table of contents


Vinton Cerf of Google on the future of the internet

Vinton G. Cerf is vice president and chief Internet evangelist for Google, where he has worked since leaving MCI in 2005. At Google, he is responsible for identifying new enabling technologies to support the development of advanced Internet-based products and services from Google. He is also an active public face for Google in the Internet […]

download transcript [vtt]
[ Music ]
This is KZSU Stanford.
Welcome to entitled opinions.
My name is Robert Harrison, and we're coming to you
from the Stanford campus.
[ Music ]
I've said it before.
It makes no difference to me whether my guests are young
or old, obscure or famous, rich or poor, enthusiastic or nihilistic.
They merely have to be able to communicate what they know in a way
that makes it worthwhile for you to listen to them for an hour.
We've had some lofty names on the show in the past few years,
Nobel Laureates, National Book Award, winning authors,
and venerable philosophers, one of whom Richard Royard
sadly passed away, not long after his on-air conversation with me in 2005.
Several of you listeners have written to me to say that one thing you like
about entitled opinions is that it keeps you wondering who's going to be on next.
Well, stay tuned. We have quite a fish on the line for you today.
[ Music ]
[ Music ]
Joining me in the studio today is Vinton Surf.
That's a name that resounds mightily here in Silicon Valley and at Stanford University,
where he earned his undergraduate degree in the '60s,
and where he also taught as an assistant professor in the Department of Computer Science
from 1972 to 1976.
Vinton Surf is known for many things, not the least of which, for being,
as the Washington Post once put it, the person most often called the father of the internet.
Thanks to his seminal work developing the internet's essential communications protocols,
TCP/IP, which he developed here at Stanford in the '70s.
When Google hired Vint Surf away from MCI in 2005 and made him vice president of the company,
Google chief executive Eric Schmidt declared in an interview,
"Vint Surf is one of the most important people alive today."
Anyone who believes that the internet is the greatest revolution in our lifetime
is likely to agree with that assessment.
For a show called "Intitled Opinions,
Vint Surf is an especially appropriate guest, since I can assure you that he has some
eminently entitled opinions about the future of the internet,
and Google's mind-boggling ambitions to organize all of the world's information,
not some, but all,
and to make it universally available to everyone with an internet connection.
In addition to father of the internet and vice president of Google,
Vint Surf has some other intriguing titles as well,
he is member at large of the interplanetary internet special internet group, for example,
and my favorite, he is Google's chief internet evangelist.
Chief internet evangelist is an interesting moniker because the term evangelist comes from two Greek words,
"Eu, good, and angulus, angel or messenger."
An evangelist is a messenger, a bearer of good news.
It is only fitting that the internet should have such a messenger,
for I've always believed that the internet is a distinctly angelic medium,
in the sense that it's a boat is in the heavens, the air, the sky,
and also in the sense that the internet at any given moment
purulates with innumerable messages, communications, pulses, and signals.
How many angels could fit on the head of a pin the medieval scholastics used to ask themselves,
an incalculable number of them they answered?
So too with the internet, every time you send and receive an email or read an article online
or listen to a podcast of entitled opinions,
a staggering number of angels, call them bits or bites,
ferry the messages back and forth across and through the ether.
In the past decade or two, this vast angelic network has brought about
far-reaching irreversible changes in our politics, entertainment, culture, business, healthcare,
the environment, to say nothing of our personal lives, and indeed the very circuitry of our brains.
The magnitude of these transformations is bound to grow only more immense and sublime in the next decade or two.
And the evangelist who joins me today in the studios of KZSU is here to speak to us about what lies ahead
in this brave new world of the internet, and to speak to us also about the decisive role that Google will play in this increasingly unlimited and
jellification of our society.
Mr. Vintenser, welcome to the program.
Well, thanks so much, Robert.
It's a real pleasure to be back on the campus, and particularly here with you on KZSU.
In fact, I wanted to ask you before we speak about the future of the internet, if you would want to reminisce a little about the
years you spent here at Stanford, and whether you remember those years in which you were developing the internet's essential communications protocol as the most exciting of your life, and if you had any idea back then when you were engaged in that project, where this new technology was going to lead.
Well, Robert, let me start out by suggesting that I'm not the only father of the internet at the very least, Robert Kahn deserves a co-father title because it was his idea to begin research into open networking.
