23:31
TED2014

Larry Page: Where’s Google going next?

Filmed:

Onstage at TED2014, Charlie Rose interviews Google CEO Larry Page about his far-off vision for the company. It includes aerial bikeways and internet balloons … and then it gets even more interesting, as Page talks through the company’s recent acquisition of Deep Mind, an AI that is learning some surprising things.

- CEO of Google
Larry Page is the CEO and cofounder of Google, making him one of the ruling minds of the web. Full bio

Charlie Rose: So Larry sent me an email
00:13
and he basically said,
00:17
we've got to make sure that
we don't seem like we're
00:18
a couple of middle-aged boring men.
00:22
I said, I'm flattered by that --
00:27
(Laughter) —
00:30
because I'm a bit older,
00:32
and he has a bit more net worth than I do.
00:36
Larry Page: Well, thank you.
00:40
CR: So we'll have a conversation about
00:42
the Internet, and we'll have a conversation Google,
00:45
and we'll have a conversation about search
00:48
and privacy,
00:50
and also about your philosophy
00:51
and a sense of how you've connected the dots
00:52
and how this journey that began
00:55
some time ago
00:57
has such interesting prospects.
00:58
Mainly we want to talk about the future.
01:00
So my first question: Where is Google
01:03
and where is it going?
01:04
LP: Well, this is something we think about a lot,
01:06
and our mission we defined a long time ago
01:08
is to organize the world's information
01:11
and make it universally accessible and useful.
01:14
And people always say,
01:17
is that really what you guys are still doing?
01:19
And I always kind of think about that myself,
01:21
and I'm not quite sure.
01:23
But actually, when I think about search,
01:26
it's such a deep thing for all of us,
01:30
to really understand what you want,
01:32
to understand the world's information,
01:35
and we're still very much in the early stages of that,
01:37
which is totally crazy.
01:40
We've been at it for 15 years already,
01:42
but it's not at all done.
01:45
CR: When it's done, how will it be?
01:48
LP: Well, I guess,
01:51
in thinking about where we're going --
01:54
you know, why is it not done? --
01:56
a lot of it is just computing's kind of a mess.
01:58
You know, your computer
doesn't know where you are,
02:01
it doesn't know what you're doing,
02:03
it doesn't know what you know,
02:05
and a lot we've been trying to do recently
02:06
is just make your devices work,
02:09
make them understand your context.
02:12
Google Now, you know, knows where you are,
02:15
knows what you may need.
02:17
So really having computing
work and understand you
02:19
and understand that information,
02:23
we really haven't done that yet.
02:25
It's still very, very clunky.
02:27
CR: Tell me, when you look at what Google is doing,
02:29
where does Deep Mind fit?
02:31
LP: Yeah, so Deep Mind is a company
02:34
we just acquired recently.
02:36
It's in the U.K.
02:38
First, let me tell you the way we got there,
02:41
which was looking at search
02:44
and really understanding,
02:46
trying to understand everything,
02:48
and also make the computers not clunky
02:50
and really understand you --
02:52
like, voice was really important.
02:54
So what's the state of the art
on speech recognition?
02:56
It's not very good.
02:59
It doesn't really understand you.
03:01
So we started doing machine learning research
03:03
to improve that.
03:05
That helped a lot.
03:06
And we started just looking at things like YouTube.
03:08
Can we understand YouTube?
03:10
But we actually ran machine learning on YouTube
03:12
and it discovered cats, just by itself.
03:15
Now, that's an important concept.
03:19
And we realized there's really something here.
03:21
If we can learn what cats are,
03:24
that must be really important.
03:26
So I think Deep Mind,
03:28
what's really amazing about Deep Mind
03:31
is that it can actually --
03:33
they're learning things in this unsupervised way.
03:35
They started with video games,
03:39
and really just, maybe I can show the video,
03:41
just playing video games,
03:44
and learning how to do that automatically.
03:46
CR: Take a look at the video games
03:48
and how machines are coming to be able
03:50
to do some remarkable things.
03:52
LP: The amazing thing about this
03:55
is this is, I mean, obviously,
03:56
these are old games,
03:58
