字幕列表 影片播放
WALTER ISAACSON: You know, we're very, very lucky here.
I mean, when I wrote this block, I
thought, who are the great heroes of it,
and Dr. Khan and Vint Cerf are among them,
and I hope you appreciate how cool it is, especially
for somebody like me who loves history,
to be able to talk the internet with you all.
VINT CERF: Well, it's a real pleasure to have you here
and to have the rest of you here in our new offices.
For those of you who came in more on time,
we kind of started just informally.
This man sitting next to me has had
quite an interesting history, so I'll
repeat a little bit of that.
He's currently the chairman of the Aspen Institute
and was formerly the chairman of CNN.
He was also managing editor of Time Magazine
and has written a whole bunch of books, the most recent of which
you have before you, "The Innovators."
The most recent before that was Steve Jobs's biography
and many others before that, all of which
are well worth reading, so we welcome you to our presence
and to an opportunity to ask you some questions about this book.
Sitting over on my right is Bob Kahn.
I was going to say the late Bob Kahn, that would be bad taste,
so I won't do that.
BOB KAHN: Wasn't that late.
Your problem is your garage says it closes at 7 o'clock,
so we were about to go find some other garage to park in.
VINT CERF: Well, since we're all here,
I can make the door open for you.
How's that?
WALTER ISAACSON: All you have to do is say, OK, glass,
open the door.
VINT CERF: That's it.
I'm wearing one right there.
BOB KAHN: Is the audience all from Google?
VINT CERF: No, no, no.
These are people from all over the place.
WALTER ISAACSON: Who's from Google?
VINT CERF: Fair number but not everybody.
So what I thought we would do, Bob,
is start out with a couple of questions about the book itself
and what Walter has discovered.
This thing covers quite a broad range
of topics about things that were highly innovative
and made huge changes in our history.
So what is it that you would distill
from what you've already discovered
in writing the book that you would like us to take away?
WALTER ISAACSON: Thank you, and let
me say that when we get further along
and we're discussing the internet,
I'm going to turn the tables because to be
able to ask questions of the two of you
about both the ARPANET, the RFCs, and then
the internet protocols.
So watch out because that's going to be a two way street.
VINT CERF: Fair deal.
WALTER ISAACSON: The main thing I discovered.
I've written biographies a lot-- Steve Jobs,
Einstein, Franklin-- and those of us
who write biographies kind of know in the back of our minds
that we distort history with the great man theory of history.
We make it sound like there's a guy or a gal in a garage
or a garret who has a light bulb moment and innovation happens.
True innovation, and there's no better example of it
then the ARPANET RFC process and then the TCP/IP process,
really comes from pairs of people, teams of people,
and from collaboration, bouncing ideas off against each other,
finishing each other's sentences,
and the internet age in particular
allows such teamwork and collaboration.
The other thing was I started this book when I said,
way earlier before we started, we
were talking about when I ran digital media for Time Inc
in the early 1990s, and I first met Vint when he was at MCI,
and I was trying to convince Time Inc to do things
with the internet.
And so the president of Time Inc says to me, well,
who owns the internet?
So first of all, I think to myself
because I don't want to get fired,
that's a clueless question.
I say, nobody owns it.
He says, well, who built it?
Who created it?
And I realized after thinking how
clueless that question was that I did not really now.
And if you feel detached from the history of the things you
use, there's sort of a detachment that leads to you
not feeling comfortable with it.
So I wanted to know, how did computers happen?
How did the PC happen?
How did the transistor become the microchip,
and then how did the internet or digital packet switch networks
come into being?
So that's why I did this book, and in doing so, as I said,
you learn about the importance of collaboration,
but you also learn that there wasn't
one person who invented the computer.
And certainly, even after doing this book,
you and I have been discussing some of the,
no, I should get more credit type feelings,
but it was done creatively and collaboratively.
VINT CERF: This notion of collaboration
resonates very well with all of us
at Google because a lot of the tools that we make and use
are exactly collaborative elements,
things like Shared Documents, things like the Google
Hangouts, things like that where multiple parties can
communicate all at the same time.
WALTER ISAACSON: That was the original intention
of the internet or ARPANET was time
sharing research computers, and then collaborating.
I even did that with this book.
When I was writing that, I said, wait,
this is how we collaborate, and I put parts of this book
online, like on a medium with Google Docs, whatever,
and said, everybody share.
Help me collaborate.
Help put stuff in.
And people like Stewart Brand, who I knew,
started rewriting the demise of the Whole Earth Catalog,
and exactly what drugs were being served at the demise
partly because I got the drugs wrong.
But people like Dan Bricklin were explaining,
no, VisiCalc was done this way.
And so they're in the book from having been crowd sourced.
BOB KAHN: Walter, I'd like to ask you a question.
Vint and I were both on a panel at the National Academy
about a week ago, and one of the questions that came up was,
what is the internet?
One of the panelists took the position
that the internet is what people think it is.
So if you think the web is the internet,
then that's the internet.
If you think the web is this.
I just wondered, do you care much
in writing a book like that of clarifying that example.
Because a lot of places in the book, you say,
the ARPANET morphed into internet, and not
exactly clear what the internet is from reading your book.
Are you comfortable with that definition?
WALTER ISAACSON: I see the internet,
and I'm going to turn it to you, because you should answer
it better, but I see it as the TCP/IP protocols.
Now, the question is when Tim Berners-Lee
puts a layer of that [INAUDIBLE] that have hypertext protocols
and markup languages, to me, that's still the internet.
Now, the question then becomes, if I'm using Twitter,
is that the internet?
And the answer technically is no.
I mean, that's not--
VINT CERF: Well, actually, we could argue that because--
WALTER ISAACSON: Oh, we could argue it,
so let me turn it back to you all.
VINT CERF: Wait, wait, wait.
You're right in the middle of chapter six,
and I wanted to at least point out to you--
WALTER ISAACSON: Start at chapter one?
VINT CERF: Well, we don't have to go through every chapter
but this starts out with Ada Lovelace, who
was Lord Byron's daughter.
And I wanted to start out by just asking, what
was it that drew you to that story first,
because it's in the 1850s in England,
and it's the earliest manifestations
of mechanical computing?
What made you decide to start there?
WALTER ISAACSON: To be honest, what
drew it to me first was my daughter had not written
her college entrance essay, and my wife, who we've talked
about, was going nuts the way Yuppie wives do, and saying,
get it done.
And one day, Betsy said, I did it.
I said, what was it on?
She said, Ada Lovelace because she's
a computer geek and stuff.
And I paused.
I kind of knew the name Ada Lovelace,
but I didn't know exactly what she did.
VINT CERF: You didn't mix it up with Linda Lovelace?
WALTER ISAACSON: No, I did not.
I'm surprised that people all remember who she is.
So I became much more interested in Ada Lovelace.
Now, Ada Lovelace is partly a symbol in this book,
because as a person, she's kind of controversial,
but what she does as Lord Byron's daughter is
her mother, Lady Byron, was optically fond of Lord Byron
by the time Ada was growing up, for reasons that Byron
fans will understand, and so had her tutored mainly
in mathematics as if that were an antidote for her becoming
a romantic poet, which Lady Byron did not want her to be.
So she embraces what she calls poetical science.
It's the ability to link the poetic and the beauty
of humanities with the technology of science.
She loves, for example, traveling the Midlands
and looking at the punch cards that
are being used in the mechanical looms of the Industrial
Revolution in the 1830s.
VINT CERF: The Jacquard looms, yes.
WALTER ISAACSON: The Jacquard loom,
and showing how they weave patterns.
Her father, Lord Byron, was a Luddite,
and I don't mean that figuratively.
His only speech in the House of Lords
was defending the followers of Ned Ludd, who
were smashing these looms because they
were putting people out of work.
And he also is there with Mary Shelley
when they do Frankenstein's monster.
This notion that technology can destroy us
was ingrained in Lord Byron.
Ada felt the opposite.
She had this friend, Charles Babbage,
who had made a pretty good calculator called
the difference engine, was trying
to conceive one which he never got
built called the analytical engine, which
uses punch cards to do the processing of the numbers.
Ada writes a set of notes that are totally fascinating.
You've got to read them.
And among the notes to this thing
is how if you use punch cards, the machine
will be able to weave patterns like Jacquard's loom,
as she puts it.
In other words, the machine will not only do numbers, she says.
You can make it do words, or even music.
I can feel her father rolling over in the grave.
Patterns, anything that can be notated in symbols, she said.
So to me, that is the core notion
of what a computer is, anything that can be notated in symbols.
She has many other things, but I'll
only mention one, which is her final note is
that they'll do anything but think.
Machines will never be creative.
Machines will never think.
And 100 years later, assuming you're
going to jump there, Alan Turing, who really comes up
with the concept of the universal computing machine,
and then works at Bletchley Park to break the German ENIGMA
code with some people.