He was in the Defense Advanced Research Projects Agency at the time, and came to Stanford to say that he had a problem making the packet switching technology of the then-arpinant,
and two other networks, a packet satellite network and a packet radio network that was operating here in the Bay Area.
He was trying to find a way to make all three of those work together, and that is why we call it the internet problem.
There are many, many others who deserve a great deal of credit, and I won't try to repeat all of them on the show, but your listeners should understand that an undertaking of this scale doesn't happen because one or two people commit themselves to it.
It's because literally millions of people have decided that this is a really interesting idea that they want to devote time and energy to.
And as you see, the results of Tim Berners-Lease work on the worldwide web, this avalanche of information flowing into the net stems from people's desire, I think, to share what they know with other people.
Let me come back to your question about what I remember of my time on the campus.
As an undergraduate, I studied mathematics, and Stanford in the 1960s was a smaller place than it is today, just in terms of physical presence.
And the distribution of women on the campus was quite limited.
So, as a student in the mathematics department, I remember reflecting on how few coeds there were in mathematics, and set about in a very logical manner to solve that problem by enrolling in historic court dance.
There were 24 women in me, that solved that problem, particularly, I appreciated that one.
My love of mathematics was overtaken by an utter mesmerization with computing.
And so I took every course I could while I was on the campus in computer science, and eventually ended up working for IBM after I graduated.
And after two years at IBM, I realized that I didn't actually know enough about the underpinnings of the architecture of networks and especially computers and operating systems, so I went back to school this time at UCLA for a PhD.
So my time as an undergraduate at Stanford is resonant for me because of the literature that we were required to read.
I look back on this now, we fussed about it taking Western Civ and reading all of these classics from Western Europe and from Greek and Roman times.
I look back on that now with great gratitude that I was required to read books that I probably would not readily find time to pick up now.
I mean reading Montesquieu about America is at this very moment extremely relevant given this sort of rebirth of excitement in this country thanks to the election in President-elect Obama.
So I remember that time with a great deal of pleasure although there was plenty of storm on the ground as any undergraduate can tell you.
And when I returned to the campus I actually resisted coming to Stanford to teach.
I was asked to come up to lecture which I did about my dissertation work which didn't have anything to do with networking and turns out.
And after I got back to the UCLA campus where I was working on the Arpanet predecessor to the Internet, I got a phone call around 1972 asking if I would come to Stanford to join the faculty.
And I said no.
And the reason I turned them down is that I knew how smart the students were here and I was terrified that I didn't have anything to teach them.
Well, it was an interesting experience because for the rest of the day the phone was ringing regularly and I was in the computer room down the hall.
And one of my colleagues kept running down the hall saying, "Oh my god, John McCarthy is calling me John had written the textbooks or Don Kluth is calling me."
It was incredible.
And at the end of the day I was a puddle on the floor and I finally said, "Yes, actually Harold Stone was the then recruiter at the campus I think he has since left and gone to Amherst."
So I ended up coming back to Stanford and you mentioned computer science department actually for us basket and I had a split appointment.
Each of us had one foot in computer science and one foot in electrical engineering.
And as our career is unfolded while we were here we shared a secretary, Carolyn Tyneye.
We realized that the likelihood of tenure was diminished by being in two different schools let alone two different departments.
And so we swapped hats.
He stayed in computer science and I went into electrical engineering.
So I actually finished my time as an assistant professor here in the electrical engineering department.
Our president John Hettice was a student at, was he a student at that time?
I'm sitting here thinking that I believe he was but I think so.
I believe he was a student either that or he had come here and maybe become a faculty member but I don't think so.
I think he had not yet joined the faculty. We should go and look this up to the Amgugli.
Right? Exactly.
Well, those are fun recollections.
There's a saying that if you can remember the 60s you weren't really there.
You remember part of it.
So I didn't partake of a lot of that other stuff because I was just too interested in the computing and mathematics and just this incredible world of scientific curiosity.
And I appreciate what you said about the Western Civ and it's now gone through various modifications.
It became culture values and I was partly because it was beginning to pick up Asian history as well.
I think that's very good because we missed a great deal of non-Western thought as a consequence.
Now it's called Introduction to the Humanities and I teach in those tracks.
I'm glad to hear that.
Good. Can I ask one of these naive questions of naive because I'm a typical humanist and I suppose many of my listeners would also want to know what is the difference between the arp and net and the internet?
We might extend your question to what's the difference between those and the worldwide web because many people can use all of them.
The arp and net was an experiment in wide area packet switching.