but the system just sees what you see, the pixels,
03:59
and it has the controls and it has the score,
04:04
and it's learned to play all of these games,
04:06
same program.
04:09
It's learned to play all of these games
04:10
with superhuman performance.
04:12
We've not been able to do things like this
04:14
with computers before.
04:16
And maybe I'll just narrate this one quickly.
04:17
This is boxing, and it figures out it can
04:20
sort of pin the opponent down.
04:23
The computer's on the left,
04:25
and it's just racking up points.
04:27
So imagine if this kind
04:30
of intelligence were thrown at your schedule,
04:32
or your information needs, or things like that.
04:34
We're really just at the beginning of that,
04:39
and that's what I'm really excited about.
04:41
CR: When you look at all that's taken place
04:44
with Deep Mind and the boxing,
04:46
also a part of where we're going
04:49
is artificial intelligence.
04:51
Where are we, when you look at that?
04:54
LP: Well, I think for me,
04:57
this is kind of one of the most exciting things
04:59
I've seen in a long time.
05:00
The guy who started this company, Demis,
05:02
has a neuroscience and a
computer science background.
05:05
He went back to school
05:07
to get his Ph.D. to study the brain.
05:09
And so I think we're seeing a lot of exciting work
05:12
going on that sort of crosses computer science
05:15
and neuroscience
05:18
in terms of really understanding
05:20
what it takes to make something smart
05:22
and do really interesting things.
05:24
CR: But where's the level of it now?
05:26
And how fast do you think we are moving?
05:28
LP: Well, this is the state of the art right now,
05:31
understanding cats on YouTube
05:34
and things like that,
05:36
improving voice recognition.
05:38
We used a lot of machine learning
05:40
to improve things incrementally,
05:42
but I think for me, this example's really exciting,
05:45
because it's one program
05:48
that can do a lot of different things.
05:50
CR: I don't know if we can do this,
05:52
but we've got the image of the cat.
05:53
It would be wonderful to see this.
05:55
This is how machines looked at cats
05:56
and what they came up with.
05:59
Can we see that image?
06:00
LP: Yeah.
CR: There it is. Can you see the cat?
06:01
Designed by machines, seen by machines.
06:03
LP: That's right.
06:05
So this is learned from just watching YouTube.
06:07
And there's no training,
06:09
no notion of a cat,
06:11
but this concept of a cat
06:12
is something important that you would understand,
06:15
and now that the machines can kind of understand.
06:18
Maybe just finishing
06:20
also on the search part,
06:21
it started with search, really understanding
06:24
people's context and their information.
06:27
I did have a video
06:29
I wanted to show quickly on that
06:31
that we actually found.
06:33
(Video) ["Soy, Kenya"]
06:35
Zack Matere: Not long ago,
06:40
I planted a crop of potatoes.
06:42
Then suddenly they started
dying one after the other.
06:45
I checked out the books and
they didn't tell me much.
06:48
So, I went and I did a search.
06:51
["Zack Matere, Farmer"]
06:53
Potato diseases.
06:57
One of the websites told me
07:00
that ants could be the problem.
07:02
It said, sprinkle wood ash over the plants.
07:04
Then after a few days the ants disappeared.
07:06
I got excited about the Internet.
07:08
I have this friend
07:11
who really would like to expand his business.
07:13
So I went with him to the cyber cafe
07:16
and we checked out several sites.
07:20
When I met him next, he was going to put a windmill
07:22
at the local school.
07:25
I felt proud because
07:27
something that wasn't there before
07:29
was suddenly there.
07:31
I realized that not everybody
07:33
can be able to access
07:35
what I was able to access.
07:37
I thought that I need to have an Internet
07:39
that my grandmother can use.
07:40
So I thought about a notice board.
07:42
A simple wooden notice board.
07:45
When I get information on my phone,
07:47
I'm able to post the information
07:49
on the notice board.
07:51
So it's basically like a computer.
07:53
I use the Internet to help people.
07:56
I think I am searching for
08:00
a better life
08:03
for me and my neighbors.
08:05
So many people have access to information,
08:09
but there's no follow-up to that.
08:13
I think the follow-up to that is our knowledge.