They build the Bombe and then Colossus.
He writes a wonderful paper.
The movie's about to come out called "The Imitation Game."
He called it the imitation game.
We now call the Turing test to address
what he calls Lady Lovelace's objection,
because Turing was fascinated by Lovelace.
And he says, well, you say machines will never think.
How would we know that?
If you had a machine in a different room and a person,
you couldn't tell their answers apart,
you'd have no empirical reason to say machines don't think.
And so he believed in artificial intelligence,
that eventually, we would have machines that think without us.
I basically use that as one of the framing devices
in the book, the notion of those who
believe that the intimate partnership between humans
and machines will get us forward and those
who believe that the pursuit of artificial intelligence
will get us forward.
Sorry for the long answer.
BOB KAHN: One of the first programs
that I ever used-- Vint actually wrote one
about PARRY and the DOCTOR.
This is about the DOCTOR program.
It made some outrageous statement,
and it was supposed to be a conversation about you.
I remember when I saw that, I responded to it, "my oh my,"
and its response back was, "your oh your?"
And I knew we wouldn't get very far.
WALTER ISAACSON: With all due respect to my Google Android's
ability to talk to me, it is really hard
to do language processing.
VINT CERF: Yes.
We have a lot of people who have confirmed that.
BOB KAHN: And we know quite a bit about it from [INAUDIBLE].
WALTER ISAACSON: You know more than anybody.
VINT CERF: The earliest experience
I ever had with trying to do language translation
was at Stanford when I was an undergraduate in the 1960s,
and we were pretty naive.
We had this Russian/English dictionary,
and we poured it into the computer.
We typed in, "out of sight, out of mind."
We translated that into Russian, then
we translated it back into English,
and it came back, "invisible idiot."
[LAUGHTER]
Which told us that there was more
to this than the dictionary.
WALTER ISAACSON: By the way, the book
ends with a chapter called "Ada Forever," which
is what you just said, this notion of understanding
a bit of language, or being able to spot your mother's
face in Grand Central Station, which
if you go to the total domain awareness system
down in lower Manhattan, they have all these cameras.
They say, oh, we can do that.
Or a robot that can walk across the room and pick up a crayon.
A four-year-old can do it but a machine can't.
It's called Moravec's paradox.
VINT CERF: It's getting better, though.
WALTER ISAACSON: It's always getting better,
but it's always 20 years away.
It's always a mirage.
VINT CERF: Look, the latest Tesla announcement,
the D thing parks itself.
We're getting closer and closer to something useful,
even if it isn't as fully creative as--
WALTER ISAACSON: I was onstage yesterday
with Elon Musk, who told me that the singularity is so close he
wants to go to Mars because machine learning will make
machines want to destroy us, and he's smarter than I am.
You all at Google are smarter, but I've
been reading that since 1955 right after Alan Turing
does his Turing test, and they always say, 10 years from now,
we'll have machines that can think better than we do.
Maybe so, but as the heroes of this book,
Lick Licklider, Doug Engelbart, Alan Kay, say, in the meantime,
let's make our machines connect to us better
than try to out-think us.
VINT CERF: Well, let's make partners out of them,
if nothing else.
I have to admit Google search is a pretty cool partner when
you think about it.
It does stuff that I couldn't do.
WALTER ISAACSON: And Google search
is done not by an algorithm that thinks without us,
but an algorithm that harvests millions of links
put every day onto websites by humans.
VINT CERF: That's right.
BOB KAHN: I wanted to ask you one question about something
that struck me when I first really understood
what Turing and company were doing at Princeton.
I gave the keynote at the 100th anniversary of his birth.
VINT CERF: His birthday.
BOB KAHN: Last year [INAUDIBLE].
It was one of a multiplicity of them.
But one of the things that struck
me back then was when Turing wrote
that paper on computable numbers, that was actually
not easily accepted by people because they didn't think
computing was something that a machine could do at that time,
really.
If you looked at what Church did with the lambda calculus,
it was a conceptual thing.
If you looked at what his colleagues, like Kurt Godel,
and there were a number of really excellent logicians
there at the time, they were thinking
of conceptual frameworks for organizing things
so that the logic made sense.
WALTER ISAACSON: You have to remember that, as you do,
uncomputable numbers was not written
to invent a universal computing machine.
That was just a thought experiment
he uses to solve one of Hilbert's
mathematical questions about whether all problems are
decidable.
BOB KAHN: But it seems to me that the one thing that really
came out of that was the idea that you could actually
create a machine that would do these computations.
They eventually showed that recursive function
theory and the lambda calculus and the Turing machine all
could do all the computable functions,
but that wasn't known at the time, nor was it obvious.
The realization that pops out at that time
that you can actually build something
around these numbers that was so fascinating,
but I didn't get that out of the book,
and I wondered, was that something you just didn't think
the public would understand, or was--
WALTER ISAACSON: I thought that was important, totally
fascinating.
I drilled down, read the entire Turing thing.
That is a sideline because it's not
about the actual what they were doing at Bletchley Park, which
I was trying to tell.
BOB KAHN: This was at Princeton.
WALTER ISAACSON: Right.
What happens there, and I would love your feedback on it,
is they are trying to figure out whether all problems are
computable, whether it's decidable to know which ones
or not, and you end up with Godel
saying there's an incompleteness inherent, and with Turing
and Church both saying, and we can't even
decide what is completable and what's not completable.
I'm oversimplifying.
So there's still a mystery that's
wonderfully inherent in math, that not all things-- I mean,
there is an incompleteness in the mathematical system.
BOB KAHN: But there was also the fact,
I believe, that what Turing had done
was not appreciated that much by the other people around him
because they were the pure theoreticians,
and Turing was more of an engineer,
if I can use that term, even though he never built anything,
much like the conflicts that we've seen in computer science
in recent days between the people who
are the pure theorists and the people who are actually
building stuff.
WALTER ISAACSON: You're absolutely right,
and I do say in the book that there
are only four or five people who even comment or review
his paper.
It's not very well thought of.
And throughout history, you see that where Penn,
they're building these computers,
and yet it seems a little beneath their dignity
because it's not an academic, theoretical pursuit,
so then they get rid of the computer.
And even at Princeton, if I may say,
the IAS finally, after von Neumann leaves,
they don't feel computing is a grand endeavor in its own way.
BOB KAHN: Well, there was physics.
WALTER ISAACSON: Yes, right.
VINT CERF: Well, everything is physics
when you get right down to it.
WALTER ISAACSON: Everything is math.
VINT CERF: Well, OK.
So we start with math and we go to physics,
and the abstraction of physics is chemistry,
and the abstraction of chemistry is biology,
and the abstraction of biology is psychology,
and it keeps on going like that.
BOB KAHN: But you remember the symposium we had up at MIT
when Richard Feynman came and a bunch of others,
and it was called "The Simulation of Physics,"
and they didn't like that one bit because they said,
physics is not a simulation.
It's not some bigger, digital thing that's creating it.
WALTER ISAACSON: You're losing control of the bus.
VINT CERF: I'll come back to a couple of things.
Well, it's impossible.
Nobody could manage these guys.
But I do want to draw you back for just a couple seconds.
First of all, the point about Turing, for example,
reinforces the notion that theory is helpful
if you can underwrite or underscore or somehow support
applications.
If you don't have much theory underneath the application,
then you don't quite know what to expect
and what you're doing.
So it's clear that the Gedankenexperiment that Turing
did is not unlike the Gedankenexperiments
that Einstein did.
A lot of his work was not math.
It was Gedanken imagery.
He thought in those terms.
WALTER ISAACSON: Visual thought experiment.
VINT CERF: Now let's go to something more concrete,
though.
You have a really interesting story
here about what happens when the transistor gets built in 1947.
The three guys who do this go off-- well, Shockley anyway,
goes off to start a company to build transistors.
And I think it's worth recounting
what happens because the focus that we got
to early in this game was collaboration
and what happens when you get groups of people all focused
on trying to do something.
So what happens in that case?
Shockley goes to the West Coast and starts his company,
brings some really smart people together,
which should have produced an important collaboration.
What happened?
WALTER ISAACSON: It goes to the theme of the book, which
is that vision without collaboration and execution
is just hallucination.
What you have at Bell Labs is this wonderful environment
where Shannon is riding a unicycle up and down doing
information theory, but you also have--
VINT CERF: Did he do that when you were there?
WALTER ISAACSON: Yes.
BOB KAHN: Shannon had left.
I started my career at Bell Labs.
One of the things I got from your book
that I found very interesting was
he started at the same place that I did,
which was 463 West Street in New York.
WALTER ISAACSON: Right on Greenwich Village's side
of the Hudson.
BOB KAHN: It was right on the Hudson River,
and you could watch the Queen Mary and Queen
Elizabeth coming in every Tuesday
and going out every Thursday.