And to think about packet switching it's not a simple concept.
If you look at the telephone network when you pick up the phone the telephone switches make an electrical path between your phone and the destination phone.
I'm simplifying a little here because it might be over the air or it might be a mobile phone but the whole idea is you have a dedicated connection.
In the packet switching world you don't have a dedicated connection.
You share broadband circuits and what you send over the circuits is roughly speaking an electronic postcard.
And so if your listeners know anything about postcards which they probably know a lot about, they actually know how the internet works because the internet
packets behave just like postcards. When you put a postcard into the postal service there's no guarantee that it comes out the other side. That's true of internet packets.
It's a best effort service. If you put two postcards in they don't necessarily come out in the same order. That's also true of internet packets.
The internet sometimes does things that the post office doesn't do. If you put one internet packet in sometimes you get too loud because they may go on multiple paths.
And more than one will hop out and you have to deal with duplicates.
So it's a best effort design.
In order to make it a more disciplined service you have to put a layer of procedure on top of what we call the internet protocol IP.
You mentioned the initials earlier TCP that stood for transmission control protocol.
And that is a set of procedures between cooperating hosts at the edges of the net.
To understand how that works just imagine that you're sending your friend a book through a postal service that only carries postcards.
And what would you do? Well, let's see. First of all you'd figure out you have to tear the pages of the book out, cut them up and glue them onto postcards.
Then you notice that the page numbers are not on every postcard because you cut the pages up.
So then you realize that these postcards won't get your friend in the right order necessarily.
So you put numbers down 123456 to allow your friend to put them back in order.
Then you realize that some might get lost so you keep copies.
Then you think, well how am I going to know whether I should send any copies?
And you think, well why don't I have my friends send me a postcard?
And once in a while saying I got everything up to postcard number 420.
You can throw your copies away up to that point.
Then you realize that that packet might get lost.
And so then what do I do now?
And you answer as well you kind of time out after you've heard nothing from your friend and you start sending copies of things that he hasn't acknowledged until he finally says I got whatever you were sending.
That's basically how TCP/IP works and that's how the internet works.
Now the ARPANET was a wire line based packet switching network.
The internet asked the question, how can I link different kinds of packet nets together, some which might use radio, especially in mobile environments,
and what others that might use satellite, and eventually of course some using optical fiber or coaxial cable.
So Bob Kahn and I were trying to solve the problem of taking any kind of packet switch net and creating a framework in which they could all enter work,
and in which every computer on every network could communicate with every other computer on every other network in a very uniform way.
And it was the design of the TCP/IP protocols that permitted that to happen.
Fantastic, it's like magic. And you said the worldwide net is a worldwide web.
The worldwide web is not the same as the internet, even though for many people it's the primary application that they use.
But the worldwide web is exactly that.
It's an application that sits on top of the TCP/IP protocols.
There are other applications that run independent of the worldwide web.
Email is an example file transfers or an example streaming audio on video or an example.
Even though they may be initiated by interactions that occur in the worldwide web context, the actual streaming audio on video, which your listeners may be getting right this moment, are being transported by protocols adjacent to the worldwide web.
Got it.
Can we talk a little bit about the future of the internet?
I read a piece that you recently contributed to the 10th anniversary of Google, in which you speculate about what it's going to be like to be.
A citizen who has access to internet connection to walk into a hotel room, for example.
And you imagine that one would have a mobile phone/computer device.
And this device would be so smart at interpreting where exactly you are.
Your room number, you can even program it to tell you to find where you might have left your briefcase and so on and so forth.
Okay, so those are the ones that of course any time scientists speculate about the future, you can bat at least a couple of things.
One is that the one-term projections will be too timid and the near-term projections will be too ambitious.
That's probably true of some in mind as well.
So that's what I think we can imagine over the next decade anyway.
First of all, lots and lots of mobile access to the net.
There are 3.5 billion mobiles in use today and about 15% are internet capable, meaning you can use them to interact with the internet.
Go to the Google homepage, for example, and search for things.
What I'm imagining is that there will be a great deal of intelligence put into the mobiles that we carry around with us.
Imagine when you walk into a hotel room, sitting in that room typically is a high resolution display already.
We think of it as a TV set, but in fact it's the same kind of display that we use on our laptops and desktops.
There's probably also a keyboard sitting there for Web TV, but there's no reason why your mobile couldn't detect the presence of those devices and expand its capability.