08:15
When people have the knowledge,
08:18
they can find solutions
08:19
without having to helped out.
08:21
Information is powerful,
08:23
but it is how we use it that will define us.
08:25
(Applause)
08:30
LP: Now, the amazing thing about that video,
08:34
actually, was we just read about it in the news,
08:37
and we found this gentlemen,
08:38
and made that little clip.
08:41
CR: When I talk to people about you,
08:43
they say to me, people who know you well, say,
08:44
Larry wants to change the world,
08:47
and he believes technology can show the way.
08:49
And that means access to the Internet.
08:53
It has to do with languages.
08:55
It also means how people can get access
08:56
and do things that will affect their community,
08:59
and this is an example.
09:02
LP: Yeah, that's right, and I think for me,
09:04
I have been focusing on access more,
09:08
if we're talking about the future.
09:10
We recently released this Loon Project
09:13
which is using balloons to do it.
09:15
It sounds totally crazy.
09:18
We can show the video here.
09:19
Actually, two out of three people in the world
09:22
don't have good Internet access now.
09:23
We actually think this can really help people
09:26
sort of cost-efficiently.
09:29
CR: It's a balloon.
LP: Yeah, get access to the Internet.
09:31
CR: And why does this balloon give you access
09:34
to the Internet?
09:36
Because there was some interesting things
09:37
you had to do to figure out how
09:39
to make balloons possible,
09:40
they didn't have to be tethered.
09:43
LP: Yeah, and this is a good example of innovation.
09:44
Like, we've been thinking about this idea
09:46
for five years or more
09:49
before we started working on it,
09:51
but it was just really,
09:52
how do we get access points up high, cheaply?
09:54
You normally have to use satellites
09:57
and it takes a long time to launch them.
09:59
But you saw there how easy it is to launch a balloon
10:02
and get it up,
10:04
and actually again, it's the power of the Internet,
10:06
I did a search on it,
10:08
and I found, 30, 40 years ago,
10:10
someone had put up a balloon
10:12
and it had gone around the Earth multiple times.
10:14
And I thought, why can't we do that today?
10:17
And that's how this project got going.
10:20
CR: But are you at the mercy of the wind?
10:22
LP: Yeah, but it turns out,
10:24
we did some weather simulations
10:26
which probably hadn't really been done before,
10:28
and if you control the altitude of the balloons,
10:30
which you can do by pumping air into them
10:32
and other ways,
10:35
you can actually control roughly where they go,
10:37
and so I think we can build a worldwide mesh
10:40
of these balloons that can cover the whole planet.
10:42
CR: Before I talk about the future and transportation,
10:45
where you've been a nerd for a while,
10:47
and this fascination you have with transportation
10:49
and automated cars and bicycles,
10:52
let me talk a bit about what's been the subject here
10:54
earlier with Edward Snowden.
10:55
It is security and privacy.
10:58
You have to have been thinking about that.
11:01
LP: Yeah, absolutely.
11:03
I saw the picture of Sergey with
Edward Snowden yesterday.
11:05
Some of you may have seen it.
11:07
But I think, for me, I guess,
11:10
privacy and security are a really important thing.
11:14
We think about it in terms of both things,
11:17
and I think you can't have privacy without security,
11:19
so let me just talk about security first,
11:22
because you asked about Snowden and all of that,
11:25
and then I'll say a little bit about privacy.
11:27
I think for me, it's tremendously disappointing
11:30
that the government
11:34
secretly did all this stuff and didn't tell us.
11:35
I don't think we can have a democracy
11:37
if we're having to protect you and our users
11:41
from the government
11:44
for stuff that we've never had a conversation about.
11:46
And I don't mean we have to know
11:49
what the particular terrorist attack is they're worried
11:50
about protecting us from,
11:52
but we do need to know
11:54
what the parameters of it is,
11:56
what kind of surveillance the government's
11:58
going to do and how and why,
12:00
and I think we haven't had that conversation.
12:02
So I think the government's actually done
12:05
itself a tremendous disservice
12:07