WALTER ISAACSON: It's still a wonderful warehouse,
and then they moved it to New Jersey.
BOB KAHN: Murray Hill existed at the time that I was there.
I think it probably didn't when he was there.
WALTER ISAACSON: It becomes a cauldron
in which you have information theorists, quantum theorists,
like Shockley to some extent, and certainly Bardeen.
VINT CERF: Bardeen for sure.
WALTER ISAACSON: Walter Brattain,
who is a great experimentalist, but also people with grease
under their fingernails who've climbed a pole to figure out,
how are we going to amplify a phone signal between New
York and San Francisco, and they collaborate.
They work together.
They have a little bit of a problem
once the collaboration was over, does Shockley get
to be in the pictures with Bardeen and Brattain,
and who gets credit and who gets a Nobel Prize,
but before you fight over Nobel Prizes, you collaborate well.
Shockley was really bad at collaboration,
and he got worse, and the Nobel Prize didn't help.
I'm glad we've not all won one yet
because it made him into a bit of a jerk.
He goes off and he forms Shockley Semiconductor,
and he gets more and more uncollaborative with his team,
and more and more top down dictatorial,
unlike, say, a [INAUDIBLE] environment.
VINT CERF: Hiding his research results from other people.
WALTER ISAACSON: Hiding his research from his own people,
making them take lie detector tests,
worrying who's going to get the credit.
So obviously, the great thing about it
is it disintegrates because of this, and the people who
fly off from this big bang are so allergic to Shockley's
way of doing things that they move to the other extreme, two
of the best examples being Gordon Moore and Bob Noyce,
who form Fairchild Semiconductor,
and then eventually form Intel.
VINT CERF: Why was it called Fairchild Semiconductor?
WALTER ISAACSON: Because they did not
have venture capitalists back then,
so they wanted to start their own semiconductor company,
but they couldn't just go to Sandhill Road
and tell Kleiner Perkins, give me an investment.
Arthur Rock was just starting, and so he
helped put it together.
Arthur Rock was not yet a venture capitalist.
He was still a banker from the east.
So he calls his east coast buddies-- IBM, Bell-- and says,
do you want a division we can start
that will make semiconductors?
And they'll say, no, we don't want that.
We don't want an autonomous division.
We don't want to give them the power.
He's just about to give up and then went
to Sherman Fairchild, who's a bit of a nutcase and a playboy
and always out in nightclubs and 21 Club.
They make a deal and Sherman Fairchild says,
yeah, I'll stake you.
I'll make you a Fairchild Semiconductor division.
They were also doing cameras and aircraft.
And of course, it becomes the biggest division
in the company, and eventually, Fairchild
exercises the right to buy them out, after which Noyce
and Moore start their own company, Intel.
VINT CERF: That's amazing.
So this is sort of like one of those black hole explosion
things that--
WALTER ISAACSON: Well, they call them fairchildren,
because Fairchild then breaks up a bit,
and if you look at Silicon Valley, you can do a Google map
and just say Fairchild, and you'll
see all the companies that spun off.
BOB KAHN: You know, there are actually
two parallel stories with Intel and the work that
went on before that are interesting.
You focused on one of them in the book,
and I knew two of those people very well.
I've met Gordon Moore, but I knew Bob Noyce very well,
interacted with him a lot, especially when
I was doing the DLSI program at DARPA
and he was fighting the VISIC program in the DoD.
And the other was Andy Grove, who was in my class
in undergraduate school, so I knew him.
WALTER ISAACSON: That's in New York.
BOB KAHN: At City College, yes.
So I knew those two well.
But the other story that's quite interesting
is who gets the credit for the microcomputer.
WALTER ISAACSON: The microprocessor
or microcomputer?
BOB KAHN: The microprocessor in this case.
VINT CERF: This is after the integrated circuit, about which
there was another big argument.
BOB KAHN: That's right.
This was in the early '70s, when Intel was trying to figure out
internally how you connect one piece to another piece,
and instead of doing a one of, they
decided, let's build a programmable thing.
So you have a guy named Frederico Faggin, who actually
did the actual implementation of it.
You had a guy named Ted Hoff, who is credited by everybody,
and Bob Noyce told me, when he was pushing
some of the nominations for awards for him,
that he was sort of the architect of the whole thing.
But you had it all done in a lab that
was run by a fellow named Les Vadez,
and Les always claimed that he set
the environment and the tone, which enabled it all to happen.
Between three of them, they don't really
necessarily all see the development the same way.
WALTER ISAACSON: You know, it's a shame,
and that happens a lot.
And in this book, I'm already getting emails,
and you are, too.
I won't name names because we're online,
but people say, he didn't quite give me enough-- oh,
actually now I've read it, he give me enough credit.
The one I love is Kilby, as you know,
Jack Kilby does it at Texas Instruments, the microchip.
I'm moving back earlier, the integrated circuit.
And Noyce does it at approximately the same time
and separately.
And Kilby wins the Nobel Prize, and unlike Shockley, it
doesn't go to is head.
And he also says, if Bob Noyce was still alive,
he'd be sharing this prize with me.
They both deserve it.
They don't give Nobel Prizes when you're--
VINT CERF: Kilby didn't get that award until he was 82,
and he passed away a couple years later.
WALTER ISAACSON: But when they finally bring Kilby
and he's awarded it, the Norwegian scientist introducing
him says, everything we have in the digital revolution
comes from what you did.
And Kilby says, that reminds me of what the beaver said
to the rabbit at the foot of the Hoover Dam,
which is, no, I didn't build it, but it
is based on an idea of mine.
One of the things about collaboration
is it gets hard to allocate credit,
and I try very hard in this book to say,
OK, here's how I would parse the credit for the computer.
I mean, is it basically Mauchly and Eckert at Penn?
Is it Atanasoff at Iowa State?
Microchip, transistor?
We all want to parse out credit, whether we're
intellectual property lawyers fighting over the patents,
whether we're Nobel Prize jurists,
or whether we're biographers, and I tried to minimize,
and you and I were talking beforehand.
There's dispute amongst the people
who conceived of packet switching, which
are three or four or five people.
They all fight over it.
They're all quoted by name on the record, will tell you,
oh, this guy's an idiot.
I try to say, no, it was the collaboration.
Packet switching was an amazing idea.
It is the heart of, obviously, what you two did,
and I was slightly annoyed by the people who
wanted to disparage their colleagues'
contribution because there is enough credit to go around.
VINT CERF: Exactly, and this emphasizes several things.
The first one is a lot of stuff happens
because it's now possible to happen.
The economics are possible or the amount of memory available
makes it possible or the speed of something makes it possible,
and so concurrent invention is not
unusual under those circumstances.
WALTER ISAACSON: Especially, there
was a miracle year of 1939, let me say,
where conceptual breakthroughs have happened, including
Turing's papers published on universal computing,
but Shannon and other people are starting
to apply Boolean algebra to how to do it in circuits.
But you also have mechanical inventions, including
the vacuum tube and telephone switchboards,
that have come together, but you also have the drums
of war, both in Germany with Herman Zuse,
and at Bletchley Park in England,
and at Aberdeen Proving Ground in the United States.
It's dawning on people that firepower is actually not
going to be as important as computing power in winning
the next war, and so that's why I'm
convinced that you have a fertile ground from 1940
to '45 for the creation of the computer.
And likewise, another fertile stew
happens, which Google is part of it
out in California, Mountain View and all,
that in the early '70s, there was just this yeasty brew
of counterculture types, electronic geeks, people who
had been in the defense industry,
people who had read the Whole Earth Catalog once too often,
all of whom are coming together and want computing
power to the people to take it back from the big corporations,
and they're creating the Altair and the Apple and all
those machines.
VINT CERF: I was waiting for you to leap into something.
BOB KAHN: I have so many thoughts going through my mind,
I'm not sure where to jump into this.
One of the questions I had about the book
was, what prompted you to choose the things
that you did to emphasize?
It's a wonderful read.
WALTER ISAACSON: That's the toughest of all questions.
BOB KAHN: About the only complaint I have is
that you couldn't touch everything
with uniform treatment.
So for example, there are things that don't show up here at all,
like, for example, the role that NSF
played in bringing the internet out.
There are people who I think played an important role that
never get mentioned, like Steve Wolf is a good example.
VINT CERF: Absolutely.
Gordon Bell.
BOB KAHN: I just wondered, this is a reflection
on what you felt you wanted to write about.
WALTER ISAACSON: I had about a 1,000 page book.
BOB KAHN: How did you get to this rendition?
WALTER ISAACSON: When I did all my notes,
there were about 1,000 pages.
Wolf is in them, and the other things.
There's a certain limit in this day and age
to books, which is approximately 500 pages.
I mean, that's--
BOB KAHN: I've read books bigger than that.
WALTER ISAACSON: I know, but I meant-- not many.
BOB KAHN: I read "Godel, Escher, and Bach" in one swoop.