So instead of being limited to the display on the mobile, you might actually be able to have the display come up on the television set, or you might be able to escape the limits of the little thumb-based keyboard and use an ordinary Web TV keyboard.
But in addition to that, the hotel rooms themselves could be announcing information, which could be picked up by the mobile or by your laptop for that matter.
Such information is what room are you in and what to address you in.
And so if for some reason that an emergency occurs in you dial 911, the information about where you are could be transmitted to the 911 emergency services operator, that's better than what happens today because when you dial 911, all you get is the physical address, but you don't have any information about where the party is in the hotel.
So we could do better with the internet than we can with the conventional phone system.
I think that we're very likely to see a huge number of sensor systems being attached to the network.
I have an example with this at home. I'm running a 12 sensor network that is collecting temperature and humidity and light levels in 12 rooms of the house, including the wine cellar.
It's very important to us. I need to keep the wine temperature below 60 degrees Fahrenheit and above about 70% humidity.
So the corksnone dry out. So every five minutes the sensors collect this data and they store it away in a little server. And I can reach that server anywhere from anywhere in the world anywhere on the internet.
It's also set to a warm. So in the event that the temperature goes above 60 degrees, I get an SMS text message on my mobile, which happened about four weeks ago when I was at our international laboratory, my mobile went on the wine cellar column.
Your wine is warming up. The cooling system isn't working. So this sort of thing is going to be very common.
And the reason I did it is that I wanted to collect a year's worth of information about the performance of my heating ventilation and air conditioning system.
So I'd see real data about how well the house was being heated and cooled rather than depending on anecdotal information like GITC-POT during the summer where the bedroom was cold last night.
We will see lots more of that sort of thing in the future. And finally, I can confidently say that we will have an interplanetary internet in operation by that time because as of the end of this year, we will have finished the first test of a deep space interplanetary internet extension working with NASA and the Jet Propulsion Laboratory.
And then next year, working with the International Space Station. So the protocols for extending internet operation across the solar system are well underway.
And they should be space qualified by the end of 2010, at which point all the space-faring nations have the option of using these protocols so that their spacecraft will be interoperable.
And the reason that's important is that over time, as spacecraft complete their primary missions, they often have lifetimes that exceed the mission plan, a good example of that.
That is the rovers, which are coming up on five years of service on the surface of Mars, even though their original mission was expected to last 90 days.
So we're hoping that what will happen is that these new protocols are adopted. Then new missions that are launched can use previous mission assets and will kind of accumulate an interplanetary backbone over the next several decades.
Who would be sending messages back from these?
Well, primarily robotic equipment in the near term, but as you know, there is an appetite to return to space.
The Chinese are already launching people into orbit and have hopes to land on the moon again.
We have some expectation of mounting a Mars, manned Mars mission sometime in the next 20 years or so.
So I expect that certainly if not in the next 10, then the next 20 or 30 years, there will be manned as well as robotic missions.
And we believe this new protocol architecture is richer by dramatically richer than point-to-point radio links and will enable some very, very complex missions to be undertaken.
So if you were wherever a Star Trek fan or a family, I am a Trekker from Japan.
I am doing one, yes.
And both the original series as well as Star Trek, the next generation.
I thought the next generation had a lot more intellectual content to it thanks to Patrick Stewart and his colleagues, many of whom are Shakespeareans as I'm sure you know.
Yes. And it always struck me that the future that's envisioned there, three centuries out, three or four centuries out, was very close to its realization in the realm of electronics, which is what I associate the computers
in the beginning for electronics. One can imagine within a decade or two or three, our world can, you've just been describing it. We can walk into a room and there's all these sensors which can, and we can have interactive computers which will respond to our verbal
Google's rather than actual testing of the technology. But that on the level of the space traveled itself, we are just as primitive as we were when Star Trek was first to imagine.
Yes, I'm afraid. So the best we can do right now is to imagine a rocket engine that's capable of getting us up to about one tenth the speed of light.
Well, why is electronics had this huge revolutionary jump forward, whereas in the hardware part of the equation?
Well, I think part of the reason for this is that the hard work, the basis for the electronics we use today came in the 1910 and 1920 period with quantum, the understanding of quantum effects, because that's what, and then the invention of the transistor, which is use those quantum effects, around 1947 to '48.
So we've had, on the order of a hundred years of time, to have a basis for some of the hardware that we use today. On the other hand, we don't have a physics that teaches us how to deal with overcoming speed of light limitations. We simply don't have any theories.