by doing all that in secret.
12:09
CR: Never coming to Google
12:11
to ask for anything.
12:13
LP: Not Google, but the public.
12:15
I think we need to
have a debate about that,
12:17
or we can't have a functioning democracy.
12:20
It's just not possible.
12:23
So I'm sad that Google's
12:24
in the position of protecting you and our users
12:27
from the government
12:29
doing secret thing that nobody knows about.
12:31
It doesn't make any sense.
12:33
CR: Yeah. And then there's a privacy side of it.
12:35
LP: Yes. The privacy side,
12:38
I think it's -- the world is changing.
12:40
You carry a phone. It knows where you are.
12:42
There's so much more information about you,
12:46
and that's an important thing,
12:49
and it makes sense why people are asking
12:52
difficult questions.
12:54
We spend a lot of time thinking about this
12:56
and what the issues are.
13:00
I'm a little bit --
13:02
I think the main thing that we need to do
13:04
is just provide people choice,
13:05
show them what data's being collected --
13:08
search history, location data.
13:10
We're excited about incognito mode in Chrome,
13:15
and doing that in more ways,
13:18
just giving people more choice
13:20
and more awareness of what's going on.
13:21
I also think it's very easy.
13:25
What I'm worried is that we throw out
13:27
the baby with the bathwater.
13:28
And I look at, on your show, actually,
13:30
I kind of lost my voice,
13:33
and I haven't gotten it back.
13:35
I'm hoping that by talking to you
13:36
I'm going to get it back.
13:38
CR: If I could do anything, I would do that.
13:40
LP: All right. So get out your voodoo doll
13:41
and whatever you need to do.
13:44
But I think, you know what, I look at that,
13:46
I made that public,
13:48
and I got all this information.
13:50
We got a survey done on medical conditions
13:51
with people who have similar issues,
13:54
and I look at medical records, and I say,
13:57
wouldn't it be amazing
14:02
if everyone's medical records were available
14:04
anonymously
14:06
to research doctors?
14:07
And when someone accesses your medical record,
14:10
a research doctor,
14:13
they could see, you could see which doctor
14:15
accessed it and why,
14:17
and you could maybe learn about
14:19
what conditions you have.
14:21
I think if we just did that,
14:22
we'd save 100,000 lives this year.
14:24
CR: Absolutely. Let me go — (Applause)
14:26
LP: So I guess I'm just very worried that
14:29
with Internet privacy,
14:32
we're doing the same thing we're
doing with medical records,
14:34
is we're throwing out the baby with the bathwater,
14:36
and we're not really thinking
14:38
about the tremendous good that can come
14:40
from people sharing information
14:42
with the right people in the right ways.
14:45
CR: And the necessary condition
14:47
that people have to have confidence
14:49
that their information will not be abused.
14:51
LP: Yeah, and I had this problem with my voice stuff.
14:54
I was scared to share it.
14:55
Sergey encouraged me to do that,
14:57
and it was a great thing to do.
14:59
CR: And the response has been overwhelming.
15:01
LP: Yeah, and people are super positive.
15:02
We got thousands and thousands of people
15:04
with similar conditions,
15:07
which there's no data on today.
15:08
So it was a really good thing.
15:11
CR: So talking about the future, what is it about you
15:12
and transportation systems?
15:15
LP: Yeah. I guess I was just frustrated
15:19
with this when I was at college in Michigan.
15:21
I had to get on the bus and take it
15:24
and wait for it.
15:25
And it was cold and snowing.
15:27
I did some research on how much it cost,
15:29
and I just became a bit obsessed
with transportation systems.
15:32
CR: And that began the idea of an automated car.
15:38
LP: Yeah, about 18 years ago I learned about
15:41
people working on automated cars,
15:42
and I became fascinated by that,
15:46
and it takes a while to
get these projects going,
15:47
but I'm super excited about the possibilities of that
15:50
improving the world.
15:55
There's 20 million people or more injured per year.
15:57
It's the leading cause of death
16:01
for people under 34 in the U.S.
16:03
CR: So you're talking about saving lives.
16:05
LP: Yeah, and also saving space
16:07
and making life better.
16:09
Los Angeles is half parking lots and roads,
16:13
half of the area,