WALTER ISAACSON: That's what, 600 pages?
Not much more.
BOB KAHN: I think it was more than that.
WALTER ISAACSON: That's a good book.
BOB KAHN: One sitting.
WALTER ISAACSON: Yeah, but you know what?
Not everybody can read "Godel, Escher, and Bach"
in one sitting.
You're not the best focus group for the publishing industry.
But I did decide, hey, I want to keep it in the package,
and I'm going to get to why we don't
need to do that in the future.
So I ended up picking 12.
I just made a list finally at the end,
12 major things I wanted to do.
I wanted to do the computer, and so I focused on those three,
and by that, I mean the people who really began it,
not Control Data Corporation and others that pick it up.
I wanted to do the transistor, the microchip, video games,
because I actually do think that they were
important in the interactive computing realm.
I wanted to do software and how original
operating system software, and the women
as well as men who did it.
And I go down the web and others.
I could have picked many other things
and I could have picked many other people
in each one of those categories, but I said, OK,
who made a conceptual leap that I really feel passionate about?
Now, I know that leaves out more than it includes,
but it makes for a narrative.
The problem with narrative history
is narrative is defined by you leave out a lot of things
so that you have a thread.
BOB KAHN: But that's really interesting,
because if you focus it all on the conceptual leap,
you minimize the contributions of the people who actually
then implemented those ideas.
As an engineer, and I think Vint would agree with this, too,
a lot of the details are in making things actually work.
WALTER ISAACSON: Right, and I really
try to emphasize in the book that you
have to execute on vision, and that's
why I focus on the engineers and the software engineers of ENIAC
more than on this conceptual leap
that Atanasoff has at Iowa State because he
doesn't implement it.
And I hope-- I mean, you can be the judge--
that with Fairchild and Intel, we
get into a lot of the engineers who are making it.
Let me leap forward to what I hope
is the solution to this, which is I would love it,
and Google is the best place to invent it
with Google Docs, networks, everything else,
and I've tried.
I've worked with people.
I'm still not there yet, but I would
be quite happy to take this book and make it a curated wiki
Google Doc, and I have done that with some of the chapters.
It's why Dan Bricklin-- you say, why
is Bricklin in there for doing VisiCalc?
Well frankly, because he gave me all of his documents
when he read an early draft and it was on Medium, which
is like a Google Doc place where people put in.
I think the book of the future will
be a collaborative, crowd-sourced,
and yet author curated, meaning you just can't run off
anywhere, multimedia, where, let us
say you want to talk about-- let's pick one thing--
how the planar process worked and how the planar process led
to the microprocessor being easily invented at Intel.
I would love the videos, I'd love the original documents,
I'd love people to upload those things, because online,
on a Google Book, I'll call it, you couldn't make it
not 500 pages but 500,000 pages and video.
BOB KAHN: Is that still your book at that point?
WALTER ISAACSON: And that's why I
would like to curate it, meaning that you're
going to have fights.
I won't name names again, but even
with your friends who were originally there
with packet switching, just like the revision wars in Wikipedia
fought with more intensity than we seem to be fighting wars
in the real world these days, you
would have to have some curator who would say, OK,
enough of this fighting about who
came up with packet switching.
Let's try to just put it all and synthesize it.
But I suspect, like Wikipedia, the crowd
would also curate pretty well.
I also think, and this is the next thing
I would love Google to do, for that to be really fair,
there has to be a payment system.
I would be happy to put this book online--
I shouldn't say this too publicly,
but after "Steve Jobs" I'm not looking for more book sales.
VINT CERF: This only just came out,
so don't say that to your publisher yet.
He'll cancel your book tour.
WALTER ISAACSON: But let us say that Vint Cerf and Bob
Kahn looked at this and said, there
are 22 things you really left out
of the RFC process and the ARPANET,
and then the TCP/IP process we went through in '73,
I think it was.
And here's all our documents, and here's
somebody who should have gotten more credit,
here's his picture, here is an interview
I've just done with him.
I want to put all that up.
You would be doing that, as you would be editing a Wikipedia
entry, just for the good of the commons,
but there would be other people who
might want to contribute and actually make
a living out of it.
BOB KAHN: So this would be like sticky notes on your book?
WALTER ISAACSON: Well, it would be
like making my book into a Wikipedia page
where everybody could put documents up,
and I would love the royalties to then
be allocated in some crypto currency way
that Google could invent, so that if people are paying
$30 for this book and they read these parts,
the royalties get allocated to whoever contributed.
VINT CERF: So that's a challenge before us.
WALTER ISAACSON: And I hope my next book,
I'd love to write about Louis Armstrong, but you need music.
I would need a Wynton Marsalis to say,
here's the 17-bar cadenza that opens "West End Blues,"
and here's how he changed it from King Oliver's version,
but that would need to be crowd-sourced and multimedia,
and I want Google to invent that.
BOB KAHN: I always thought that one of the really nice things
that historians rely on are crafted treatises
on various topics.
There's a lady that Vint and I both know
who wrote some monographs on infrastructure development.
She would not write a single sentence in there
that she couldn't curate from some well known historian who
quoted something and made a point.
So how do you treat something like this
when the only real authority behind it
is the party who put it all together
and everything else is all this other stuff that's
stuck in there?
WALTER ISAACSON: This is much more footnoted,
source noted than, say, the Steve Jobs book,
which mainly came from interviews with Steve
and 60 or 70 other people around him.
I am probably not the best reporter
in the history of the world.
I just saw Bob Woodward when I was walking here.
He's much better.
I'm probably not the best academic historian.
I mean, somebody like you just talked about who does that.
I do have, I hope--
BOB KAHN: Yeah, but you write so well.
WALTER ISAACSON: Thank you so much.
But I also have, I hope, the ability
to combine historical resource, like looking at the monographs
or whatever, with just being able to call Gordon Moore
and saying, I'm going to get in the car,
I'm going to drive to Woodside, you've
got to spend all day with me, and I'm
lucky because I've been at Time Magazine.
Gordon Moore, who would not take the phone
call of my second cousin from Louisiana if he said,
I want to drive up to Woodside, I get to see these people.
So I do a lot of interviews.
There are about 100 or so, including you and others,
in the book, but I also go back to the papers and the IEEE
and everybody, your collaborated document
on the invention of the internet, a little original Len
Kleinrock and Paul Baran papers.
There's no quote in this book, I think,
or no major quote in this book or assertion that's not source
cited by name on the record, but all history
is just the next draft of history.
There will be people who will build upon this.
I think I bring some journalistic skills.
Other people bring better historical-- Janet Abbate,
others who have done great academic histories
of the internet.
And someday, when it's all part of this curated, crowd-sourced
system, you'll be able to hop around and get
whatever you want.
VINT CERF: So I have a suggestion.
He was threatening to turn the tables us,
so maybe we should do that, and I'd
like to let everybody else have some time to ask questions.
BOB KAHN: Are you trying to filibuster against that?
WALTER ISAACSON: No, I'm trying to--
VINT CERF: I figured I managed to--
WALTER ISAACSON: I have spent 30 years as a journalist.
I love asking the questions.
Here I am having the people who actually invented
the internet ask me questions.
I'm going, that's wacky.
I should be asking them questions.
BOB KAHN: Actually, when you read the book,
the impression that it gives, which
I thought is the one you were trying to give,
is that collaboration is really everything
and all these people really created the internet.
WALTER ISAACSON: Right.
I really wanted to, and it is true.
I'm trying to show collaboration, enough
credit to go around, and it is true, I sort of think.
And then afterwards, people say, you didn't give me
quite enough credit, or worse yet, they're
suing [INAUDIBLE] over the patents.
But generally, collaboration is really cool.
And I'll start, there's two real segments of the internet.
BOB KAHN: So when you point out Vint and myself,
what makes you single us out?
I mean, we could give you our views,
but I'm wondering what your views are.
WALTER ISAACSON: I will turn it around
and make you give the views.
I'll say that there's two great phases
in the creation of the internet.
Phase one was the creation of ARPANET.
I think you were at BBN probably--
BOB KAHN: I actually was a system designer.
WALTER ISAACSON: Yeah, at BBN, right?
Bolt, Beranek, and Newman, which was a private company but still
very associated with MIT to some extent.
BOB KAHN: And Harvard.
WALTER ISAACSON: And Harvard, and to some extent
with the government, because you had from Larry Roberts,
and I guess the original Licklider,
you had contracts to do what was unfortunately
named IMPs, Interface Message Processors, which--
BOB KAHN: Why is that unfortunate?
WALTER ISAACSON: Well, because when BBN got the contract--
BOB KAHN: I actually have the little slide
that Ted Kennedy sent.
WALTER ISAACSON: Yeah.
I'm going to do a Ted Kennedy joke here,
which I'll let Dr. Kahn do the punchline.