We have string theory with 11 dimensions, but we don't know whether we could travel in the little curled up dimensions instead of the big three-dimensional and four-dimensional space time continues.
So we just don't have a clue right now about how to overcome those limitations. We don't know how to make a wormhole. We don't even know how to find one.
So we're just a long ways away from being able to undertake that part of Star Trek.
But I have to, I'm amazed about it in various, is a good friend. And I have to say that the philosophy behind Star Trek, which she reflects too, and around thinking, is very positive.
And in eight belief that the human race is not self-destructive, no matter what we might, the needless on your show might think, I am still very optimistic about what we can do as a species.
Right. Well, I was going to save this question for later in the program, but I might as well throw it out now that we're talking about Star Trek.
Because in the next generation, there is this enemy, which is the Borg.
And the Borg is very fascinating because I always thought it comes out of the unconscious, or they say the subconscious of a highly technology civilization like the Federation of Plants that the Enterprise belongs to.
And that is a world where the members of that collective are synthetic organic creatures. I mean, they're, again, as organic.
But they have all these prosthesis, which connect them to this collective.
And they are always hearing voices in their heads because they're forever connected.
And the ideology of Star Trek is that no, we humans believe in the absolute value of individuality, thinking for ourselves, and so forth.
Is there a danger that the future of the internet, we could be so interconnected one with the other, and that everywhere we go with our mobile devices, we're always hearing voices in our heads from the collective that the Borg
starts becoming a menace to us.
Well, that's a very dark view of the potential.
I don't know, to be honest, whether I buy that.
Although I have to say, if you've read the singularity is near by Ray Kurzweil, you'll appreciate that his model of the world is that our computing capacity will have increased by around 2029.
And so it is essentially more capable than the human brain.
At which point that device will then begin to design its successor, which will be more capable than the human race.
At which point that successor will design its successor, and it will be more capable than all the human beings that ever lived.
At which point Ray intends if you still align to upload himself into the system, and thereby achieving mortality as would everyone else be able to.
And his thought is that's how we will overcome the space travel limitations we simply will live forever.
And that means we don't mind taking a few hundred thousand years to get from here to the next star.
We live as brains though, or as consciousness not as bodies.
Not his physical body. No, he would upload his consciousness in effect.
I think that's probably a little farther out than I'm prepared to endorse, but it makes for an interesting thought.
Your concern is interesting and slightly flamboyant of course, which is the intention I'm sure.
There are some interesting technologies that get us closer and closer to a more like capability.
I'll give you an example. My wife has two cochlear implants. She was deaf for 50 years.
And she had a cochlear implant. And of course, life transforming because suddenly she can hear her.
She wears a speech processor, which is about the size of a mobile.
The speech processor takes in sound from a microphone. Does a 48 transform to figure out what frequencies are present.
And then it figures out which electrodes in the inner ear in the cochlea to stimulate to emulate what the nerves would have done if they were working properly.
And the brain interprets that as sound. Now, my next chore is to reprogram the speech processor so it can be on the internet.
You're using the TCP/IP protocols and voice over IP. What could we do with that? Well, if she were to voice a question, a Google kind of question, we could take the sound into the speech processor, which it normally would go.
And then digitize it and send it out into the internet to a computer that can understand speech.
And I can tell you that speech understanding computers are becoming increasingly capable.
At which point it could generate a response, text response, which could then be transformed from text to speech.
And then relay back as a series of internet packets, which land in the speech processor, and then are regenerated as electrical pulses in her head.
No one would hear the answer except segued. Now, I haven't done that yet, but the idea that you could effectively have a computer communicating with you in your head and hearing voices in your head that way is not too far away from your board.
I'd like to think that that conversation is a constructive and positive one as opposed to this individual suppressing Borg-like phenomenon.
I don't think that the internet, despite what you're seeing now in the younger generation, tutoring at all hours of the day and night, necessarily translates into Borg-like outcomes.
For one thing, these phenomena, the texting and instant messaging and the like, which the current generation likes, works only because they form a cohort which are in nearby time zones that is saving aren't too far away from each other.
That's because they haven't grown up yet. But as they get older and their friends move more and more apart, they will find that those kinds of real-time interactions don't work because the hours are three in the morning and the other.
And so they will probably adopt other older practices like email because of the time, what we're coming the time delay that these asynchronous communications permit.