16:17
and most cities are not far behind, actually.
16:19
It's just crazy
16:22
that that's what we use our space for.
16:24
CR: And how soon will we be there?
16:25
LP: I think we can be there very, very soon.
16:28
We've driven well over 100,000 miles
16:29
now totally automated.
16:33
I'm super excited about getting that out quickly.
16:37
CR: But it's not only you're
talking about automated cars.
16:41
You also have this idea for bicycles.
16:43
LP: Well at Google, we got this idea
16:45
that we should just provide free bikes to everyone,
16:48
and that's been amazing, most of the trips.
16:51
You see bikes going everywhere,
16:54
and the bikes wear out.
16:56
They're getting used 24 hours a day.
16:57
CR: But you want to put them above the street, too.
16:59
LP: Well I said, how do we get people
17:01
using bikes more?
17:02
CR: We may have a video here.
17:04
LP: Yeah, let's show the video.
17:05
I just got excited about this.
17:07
(Music)
17:10
So this is actually how you might separate
17:16
bikes from cars with minimal cost.
17:18
Anyway, it looks totally crazy,
17:26
but I was actually thinking about our campus,
17:28
working with the Zippies and stuff,
17:30
and just trying to get a lot more bike usage,
17:32
and I was thinking about,
17:35
how do you cost-effectively separate
17:36
the bikes from traffic?
17:39
And I went and searched,
17:40
and this is what I found.
17:42
And we're not actually working on this,
17:43
that particular thing,
17:45
but it gets your imagination going.
17:46
CR: Let me close with this.
17:48
Give me a sense of the philosophy
of your own mind.
17:50
You have this idea of [Google X].
17:52
You don't simply want
17:55
to go in some small, measurable arena of progress.
17:58
LP: Yeah, I think
18:03
many of the things we just
talked about are like that,
18:05
where they're really --
18:07
I almost use the economic concept of additionality,
18:10
which means that you're doing something
18:14
that wouldn't happen unless
you were actually doing it.
18:16
And I think the more you can do things like that,
18:19
the bigger impact you have,
18:22
and that's about doing things
18:24
that people might not think are possible.
18:27
And I've been amazed,
18:31
the more I learn about technology,
18:33
the more I realize I don't know,
18:35
and that's because this technological horizon,
18:37
the thing that you can see to do next,
18:40
the more you learn about technology,
18:43
the more you learn what's possible.
18:45
You learn that the balloons are possible
18:48
because there's some material
that will work for them.
18:50
CR: What's interesting about
you too, though, for me,
18:52
is that, we have lots of people
18:55
who are thinking about the future,
18:56
and they are going and looking
and they're coming back,
18:58
but we never see the implementation.
19:02
I think of somebody you knew
19:04
and read about, Tesla.
19:05
The principle of that for you is what?
19:08
LP: Well, I think invention is not enough.
19:12
If you invent something,
19:14
Tesla invented electric power that we use,
19:15
but he struggled to get it out to people.
19:18
That had to be done by other people.
19:21
It took a long time.
19:23
And I think if we can actually combine both things,
19:24
where we have an innovation and invention focus,
19:28
plus the ability to really -- a company
19:32
that can really commercialize things
19:35
and get them to people
19:37
in a way that's positive for the world
19:38
and to give people hope.
19:40
You know, I'm amazed with the Loon Project
19:42
just how excited people were about that,
19:45
because it gave them hope
19:48
for the two thirds of the world
19:50
that doesn't have Internet right now that's any good.
19:51
CR: Which is a second thing about corporations.
19:54
You are one of those people who believe
19:56
that corporations are an agent of change
19:59
if they are run well.
20:01
LP: Yeah. I'm really dismayed
20:03
most people think companies are basically evil.
20:04
They get a bad rap.
20:08
And I think that's somewhat correct.
20:09
Companies are doing the same incremental thing
20:12
that they did 50 years ago
20:15
or 20 years ago.
20:16
That's not really what we need.
20:18
We need, especially in technology,
20:19
we need revolutionary change,
20:22
not incremental change.
20:24
CR: You once said, actually,
20:25