BOB KAHN: I gave it to Licklider's wife
at the ceremony.
VINT CERF: Wait, let him do his punchline.
WALTER ISAACSON: They don't know the joke
before you do the punch--
VINT CERF: It's his punchline.
Go ahead.
WALTER ISAACSON: No.
It's the Interface Message Processor,
and it's awarded to BBN, which is in Ted Kennedy's district,
and he sends a telegram congratulating you
on getting the--
BOB KAHN: Inter-faith message processor.
WALTER ISAACSON: Which, by the way,
is a wonderfully telling slip, because it was inter-faith,
in many ways.
VINT CERF: Well actually, he thought
he was congratulating them on what
he thought was an ecumenical effort.
WALTER ISAACSON: It was.
It was the ultimate ecumenical effort.
VINT CERF: An inter-faith message processor,
the Jews and the Muslims and the--
WALTER ISAACSON: We basically call them routers now,
but these were the machines that were going to--
BOB KAHN: Actually, I would say we call them packet switches
now.
WALTER ISAACSON: Yeah, but these were the machines
that would be sent to the original nodes on ARPANET that
would be an inter-faith or interface between the host
machine and the rest of the network.
BOB KAHN: So Vint, you mean if it wasn't ecumenical,
we wouldn't have gotten that thank you note?
VINT CERF: I don't know.
It was supposed to be something that was nondenominational.
The whole idea was it's non-proprietary,
and it was, from Kennedy's choice of words,
nondenominational, and that's very true.
The IMPs were there, packet switches,
to form this uniform, homogeneous network
to deal with an extremely heterogeneous collection
of computers that were on the edges.
And that demonstration, to take your point
about these two phases of the whole process.
The first phase demonstrated that heterogeneous computers
and operating systems could be made
to communicate across a homogeneous network,
and that was the beginning of understanding
what was possible.
WALTER ISAACSON: And it was the ultimate collaborative process,
which is what I'm going to ask you about,
because if somebody had to say, if you could show me
the true nature of a collaborative process-- this is
early on, before we're getting to the internet TCP/IP
process-- is, how are these IMPs going to communicate
with host computers and each other?
And everybody vaguely expected that, either
from the bosses of BBN or the bosses in the Pentagon,
the word would be sent down, but instead, it's
done collaboratively mainly by graduate student,
including you were a graduate student.
BOB KAHN: There are actually two parts of that process.
The one that had the computers talking to the other computers
was what led to the host to host protocols.
That's one Vint was very involved in,
along with a number of other people.
And the second part that had to do with how the machines talked
to these interface message processors was in a BBN report
that I authored called "BBN Report
1822," so that interface was all called the 1822 interface,
and that was a combination of how the hardware would work,
because you had different word size machines, how
the software would work, and that
was a precursor to having the machines talk to each other.
WALTER ISAACSON: Right.
And the one I focus on, because you
asked, why do you focus on things,
is the request for comment process,
because it was so collaborative and collegial, even the phrase
that Steve Crocker, who you turned me onto and I went
and interviewed for the book.
He's just a graduate student with you,
and instead of having the graduate student say,
here's the way we should do it, he said,
I had to think of a name that would
make people feel like they were collaborators in this.
And so when we sent them out, I didn't want to say,
here's our plan or our proposal.
He was standing in the shower at his girlfriend's house,
he said, and it came to him that he
should call it a request for comment
because it was so open and inviting, as opposed to, here's
what you're supposed to do.
Let me let Vint tell the story from there, and then Bob,
but you were in the RFC process, weren't you?
VINT CERF: Yes, but to draw a little bit further
on what Bob was talking about, in order to have communication
at all, you need to have commonality, and so the 1822
hardware and software interface between the host that
were heterogeneous and the IMPs that were homogeneous
was the first important piece of commonality.
The next piece of commonality is that the computers
at the edges of the net needed to share something
in common so they could communicate
with each other, packet formats and protocols,
handshaking, and so on.
That's what Steve and I and others
worked on, the host to host protocols.
Steve and I and the others, Jon Postel,
were graduate students in Len Kleinrock's laboratory at UCLA,
and we kept thinking that somebody, like Larry Roberts,
who was in charge of the project at ARPA,
would come out and tell us what we were supposed to do,
and he didn't.
So we thought, well, we have to do something.
BOB KAHN: He didn't know.
He was expecting you to tell him.
VINT CERF: Len didn't tell us either.
So basically, it's graduate students who
are scratching their heads saying, well,
we have to do something.
So Steve did a beautiful job, exactly
as you say, organizing this group, which
spanned across a dozen different universities,
looking at not only how to get the computers to exchange
postcards, so to speak, but also to build on top of that.
So this notion of layering came out very early
on in the ARPANET history.
We learned an enormous amount from that experiment.
Email came along in 1971, two years after we turned the IMPs
on at UCLA, Santa Barbara, SRI International, and Utah.
BBN, where Bob was still, was pumping out IMPs,
and they were popping up on the network, one
a month for quite awhile.
So we were experiencing utility from this homogeneous network,
and got pretty excited about that.
So that's the first phase, as you said.
WALTER ISAACSON: And to me-- and tell me if I'm wrong--
but the notion that you all did this to an RFC process
I found fascinating because it almost
seems like it's ingrained in the genetic code of the internet,
which is a collaborative system, built collaboratively.
VINT CERF: This, I think, is fairly important.
Bob, I think you've been very articulate about this.
Not only did we have a bunch of people
who shared this common excitement
and view that we could make this stuff work,
but the institutions that have been created,
arising from the ARPANET project and the internet project,
were equally bottom up and collaborative,
and this notion of multi-stakeholder models,
where all the parties who have a stake in something
participate in the development, whether it's
the technical development or development of policy,
infuses all of this internet environment.
I have to say that Steve's trailblazing--
WALTER ISAACSON: Steve Crocker.
VINT CERF: Steve Crocker's trail blazing-- that's right.
You would think Jobs, wouldn't you?
So Steve Crocker's trail blazing I think
helped set that tone, and all the institutions who came out
after that had a very similar character.
BOB KAHN: The counterpart to that,
though, is that these computers were all linked together
through a single network called ARPANET,
and the process by which the ARPANET was built
involved DARPA, generally in the form of Larry
with a group that be coordinated with,
essentially writing the requirement specs.
Half a second delay across the net, 1,000 bit packets or less,
and so forth.
But the actual building of that net
was much more insular than any of the other things that
went on.
In fact, there was almost a desire by BBN
to keep it that way.
In addition to playing a technical role there
that was pretty important, I was the only one
that was assigned to deal with the external community.
That was my job.
If anybody had a question, I was the only one
they knew of because nobody else,
they were not interested in having
them be deviated from a very short time frame
to get something done with questions from the community.
And this BBN Report 1822 was one of the hardest things
I ever wrote my life because I had to make a document that all
of the sites that would connect had
to be able to read and understand,
and I have to solve their problems,
at the same time abstracting out everything
about the net that they didn't need to know about.
So it was an intellectual exercise of the first order
to actually write that report in a way that
didn't get into so much detail.
I get so many questions about, exactly how does the routing
work, and exactly how does the buffering work,
and how does the [INAUDIBLE] control work?
WALTER ISAACSON: You just answered the question
of why I have to leave some stuff out, too.
Whenever you're writing something like that.
I want to ask you all a question, before we
get to the internet TCP/IP, which is the next step,
is there's a great controversy, which I show
both sides of in this book, as to whether or not the ARPANET,
I'll call it, was designed with partly surviving
a nuclear attack in mind.
BOB KAHN: Yeah.
Vint wasn't internal to DARPA at the time,
and I wasn't there when it first started,
but I read a lot of the internal documentation on that
and talked to a number of people.
There are two or three points of view on this.
One is, and you call it out in your book, Charlie Herzfeld,
who Taylor likes to talk about the 10 minute story,
and you clarify that it was really
multiple years that [INAUDIBLE].
WALTER ISAACSON: I hate clarifying stories
that are so good you want them to just
be told the way people like that.
BOB KAHN: That one, and that the Kleinrock story about sending
the L and the O. That was years before we had the protocols,
years before we had email systems and the like,
but they're nice stories.
When it comes to the actual implementation, what
was motivating the folks in the office
was the economics of sharing computer resources that
were very expensive.
I can recall Harvard got a big disk.
At the time, it was probably $1 million,
and so some other university like Carnegie Mellon said,
we need one, too.
Well, why spend another $1 million on Carnegie
when they could share the one at Harvard, save DARPA some money?
Part of it was the economics, but if you actually
look at the documentation, it talks
about needing to build this net to facilitate computer resource
sharing, so there were some interesting technical problems.
It points out that the network itself
could have a big effect on communications.
But this was at the office level,
so there was nothing that pervaded the office activities
that had anything to do with nuclear survivability or any
of the military aspects.
