projections; so project onto your glasses, project on a
goggles, that kind of thing. I think of them as heads up display kind of approaches. And
I’m not sure you really want to walk through town with these odd glasses on, you know,
with–looking like an air force, you know, jet fighter or something, but I’m sure people
will. I mean, as I was driving by here, I saw people riding their segways, you know,
looking like normal pedestrians.>>BENNET: Augmented humanity, you’ve described
as letting computers do what they do well, letting human beings focus on what they do
well. What is it you think human beings do well?
>>SCHMIDT: Hard question. One way to think about, I think, computers and people is that
the computers are really good at some things that we’re no good at, at all. They remember
billions of things incredibly accurately and none of us, even the most brilliant of all
us, do that. And further more they can search across that in literally milliseconds, and
come up with all sorts of interesting results that even us, even we, with all of our wonderful
reasoning and intuition, can’t do. Now, humans on the other hand have judgment, have intuition,
and have the ability to recognize patterns in ways that computers are unlikely to for
a reasonably long time. In fact, many of the algorithms that we use are in fact computer
trained. What you do is you would sort of–the computer starts off, you sort of program it
in general and you say, “Oh, yes, that looks right,” or “That looks wrong.” many of the
language algorithms, we do automatic translations, start off with those kinds of approaches.
So computers can help, sorry, humans can help computers get better and humans can help computers
get better. And ultimately if you take this forward for some decades, it’s obvious that
the computers will be our assistants, all right? That they will wander around with us,
if you will, they know where we are and they’ll care what we’re doing, they’ll be able to
track all these kinds of things. And this means that, for example, you don’t need to
remember as much as you used to. I was growing up here in Virginia and at that time, in seventh
grade, they ordered me–the teacher told me I had to memorize all 50 county capitals in
the state of Virginia which I dutifully did, and like, why? It bothers me to this day,
at the age of 12, you know, I had to do this. You don’t have to this anymore, you’re never
lonely and you’re never bored now.>>BENNET: Yes. I would actually wish it was,
sometimes.>>SCHMIDT: By the way, there–by the way
the off button is still working.>>BENNET: Yes. Off button.
>>SCHMIDT: There is an off button and you can turn it off.
>>BENNET: But you just described, a moment ago–you said intuition and judgment are left
to humans, but a moment ago you described a system that–in which the computer will
be intuiting what I think as I’m walking down the street…
>>SCHMIDT: That’s not–no…>>BENNET: …oh, I’m just–I’m stretching
words here but I’m trying to figure out where the line is.
>>SCHMIDT: That would be–that would be on the other side–that would be on the other
side of [INDISTINCT].>>BENNET: It’s intuiting, it’s guessing what
I’m interested in…>>SCHMIDT: Yes.
>>BENNET: …and it’s making judgments about what I’m going be interested in, and what
I’m going to want to do next, what I’m going to want to stop at the street and look at.
So, that’s beginning to intrude on what you described as the–as the kind of human sphere
of really unique attributes and I mean as–there are really two strains of thinking, running
through American history about technology; there’s the utopian strain about looking backward,
[INDISTINCT] strain; and the more dystopian strain about–scary, what computers can do.
Does it keep you up nights? Do you worry about that second category? And what’s going to
keep us from that dystopian vision of the future?
>>SCHMIDT: I worry a lot about the ladder simply because there is a history of technology
being misused by evil people for their own purposes. And, if you follow the line of reasoning
about we know where people are, obviously a bad government, you know, could misuse that.
At Google, we’re very, very focused on letting the end user make those decisions, and so
the way we–the way we ultimately choose is we say, “Let the user have control over this
kind of information.” Do we collect it? Do we use it? The off button, and so forth and
so on. Ultimately because the technology is available doesn’t mean you should be forced
to use it nor should you be subjected to it if you don’t want it. The fact of the matter
is that technology is misused by evil people. Cell phones for example are used by criminals,
the Internet is used by criminals, Google I’m sure is as well, and it’s a terrible thing.
And to the degree that we can detect it and prevent it, we’ll do our best. There are probably
ways in which we can also automate–understand–I’m now speculating, we can also, for example,
look at bad behavior and try to modify that using computers. We can probably detect a
little bit of it. You could imagine your phone in that scenario detecting maybe this is a
place you shouldn’t be.>>BENNET: And telling somebody?
>>SCHMIDT: Well–and you could–presumably the bad person would just turn the phone off
at that point.>>BENNET: What have you–and I’m sorry,
this is another vague, philosophical question, but all the information that passes through
Google, everything you are able to see, people uploading 24 hours of YouTube video every
two minutes or something…>>SCHMIDT: One minute now.
>>BENNET: One minute now? That’s probably two billion views a day or whatever it is,
what do you…>>SCHMIDT: Think about the amount of time
being spent on YouTube?>>BENNET: Well, and that’s where I’m going
at. What–in your job, what have you learned about human nature?
>>SCHMIDT: A lot. We all live in–well, first place, a couple of things, the first message
about human nature is that human nature is overwhelmingly positive. And the second is
that all of the people we don’t know, who we either directly or indirectly think are
not at our level, have all the same aspirations that we do. Just because they’re poor or uneducated
or in a different language or “Oh my god, not Americans?” it’s amazing that they all
want exactly the same things, staring with Britney Spears.
>>BENNET: Right.>>SCHMIDT: You know, it’s truly amazing how
global these markets are. And so, to me this message is largely a unifying one, that people
are pretty much the same everywhere. And that means they also have some of the bad things
as well. People want to be entertained and I think we can decry, and especially people
above a certain age, complain about what they view as the ‘Bowling Alone Phenomenon’, you
know, the changes in society. I would argue that those are–those criticisms are ill-informed,
misinformed if you will. That in fact the people that I work with are highly inter-connected,
and they’re highly communicative, and they’re highly social, it’s just in a way that’s different
from people of my generation. And then, indeed they are much better prepared, if you will,
to lead the world than we are. All right? Because they’re much global, they’re much
more current, they’re much quicker, they’re much more current in the technology that’s
going to affect all of us.>>BENNET: Thank you very much Eric Schmidt
for your [INDISTINCT].>>SCHMIDT: Yeah. [END]