So I don't think we're going to be twittering to each other by thinking in our heads anytime soon.
However, if one does develop a computer which is more intelligent than the human brain and where one can access the human brain and access the greater intelligence of such a machine, and here I'm going to be even more provocative.
And quote the unibomber, when he wrote that manifesto, because he said, you know, technology has always begin as optional and then they quickly become necessary.
The question then would arise, will it be an obligation of everyone to upload as your friend's agency?
But anyway, we can...
Well, as long as you're being provocative, I feel like I have to respond.
In some respects, what Google hopes to do is to make virtually all information that we have as a human race available somehow.
We know just from watching what's happened on the World Wide Web that people do want to share their information.
It's just an avalanche of it.
And YouTube, for example, is getting 13 hours of video uploaded per minute.
I have no idea who's watching all that video, but there's an awful lot of it getting into the net.
So in a sense, we're already showing an appetite for getting access to a great deal of information conveniently.
And we're looking for easier and easier ways of expressing our interest in it.
So I expect to see exactly what you were describing that is to say oral interactions with this quantity of information over time.
And I also expect to see collaborative interactions.
Now today, the collaboration takes place between human beings.
We talked to each other.
So in our mind, we are looking at the same documents, editing them together, modifying specialties together and chatting with each other literally orally.
As we do it, and we do this in groups of people that are scattered all over geographically distributed.
It is imaginable that some of those interactions could occur with a computer which is capable of understanding speech and generating speech.
And so the collaboration that we start with from human to human could include some programs that are also a part of the discussion.
And I find that rather intriguing.
It is very intriguing.
And if we can talk about Google's role in this future, Google is a company that is just awesome.
It has completely fearless. And its ambitions are sublime in the most technical definition of the sublime, namely something that transcends any kind of finite limits.
It wants, as you said, to organize the totality of human knowledge.
It has projects which are truly awesome.
Like to scan every book ever put into print, which is for us scholars, it is a beautiful prospect that all these out of print books can be one's exactly the term out of print.
I hope we will become obsolete.
And to it has the biggest photo album ever compiled where every corner of the earth, you can have a picture of it.
And it is increasingly, I think, that it is going to release or has released a mobile phone.
That is a T-Mobile, Android.
Well, Android is the operating system and T-Mobile has simply used this open operating system for their G1 mobile device.
The first one to use the Google Android operating system.
And if that phone develops along the lines that you were suggesting earlier, then it is going to be able to perform all the functions that are laptop.
Oh, yes. I would think so.
Of course, there are things that the laptop does that mobile phones don't do yet, including the larger displays which gives you some context.
I think it is probably important for you to appreciate that the founders of Google are unafraid of significant ambition.
And when they set out to simply download and index the entire internet, it is indicative of their willingness to push the limits.
The choice of the name of the company also is indicative of that because Google, G-O-O-G-O-L, means 10 to the 100th in the language of mathematics.
And the company's name was misrecorded by someone.
I don't know the details of the story, but it was written G-O-O-G-L-E, Google instead of Google.
And it is sort of stuck.
But when Larry and Sergey and Eric Schmidt are interacting with the rest of the company, their natural tendency is to challenge everyone to push much harder than they might naturally have chosen to do.
And so if you set a particular target, there is often they will "why not?" 10 times bigger, "why not?" 100 times bigger or faster.
And I think that is good. It is healthy to be challenged. And that is what makes Google such an exciting place to be.
But is it's ambition ultimately to not only organize all of the knowledge available, but to re-centralize it all in a supercomputer that people are calling the cloud, where everything will pass through a specific network which is owned by Google or which is operated by Google,
and would be this kind of central network in the angelic sphere of the cloud.
Well, what is the cloud?
Okay, so let's see how do I get the "We need some vocabulary here."
The term "cloud" originally I think came from networking because we all used to draw networks as if they were clouds.
Each network was its own cloud and they were interconnected with each other.
Then what has happened at Google and other places as well in Amazon and Microsoft and so on is that very, very large-scale computing standards have been connected to the Internet.
And they represent large-scale resources that are reachable through the Internet.
And so people talk about cloud computing and the implication is the computing resources are in the network cloud.
So Google does have multiple data centers, they're scattered around the world, they are extraordinarily powerful computing engines, huge amounts of storage.
But what should be important to you and your listeners to know is that Google, for the most part, does not hold a lot of the information which it indexes.
In fact, it doesn't hold most of it.