What Steve Lukasik wrote about, and he's written very clearly
about this over the years, was that at his level and above,
the reason they thought this was a good investment for DARPA
to do is because it could have all of this impact.
VINT CERF: Distributed.
There's no central hub to the internet.
BOB KAHN: It could be reliable communications.
VINT CERF: And it will route around any Soviet attack.
BOB KAHN: It could be a better solution for command
and control and so forth.
So to say that that wasn't there because the office didn't
see it would be denying what was going on elsewhere.
VINT CERF: If I may read, there's
a very funny exchange in here about that.
Go ahead.
WALTER ISAACSON: This is the one I think you're talking about,
which is when Lukasik, who said, hey,
we were doing it to survive a nuclear attack.
You down at the office level didn't know that,
but that's why we were funding it, and Crocker says, no,
you weren't.
So anyway, I interviewed them both, and--
BOB KAHN: Yeah, I know.
Crocker changed his mind.
WALTER ISAACSON: And Lukasik says, "I was on top
and you were on the bottom," this
is what he says to Steve Crocker,
"so you really had no idea of what was going on
and why we were doing it."
To which, when I told Crocker this,
Crocker responded, "I was on the bottom and you were on the top,
so you had no idea what was really going on."
That's the way life is sometimes in a collaborative process.
VINT CERF: I get to add one little codicil to this
because the original work that Paul Baran did
on survivable networks, the 11 volume series
called "On Distributed Communication,"
was all about building a survivable command and control
network.
It was digital in nature but it was caring digital voice.
Think about this.
He's doing this around 1960 before we
have integrated circuits or anything else.
So he's looking at survivable command and control,
then along comes this ARPANET thing and for many of us,
it was a resource sharing project.
Then internet comes along, we come to that,
and I remember thinking, when it was my responsibility
to manage the program, that we should actually demonstrate
that the TCP/IP protocols would in fact allow
you to survive a nuclear attack.
So although I may be jumping ahead a little bit,
towards the end of my tenure at DARPA,
we actually put packet radios in the strategic aircraft
and flew them around and artificially disabled
the ARPANET, broke it up into pieces
and reassembled it using packet radios in the air
and on the ground using TCP/IP, to demonstrate
that you could actually make a survival network.
WALTER ISAACSON: And since women programmers
don't often get their due in history,
you had a great woman, Perlman right?
VINT CERF: Radia Perlman did a fabulous job
of designing the routing system that would recover,
by figuring out what the broken segments were
and how to glue them back together.
She did a beautiful job.
WALTER ISAACSON: Let me, Bob, if I may, leap to I
think what would be spring of 1973, if I remember correctly.
What you have is ARPANET, and as you say,
I don't do enough with the NSF net,
and there are other nets, and ALOHAnet, and all--
VINT CERF: Of course, they don't exist yet.
BOB KAHN: I want to stipulate for the record,
I didn't think you could cover everything.
I was just asking why you picked what you did.
WALTER ISAACSON: But anyway, all of a sudden,
there are a whole lot of packet switch networks,
but they don't talk to each other.
VINT CERF: There were only four.
WALTER ISAACSON: Only four.
ALOHAnet--
VINT CERF: At that point, you had ALOHA,
you had packet radio, you had packet satellite,
and you had ehternet.
WALTER ISAACSON: And then you had ARPANET.
VINT CERF: And you had ARPANET, so there's five.
BOB KAHN: You're going at the story from the back end.
I would like to tell it from the front end.
WALTER ISAACSON: Tell it from the front end.
BOB KAHN: The front end was we had
the ARPANET was DARPA had commissioned that.
We built it.
I then go to DARPA and I show up there
in roughly the end of October of 1972,
and I go there to start a program
that automated manufacturing, which the Congress later kills.
VINT CERF: They're back at it again, though.
I mean, automated manufacturing.
WALTER ISAACSON: Or killing programs.
VINT CERF: Have you seen how the Tesla-- anyway, go ahead.
BOB KAHN: So when they kill the program,
I'm thinking, all right, what I signed up for here,
I'm going to go back to MIT, where I had a standing
offer to go join the faculty again
because I had started my career out there after getting
a doctorate from Princeton and before going
to BBN on a leave of absence.
So I was thinking.
Larry Roberts, who's running the office, persuades me, stay on.
You've been so involved in networking, don't do that.
So I agree to give it a shot, but only if I can have nothing
to do with the ARPANET because I did that before.
I wanted to do something new.
VINT CERF: Well, you would have been constrained,
in some sense, because you were at BBN
and they were still a contractor,
and you can't manage them as a program manager
after having been at BBN--
BOB KAHN: This was before we had conflict of interest rules.
I mean, I got my own ability to manage conflict, so not
a problem for me.
VINT CERF: No conflict, no interest.
BOB KAHN: Not a problem for me.
But I finally said, all right, I'll give it a shot.
I started the packet radio program after I got to DARPA,
and I started the packet satellite program as a program.
Now, let me just say, Larry had started
to fund BBN to turn a satellite IMP into one that
could use a satellite as a modem relay.
And one of the issues that I had to deal with was, was
that a good idea or not?
I thought, no, it wasn't.
Let's make the satellite net a separate net
by splitting it out, and this was
an interesting technical discussion we had with BBN,
because they were of the opinion that this would be much more
reliable if everything was in an IMP,
but that meant you could never do the equivalent of putting
a router in between it.
You'd just lose complete control because everything
was memory to memory transfers.
So we ended up building a separate net
on a satellite net.
When Larry funded them, there were no domestic satellites
at all.
There was no plan to use a satellite anywhere.
And so I ended up negotiating with COMSAT,
and then an Intel sat to actually set up
tariffs that would enable this to happen,
and then we set up the program.
The one who ran that program technically was a fellow
named Irwin Jacobs, who later became the founder of Qualcom.
I was sort of chief architect of the packet radio program, which
became essentially the forerunner of today's CVMA
technology.
WALTER ISAACSON: The point is that you have these
all of these networks and they don't interconnect.
BOB KAHN: These three networks show up,
and that was what motivated me to work on the problem
of inter-networking them, and I had some architectural--
WALTER ISAACSON: Let me pause right there because when I
heard the word "inter-networking," I said,
I get it.
That's where internet comes from.
Is that right?
BOB KAHN: That's where it came from.
And so I had that as a goal.
I had some architectural principles.
We knew about putting-- we didn't call them routers then,
we called them gateways.
But I came from a communications background,
and so I was very comfortable with building things that
could move the bits around, but I was less comfortable
what the computers would do with them when they got that.
And Vint and I had collaborated a lot
on the ARPANET side of things, me on the communications part,
him on the computing part.
He knew more about how you embed these protocols
into the end computers.
I approached them at one point and said,
look, here's what I'm trying to do
and here's a whole architectural approach.
Here's what I really don't know.
I don't know about what it will take
in these different machines.
We agreed to collaborate, and based on that collaboration,
I think both parts of it came out better.
We were able to improve the communications part,
we were able to improve the end part.
I've had many collaborations, but the one with Vint
was probably one of the best, and we've
stayed pretty good friends over the whole time.
VINT CERF: In spite of that.
BOB KAHN: In spite of everything, right.
WALTER ISAACSON: That really does
show that this was a partnership,
this was collaboration.
BOB KAHN: Bit that's where it came from.
It came from the need to connect all these different nets.
WALTER ISAACSON: So explain how you turned that into TCP/IP.
BOB KAHN: Well, we did it together.
VINT CERF: He and I spent six months trying
to sort all this out, and eventually, this notion
of gateway-- the system sort of emerged out of requirements.
You have a bunch of networks and we
weren't allowed to change any of them,
and the question is, well, if they
can't know that they're part of a distributed system,
how do we link them together?
The conclusion was you needed something
in between that knew how to talk to both networks
and would glue them together even though they
didn't know they were connected.
Then the other problem is that the hosts at the edges
needed to be able to say something like, send this
to another network, but there was no language for that
because--
WALTER ISAACSON: It was like putting a postcard
in another envelope?
VINT CERF: Well, the first problem
is language because the network that you were connected
to locally had no vocabulary for even understanding,
take this thing and send it to some other net.
Every net thought it was the only net in the world.
BOB KAHN: I have to tell you what addressing
was like in the ARPANET, 16-bit addresses.
When you sent something, you said, essentially,
send it to this wire.
Imagine that every wire are on the network was numbered.
So you say, send this to wire 17,
and what was supposed to be attached to wire 17
was a big computer, so that worked fine on the ARPANET.
But if wire 17 then went to some other net, which
had a whole bunch of computers, which of those computers
was it for?
There was no way that language could describe it.
VINT CERF: You couldn't say that,
so we had to invent the internet protocol addressing.
BOB KAHN: That's where IP addresses came from.