as I think I've got this about right,
20:26
that you might consider,
20:28
rather than giving your money,
20:30
if you were leaving it to some cause,
20:31
just simply giving it to Elon Musk,
20:35
because you had confidence
20:37
that he would change the future,
20:38
and that you would therefore —
20:40
LP: Yeah, if you want to go Mars,
20:42
he wants to go to Mars,
20:43
to back up humanity,
20:45
that's a worthy goal, but it's a company,
20:47
and it's philanthropical.
20:49
So I think we aim to do kind of similar things.
20:51
And I think, you ask, we have a lot of employees
20:54
at Google who have become pretty wealthy.
20:57
People make a lot of money in technology.
21:00
A lot of people in the room are pretty wealthy.
21:03
You're working because you
want to change the world.
21:05
You want to make it better.
21:07
Why isn't the company that you work for
21:09
worthy not just of your time
21:13
but your money as well?
21:14
I mean, but we don't have a concept of that.
21:17
That's not how we think about companies,
21:18
and I think it's sad,
21:21
because companies are most of our effort.
21:22
They're where most of people's time is,
21:26
where a lot of the money is,
21:28
and so I think I'd like for us to help out
21:30
more than we are.
21:33
CR: When I close conversations with lots of people,
21:34
I always ask this question:
21:35
What state of mind,
21:37
what quality of mind is it
21:39
that has served you best?
21:41
People like Rupert Murdoch have said curiosity,
21:42
and other people in the media have said that.
21:45
Bill Gates and Warren Buffett have said focus.
21:47
What quality of mind,
21:50
as I leave this audience,
21:52
has enabled you to think about the future
21:53
and at the same time
21:57
change the present?
21:58
LP: You know, I think the most important thing --
22:01
I looked at lots of companies
22:02
and why I thought they don't succeed over time.
22:04
We've had a more rapid turnover of companies.
22:07
And I said, what did they fundamentally do wrong?
22:10
What did those companies all do wrong?
22:13
And usually it's just that they missed the future.
22:15
And so I think, for me,
22:18
I just try to focus on that and say,
22:21
what is that future really going to be
22:23
and how do we create it,
22:25
and how do we cause our organization,
22:27
to really focus on that
22:32
and drive that at a really high rate?
22:34
And so that's been curiosity,
22:38
it's been looking at things
22:39
people might not think about,
22:41
working on things that no one else is working on,
22:42
because that's where the additionality really is,
22:45
and be willing to do that,
22:49
to take that risk.
22:50
Look at Android.
22:52
I felt guilty about working on Android
22:53
when it was starting.
22:56
It was a little startup we bought.
22:57
It wasn't really what we were really working on.
22:59
And I felt guilty about spending time on that.
23:02
That was stupid.
23:04
That was the future, right?
23:05
That was a good thing to be working on.
23:07
CR: It is great to see you here.
23:09
It's great to hear from you,
23:10
and a pleasure to sit at this table with you.
23:12
Thanks, Larry.
23:14
LP: Thank you.
23:15
(Applause)
23:17
CR: Larry Page.
23:21

▲Back to top

About the Speaker:

Larry Page - CEO of Google
Larry Page is the CEO and cofounder of Google, making him one of the ruling minds of the web.

Why you should listen

Larry Page and Sergey Brin met in grad school at Stanford in the mid-'90s, and in 1996 started working on a search technology based on a new idea: that relevant results come from context. Their technology analyzed the number of times a given website was linked to by other sites — assuming that the more links, the more relevant the site — and ranked sites accordingly. In 1998, they opened Google in a garage-office in Menlo Park. In 1999 their software left beta and started its steady rise to web domination.

Beyond the company's ubiquitous search, including AdSense/AdWords, Google Maps, Google Earth and the mighty Gmail. In 2011, Page stepped back into his original role of chief executive officer. He now leads Google with high aims and big thinking, and finds time to devote to his projects like Google X, the idea lab for the out-there experiments that keep Google pushing the limits.

More profile about the speaker
Larry Page | Speaker | TED.com