Most of it's out there on other machines on the Internet and Google's job is to simply crawl through all those web pages and index them and help people find which web pages are of interest.
With regard to Google Book Search, the same is often true.
We index the book content so you can do a full text search of the books that we've scanned.
But the book content may not necessarily be visible to you.
It depends on what arrangements the publishers are willing to undertake.
In some cases they'll say, "Well, you can put a small snippet around the words that are associated with the search, but the user has to go find a copy of the book somewhere in the library or buy it from Amazon or near a bookstore."
In other cases, the books are known to be out of copyright or be on copyright and therefore their content can be put up on the net freely.
In other cases, negotiations have occurred and maybe some portions of the book can be put up, maybe some chapters can be put up.
So there's a spectrum of access.
Google doesn't necessarily hold the contents of every book except in the sense that they've indexed as much as they can.
Plainly, every one of these initiatives has limits and even Google knows that there is a finite limit to how much we can achieve in the course of a year or two or ten.
But our ambition level is to try and stimulate this kind of scanning and capture of content and retrospectively so as well.
Books that have never been online before being picked up by scanning an optical character recognition.
What I hope will happen in the future, however, is that the publishers will voluntarily decide that the book content should be available in multiple forms.
And that it is as useful and perhaps even more useful to make the books available in online form as it is to make them available in print form.
There are some evidence that when you make books available in online form that you increase the sales of both forms that you can be people will find the books that they're interested in and maybe even use the online form to search the book and find the chapter they're interested in and then they go get the real book and they read it that way.
So there may be a variety of ways in which schools, books or capability will enhance as opposed to degrade the sale of conventional books as well as online versions.
I'm convinced of that. I think it's the same thing in music.
When you hear something either on the radio or you can download a CD that you like, the tendency is to go and buy the CD so that you have regular access to it.
But particularly if you're not online at the time. There's a band called Radiohead and it might even end with one of their songs and the outgoing music and they, their album in Rainbows was the first album to be available to download free on the internet.
And as a result it was downloaded by millions of people but it's immediately sold to million copies.
That's wonderful. That's a very good example. I like that.
So I was reading, you know, in preparation for the show some things and there is a editorial in our San Francisco Chronicle by Richard Schloss, who said that microsoft's founding mission was to put a computer on every desk and in every home.
And in that regard to decentralize computing, take it away from mainframes. Whereas Google is resentralizing computing by trying to bring it all back into some kind of supercomputer cloud thing. Is that accurate?
Well, it's not exactly accurate but there are pieces that are worth exploring a bit. One of the side effects of having software in every laptop and every desktop which has to be maintained and updated and so on is a certain amount of expense.
You have to license each copy, you have to make sure that you upload or download the changes and updates and so on.
And so the maintenance activity is actually fairly intense. If you turn this around on the other hand and put most of the operational software, the functional software in the cloud in a computing facility, then any changes and updates will occur for everyone.
So, famously as opposed to being a very uncertain, spotty kind of thing, sometimes driven by economics or inconvenience.
So, Google's hope here is to not only make the computing more uniform by using a computing utility concept, which by the way is not new. If you were ever open a book, the 1960s you'd hear about the computing utility as a gigantic building, the smoke coming out the top of that.
And Google's data centers and other people's data centers are much like that, except instead of a single big supercomputer, it's actually tens of thousands of small little machines all stacked up in racks.
So, there isn't interest on the part of Google to do that. There's another reason for this as well. It's convenience for the users.
For example, when you and I are working on a common document, if we both have separate copies of it that we're editing, then figuring out who has the up to date copy as a problem and expose our ten of us and not two.
On the other hand, in the online environment of Google, there's only one current copy, it's the one that's on the net.
And as we edit it, it's being updated so that everyone's instance of it, everyone's view of it, is modified appropriately.
Our ability to work together is enhanced by this ability to kind of meet in the cloud.
So, I think Google's interest here is in making software more reliable, more uniform, more easily adaptable and maintained, and to enhance our ability to collaborate and share information.
All of those things seem to come naturally with this cloud-like concept. We replicate the information in multiple data centers, and we do that for reliability. We also do it for response time, responsiveness.
So, if you are in a particular location topologically, we'd like to connect you to the nearest data center so that your interactions are swift and responsive.
And so, we replicate the information and the data centers around the world.
I just wanted to correct the name of that column I wrote. It was written by Randall Strass, not Sasha, and he actually wrote a book came out this year called "Planet Google Free Press 2008. I don't know if you're aware of it."