VINT CERF: And then we realized that, in addition to that,
of course, we had to make sure that all the data that we sent
in these little postcard size pieces
would be received and reassembled at the other end,
and we knew that some of the networks,
like the packet radio net and the ethernet, would be lossy.
There would be radio shadow, there
would be jamming and all kinds of things,
so we had to, on an end to end basis,
make sure that the communication was reliable.
WALTER ISAACSON: So how did you do that with the transmission
control protocol?
VINT CERF: Transmission control protocol.
Actually, the first papers that we wrote had only one protocol.
It was called TCP, and it did everything.
BOB KAHN: It had IP integrated inside.
VINT CERF: It solved the addressing problem.
It solved the end to end retransmission and filtering
out of duplicates, flow control, and reordering,
and everything was all part of the one protocol.
But after a few iterations, finding some bugs
in the design, several people beat us over the head and said,
we really need a way of delivering things really,
really quickly even if not everything gets there.
So we split the internet protocol out
from the TCP part, the end to end reliable part,
and then created something called User Datagram Protocol
so we could handle real time traffic.
WALTER ISAACSON: Let me ask you a question I didn't
ask for the book, which is, if you had to do it all over
again knowing what you know now, what would you
have designed differently?
BOB KAHN: I would have started with the digital logic
architecture that I've been working on for the last 20
years, but that's not fair because it's like saying,
how would radio communications in this country
have evolved if the semiconductor
chip was available in 1915?
You can't run that Gedankenexperiment anymore.
VINT CERF: But there are things that I
wish that we could have done that we didn't,
like I wish that we had picked 128-bit address space instead
of a 32-bit address space.
We wouldn't have to go through this pain
of the IPv6 transition.
BOB KAHN: What makes you think 128 bits is enough?
VINT CERF: Well, it might not be,
but it'll be enough until after we're dead.
Then it doesn't matter.
WALTER ISAACSON: Let me talk about more consumer
friendly things, rather than addressable space, which
is things like, is there too much anonymity built
into the internet?
Would we be better off if people were a little bit more
responsible for what they put up?
BOB KAHN: I think that identity in the internet
is really important.
I would like to see us come to the point
where everybody has unique and persistent identifiers that
resolve to reasonable state information about them,
and they can have multiple identities,
depending on the role they want to play.
And one of those roles we must absolutely support, I think,
is the role of anonymous.
WALTER ISAACSON: So you'd have an anonymous, pseudonymous
thing.
BOB KAHN: You can decide how you want to participate.
If I want to participate as a government employee,
I'm probably going to give a government employee identifier.
If I'm going to participate as a representative of my company,
that's a different one.
And then maybe the services and the opportunities you get
are determined by how definitive your identity is.
WALTER ISAACSON: But in other words,
you could create an internet in which there's secure ID,
and when I'm dealing with my bank,
I would have to use that identity when
I'm dealing with them.
BOB KAHN: Well, if this were back in 1972 and Vint
and I were doing it, I would say, yes, we could do that.
Today, I don't think anybody can do that anymore
because there's so many places and parties around the globe.
We can work toward it.
VINT CERF: I actually think it's possible to do some things,
to create strongly authenticated identities,
and also to achieve some anonymity,
but we're running out of time, it turns out.
BOB KAHN: But Vint, I agree with that.
I was harping on his question about you.
You could do it.
VINT CERF: We've got to let these other folks ask
questions, too.
WALTER ISAACSON: Having three moderators
is always hard, right?
VINT CERF: So I've got one question.
Patrice?
And then anybody else, get your hands up.
BOB KAHN: This is also my wife.
I saw her come in.
AUDIENCE: I'm going to throw a softball.
I've been sitting here looking at your title, "Innovators."
I started life out years ago as a copyright attorney.
I'm very much interested in the creative process,
and I've heard a lot tonight about collaboration
as if the spark of innovation doesn't quite matter anymore,
that it's some kind of muddled group, they get together,
and they come up with the idea.
Now, I've worked with Bob and Vint over the years,
and I was listening tonight to their discussion
of the early days, which that's an interesting story,
but they each brought their own innovation to it
so you could have a finer level of granularity.
So you can have a collaboration, but if you
have a bunch of dullards and one really bright card, then
the collaboration really is a useless concept.
You have to look at the innovation, what
is the creative force?
I'm trying to nuance a bit because I
heard the drumbeat of collaboration a bit much.
WALTER ISAACSON: That's why it's a long book in a way
because a lot of collaborative teams
don't work particularly well.
I loved both the RFC process, and then
what you all talk about, about the TCP/IP process, which
is two people each bringing, you're
slightly more of a communications mindset,
whatever, but it ends up being a collaboration where
innovation really happens.
Now, by the way, that happens with Steve Jobs
and Steve Wozniak.
They are bringing different things to the party.
It certainly happened with Andy Grove and Bob Noyce.
One's a congregationalist who loves singing madrigal songs
in a congregational church, and the other
had to swim to escape the Nazis and the Communists.
I even tried to do that with Benjamin Franklin,
because he was not the smartest of the founders--
I mean, Jefferson, Madison, and all--
but he was the one who was the glue who could bring together
a Washington who's indispensable,
passionate people like John and Sam Adams,
smart people like Madison and Jefferson.
It's almost like a baseball team.
You need a pitcher, a catcher, you need the manager,
you also need utility infielders.
What makes for a collaboration that has both a visionary
but also executors?
BOB KAHN: Walter, who was the smartest of the founders?
WALTER ISAACSON: The smartest of the founders
may have in-- I don't know.
It's a good question, and "smart" means
many different things.
I think Franklin was really smart
because he knew how to bring smart people together,
which is a skill.
VINT CERF: Yeah, but he was a twit.
WALTER ISAACSON: I think Madison was intellectually-- what?
VINT CERF: He was a twit.
WALTER ISAACSON: A twit?
Franklin?
VINT CERF: Yeah, Franklin.
Well, I read your book.
You spelled it out.
WALTER ISAACSON: John Adams was a twit, not Franklin.
I think Madison was the most creative mind there.
Obviously, Jefferson is very fertile mind,
but we don't have to get into a debate on the founders,
as much as I have my problems with Jefferson that
go to the limits of intelligence.
VINT CERF: Let's let the audience ask questions.
WALTER ISAACSON: Susan?
AUDIENCE: No, no.
Who else has a question here?
WALTER ISAACSON: You're the moderator.
I shouldn't be able to pick out people.
AUDIENCE: I'm Alex Howard.
I've enjoyed reading your books, following along, and using
the internet for some time.
I couldn't help noticing as I opened this up that you have
Ken Kesey in there, and Stewart Brand, who you mention.
If you look at the people in this book,
it's an American and British story in a lot of ways.
Since we're in DC and people think about public policy
here a lot, how you create the conditions for innovation,
if you look at that story, how do you do that now,
as people are thinking about national competitiveness,
about education, about being inclusive,
because STEM is a big issue?
The impact of you writing about this
has caused quite a stir, too, to recall the limit of ENIAC.
It's important that the original computers were women.
When you think about the conditions for innovation,
what have you learned from that, and what have you
all seen, too, in terms of the things being focused on?
BOB KAHN: This is not a question about genetics, right?
AUDIENCE: Absolutely not.
No, it's a condition about how we create the next big thing.
WALTER ISAACSON: But it is a question
about the primordial stew from which the DNA emerges.
VINT CERF: So at Google, I think the most important ingredient
for innovation is permission to fail
and the willingness to let you try crazy ideas.
The founders are willfully determined
that we will push the envelope.
That's why we have a Google X and we've
got a moonshot experiments and things
that, like self driving cars.
I think this attitude also shows up in the Silicon Valley,
where venture capital is willing to accept risk.
They know that maybe only 10% of the things they invest in
will actually work out.
I don't think it's a peculiarly American attitude,
but it is an American attitude, and I
don't think we've lost that.
I think we still have it and we should use it.
BOB KAHN: And that's also true in DARPA to a large extent.
VINT CERF: Oh, absolutely.
BOB KAHN: They shoot high and are willing to fail,
even though they don't fail too often because they often
don't shoot as high as they start out shooting for.
WALTER ISAACSON: I'd like to ask you
a question on that, which is--
VINT CERF: They're supposed to be asking questions.
WALTER ISAACSON: Oh, sorry.
BOB KAHN: A lot of success stories.
WALTER ISAACSON: Well, I'll just make a quick statement
that I think we have blown it a bit in government
because it used to be you were allowed to fail, and now
any little gaffe you make in this day and age,
you'll never get confirmed, you'll get run out of town.
In Silicon Valley, at least, or in other innovative places,
you're allowed to fail.
I have Ken Kesey in the book because it
was a rebellious attitude that happens
in the late '60s and early '70s in the Bay Area that involves
"The Electric Kool-Aid Acid Test"
and Stewart Brand, and also Doug Engelbart, as you know,
with the mother of all demos.
It's almost like one of the electric Kool-Aid acid tests.
It has Brand as the impresario.
So part of that cultural stew was
the rebellious counterculture, anti-war, anti-authoritarian,
hippie, acid test mentality that has been written about
by Fred Turner very well, by Steve Levy, John Markoff.
Many people have written about that hacker mentality,
and it ties into the rebellious as it allows you to fail.
VINT CERF: Let's get two or three questions
asked, and then we'll try to respond
to them as a way of getting more.
Let's start over here.
AUDIENCE: Hi, Jim.
On the theme of collaboration, with the internet
hopefully going to be accessible by the next four billion people
on the earth's surface and more than half the population,
in terms of the global innovator community
and people's ability to solve their own problems,
what would you like the internet to help do in the future
to solve global problems, to connect people to collaborate
on problems that seemingly are insurmountable?
VINT CERF: So global collaboration,
what can the internet do?
Let's get a couple more.
AUDIENCE: I really love "Steve Jobs," the book, rather,
and I'd like to ask Mr. Isaacson, what
are your impressions of innovation
at Apple after the passing of Mr. Jobs,
because it seems to highlight that tension
between a strong singular visionary and then
a collaborative team?
I think there is a storyline that there's
been some drift since you lost that visionary.
VINT CERF: OK.
So we've got collaboration at Apple
and we have global collaboration over the internet.
One more if there are?
If not, we'll try to tackle those two.
There's one more here.
AUDIENCE: So Mr. Isaacson, how did you
pick who the four people were going to be on your book cover?
VINT CERF: All right.
AUDIENCE: I appreciate that there's a woman up there.
VINT CERF: I'm pretty sure we're not
going to be able to integrate one answer to all that.
Let's start with the last one, book cover.
How did you do that?
WALTER ISAACSON: Boy, it was very
hard to do this book cover.
I'm used to writing biographies.
The Steve Jobs cover was probably the most iconic cover,
Steve Jobs and a picture, so that's the easiest one.
When you're doing something that involves, say, 12 grand steps
and 30, 40 innovators, how do you pick them?
In the end, it's Alan Turing, if you're wondering,
who's on the bottom right.
You know the others, Ada Lovelace.
VINT CERF: They're there the bookends,
in some sense, and then the rich guys.
AUDIENCE: [INAUDIBLE] that's what I saw.
WALTER ISAACSON: I think that Steve Jobs and Bill
Gates are very emblematic characters
of how you both have vision.
Also, you execute, and business execution is part of it,
and collaboration.
If I'd had 10 people, I could have put 10.
It was not a particularly easy cover to do.
If you go back and you can find, Google things
that are cached from earlier times
on Amazon, there's probably eight
covers that they went through before they got to this one.
VINT CERF: Actually, maybe you should
consider publishing a book with multiple covers.
WALTER ISAACSON: Yeah.
At Time Magazine, I used to do that.
It was a real gimmicky, but especially every now and then,
you'd have four covers for four different--
VINT CERF: If collaboration was the intent,
I don't see collaborators on the cover at all.
Anyway, just a gripe.
WALTER ISAACSON: Turing.
Turing was a collaborator.
VINT CERF: Well, mumble.
OK, so let's go back to this collaboration question.
You've got this question, how do you use the internet
to help solve global problems?
BOB KAHN: You want me to take that one on?
VINT CERF: Go ahead.
BOB KAHN: I would let's say you're
thinking of the internet writ large now.
When I think of the internet, I think about it
in terms of the protocols and the procedures
that let you connect things.
That gets back to the definition of what it is,
and then you can use it for different things.
So it really gets into how people
are going to get together to solve some
of these global problems and what role
the internet can play in it.
You asked the question the other way around,
what can the internet do for that,
and I would say that's the wrong way to get at it.
The question is, what global problem
are you trying to solve?
Is it clean water in parts of the world that
don't have access to clean water?
Is it how to get rid of Ebola viruses?
Is it how to do checkout of people
in airports when they're coming in?
Is it how to deal with better targeting
in the field for the military?
You can figure out what problem you're concerned about
and then you can figure out what a solution is,
and then you can ask, is these an internet role
in all of that?
And most likely, there will be because everything is disparate
and disconnected and you've got to bring them all together.
I wouldn't phrase it exactly the way
you did, starting by asking about what
the internet can do for you.
It's the Kennedy thing flipped around.
VINT CERF: Ask not what the internet can do for you.
Ask what you can do for the internet.
I actually want to draw out of this another notion,
and that's the notion of an enabling technology,
because if you look at these stories,
all of them, what you see is this scheme of things
linking the stories together.
Somebody creates something, which
enables someone else to do another thing.
This whole story is all about enabling,
which I think is part of the origin of your question.
What could the internet enable?
The thing it does better anything
is it connects stuff together that
was never connected before.
This casual idea that I can take this can of Coke, for example,
or this book, put a chip in it, and connect it to the internet
somehow is the beginning of ideas
that people never thought of before
because it wasn't possible.
Now it's possible.
BOB KAHN: Maybe you don't connect it to it.
Maybe it's automatically part of it when you put the chip in.
VINT CERF: Well, if we ever get to the point
where we can have self assembling devices
through biotechnology, then when people say, what do you think,
computers grow on trees?
And you say, yeah, let me show you.
Let's go back to the collaboration at Apple,
though, because I think that was a specific question for you
to try to tackle.
WALTER ISAACSON: Yeah.
I mean, when I first started writing about Steve Jobs,
I thought he was an exception to the rule about collaboration,
that he was a singular visionary and his visionary did things.
And I asked him at one point, what
was your most visionary thing that you created,
and I thought he might say the iPhone or iPad or the Mac.
And he said, no, creating a particular product
or having a vision for it is hard,
but the really hard thing is creating a company that
lasts in which people all know how
to execute with vision on various things.
So the most important thing I ever created
was Apple because I realized it wasn't about me.
This is him talking.
I realized it was about a team of people where I could walk in
and say, this sucks, or this is great, or whatever,
but eventually, out of that sturm und drang
would become a particular product.
And I think the most important thing,
and it's probably a [INAUDIBLE] as here,
whether it's Google or Apple or whatever
else, is to look at the companies in which ingrained
in the genetic code was a willingness to take risk
and to fail, and importance about connecting beauty and art
and humanity, making our machines more personal
so we felt intimate with them.
There were a lot of companies that came up with great ideas--
Xerox at Xerox PARC, Bell at Bell Labs, IBM, whatever.
They didn't innovate as a team and take risks.
I see Google doing that.
I'm not just saying it because I'm here with you all.
Google is one of the few really big companies in which it's
still like, let's shoot for the moon,
let's try some weird things, and let's understand
the importance of teamwork.
I think Apple does that.
Apple still has a team.
What Steve did was not just create a company.
People read my book and they say,
Steve was pretty much of a jerk at times, wasn't he?
I say, yeah, but don't forget how it turns out.
He has the most loyal people around him
who will march through walls with him, who will collaborate,
who will stand up to him.
That team at Apple is a lot tighter
than the teams at much friendlier companies like HP.
BOB KAHN: Walter, one of the things
that doesn't show up in your book,
and yet one is often drawn to a comparison between Apple
and Oracle, is Larry Ellison.
Is he out of the picture here because you just
didn't have time to get to him, or do you
think he doesn't play a role here?
WALTER ISAACSON: I don't know.
You tell me.
BOB KAHN: No, you're the one who wrote the book.
Why did you leave him out?
VINT CERF: Why don't you [INAUDIBLE]?
WALTER ISAACSON: I don't know that Oracle
was one of the 12 great leaps of the digital age.
I think Larry Ellison's a brilliant person,
but I don't find that the creation of the Oracle systems
was up there with the transistor, microchip,
the personal computer.
If I may, I would love to get back--
BOB KAHN: I was just curious as to your view.
WALTER ISAACSON: So what Steve is able to do
is really show how you get every player on a team,
somebody like a Tim Cook, who knows
how to make things happen, somebody like a Johnny
Ives, who understands the design,
and he creates this tight team.
I think at Google, too, there is this ability,
when you watch Larry and Sergey, but then
all the people around them and all the people here.
We can take risks, we can take time off, we can collaborate,
and I think that the most important thing for innovation
is not creating the next new product
but creating the Apples, the Googles, the Amazons, whatever
it may be, who are always trying to do different things.
VINT CERF: I'm sorry.
It turns out that this thing says 0S,
and I'm pretty sure that doesn't mean osmium.
WALTER ISAACSON: It's either zero seconds
or operating system.
VINT CERF: Right.
Or something.
So I have a suggestion.
Why don't we finish up here?
We have all kinds of refreshments
out the doors here.
Why don't we thank our author over here, Walter.
[APPLAUSE]
WALTER ISAACSON: And let me make sure we
thank the two great heroes of the theme of this book,
but also of the digital age, Bob and Vint.
Thank you all.
[APPLAUSE]