Actually, I'm a professor. I am ready yet.
So, I shouldn't plan it to Google, you say. I don't have a look at that.
It says one company's audacious plan to organize everything we know.
And I like that audacity. In fact, ever since Google was provided to a Stanford faculty member as part of our online endowment.
It's been, again, like magic, that you could have a technology, a kind of search engine that would so instantaneously provide you with the reference that you're in search of.
And this boldness is very appealing. Of course, it also sends little tremors of anxiety in people, of course, because it's so bold that it's done.
So, let's analyze this in two or three different dimensions. Robert, first of all, Google is not the only search engine around, and we're quite aware of that.
And we know that any of our users could switch to another search engine with the click of a mouse, and they do.
And that's good. We see that as a competitive environment, which forces us to push ourselves harder.
The second observation I would make is not everything that you find on the net is necessarily accurate.
But Google can't guarantee accuracy. All I can do is say this is what I found. And I think in classroom settings, it's awfully important for students to understand exactly that.
And to become more critical about the information that they read, whether it's on Wikipedia or, you know, any other web page.
I would love to have exercises in the class operated by the teachers that say, let's look at this web page for some web page.
Your homework assignment is to analyze the contents of that web page and explain why we should or should not be confident in the data or content of that web page.
By the way, your homework assignment is not done if you only use online sources.
I expect you to go to a place called a library. It has things in it called books. You're supposed to go and find out whether there are any books that are relevant to that.
Now, we hope we can help with that at Google by doing Google Book Search. But you still need to show up in the library.
And I want students to understand the critical thinking is part of their survival kit in a world of uncertain information.
That's a vision that I applaud and I certainly support it.
And I think I might try it out to tell you the truth because it's an excellent exercise to at least show the students what is the difference between a vetted, a piece of scholarship and unvetted.
When I'm amazed at how good are a lot of the unvetted things on the web.
It is quite very surprising.
But the problem here is like newspapers. You read something about which you know nothing.
And you can't tell that it isn't accurate just like a newspaper story.
Google has started a project called No Can, a Well, which sort of stands for the Noell is a bit of knowledge.
And it's intended to have characteristics not unlike Wikipedia except that the authors must declare themselves.
And you have an opportunity to evaluate the source of the information where isn't the Wikipedia that tends to be anonymous contributions.
Finally, since we've gotten to the end of our hour, the motto of Google is very interesting to me. Don't be evil.
I take it that there's something in that motto that met alludes to the Hippocratic Oath, which the first principle of which is don't do harm.
Do not do harm. Do not do harm. Do not harm. Do not harm.
That's right. Of course, the difference between do no harm and don't be evil is that the first one is about action.
And the Google motto is about being good or, let's say, don't be a key.
But not being evil.
Not being evil for some definition of evil.
Right. I think that we believe it at Google.
I mean, our intentions are not evil even though I think many people see a dark side of all the information that is accessible.
Through Google, it's interesting reactions that people have.
We provide an enormous number of free services on Google.
And the only reason we can do this is because we have a very effective advertising engine, which allows us to pay for the cost of all those services.
No one is required to use them. No one is forced to use them.
We don't have a monopoly on anyone's decision-making.
Unlike some other conditions or circumstances of other companies.
So we feel like our contributions are not evil in any sense.
It's true that the accumulation of information about what people do on the net could be viewed by some as a kind of privacy threat.
We've gone out of our way as much as we can to protect people's privacy.
We have fought court battles for months in cases where demands have been made on us.
We've been careful about the way we've deployed our system in China to avoid putting at risk the Chinese that we don't do blogging and we don't have email there because we don't want to be forced to disclose under some government mandate in China, the sources of the content.
So the company is certainly not perfect, but we really believe that motto and we try to live it.
I'm a huge fan of Google, I have to say, and I'm also proud that it has a Stanford Matrix, if you like.
And of course there are these concerns that it's only natural that some humanists who avail themselves of what Google offers every day.
But not feeling in any sort of control of where it's going and where it's heading.
One naturally might wonder whether there is anyone out there who is going to be able to control the technology itself.
But Google does this good job as anyone I know, so thank you very much for being on the show. It's been a fascinating hour.
Robert, I've enjoyed it very much and I certainly will make a point of listening to the KZSU podcast as I have an opportunity to do so.
Thank you for doing that. Bye bye. Bye bye.