>>Jonathan Rosenberg: Thanks everyone for
coming. So, before I introduce our distinguished guest, we actually have a couple of additional
guests that I thought I would mention. So, we have Steven Zierk and his mother, Marina.
Actually, I'm not sure his mother's name really matters because from this moment on, you're
gonna be known as Steven Zierk's mother, for the rest of your life. So, Steven has just
returned from Greece, where he won the World Youth Chess Championship sponsored by the
International Chess Federation. [applause] Turns out that Steven is actually a senior
at Los Gatos High and took classes with my son, my nephew and I'm sure is looking out
for my younger daughter, Hannah, who is a sophomore. So, Steven, thank you for joining
us. You can sit down. [laughter] So, back in March, I read an article by Garry
Kasparov in the New York Times Review of Books and it was a review of this book called "Chess
Metaphors," by the Spanish author, Diego Rasskin-Gutman, and I'm passionate about issues related to
innovation, as most of you know, and a bunch of Garry's comments in the article basically
focused on the poor state of innovation in computer chess. And one of the things that
he observed was that because brute force programs played chess so well, there was little innovation
on any other kind of approaches. And another one of the things in the article that many
of you guys mentioned was that it talked about how mixed human plus computer teams could
compete in chess and the outcome of one of these tournaments was that amateur players,
with basically garden-variety PCs, won against much more formidable opponents. So, it turned
out, if you read the article, that weak humans plus machines plus very, very good process
trump a strong computer by itself, but even more surprisingly, a strong, even more surprisingly,
a strong human plus a computer with not as good process, which is a compelling argument
for very good UIs I think. But there was lots of applicability, I thought, in this article
to some of the things that we think about at Google and many people suggested that we
should bring Garry here, so, we did. If you don't already know, Garry Kasparov became
the youngest undisputed world chess champion ever when he won the title at age 22, in 1985.
Steve, you have four more years. [chuckles] So, 20 years later, he retired from chess
as the highest rated player in the world; many people consider him the greatest player
to ever play the game and he's a widely published author on a whole bunch of different subjects,
from leadership, on leadership and strategic thinking and he's a significant player and
active in Russian politics. So, I thought we'd get started with Udi. I think we have
a slide. Udi-- [laughter] played a simultaneous match with Garry in
2005. I suppose I could ask Garry what the outcome of the match was, but I don't think
it was particularly remarkable. [laughter] So, Udi, I'll just ask you. How many moves
did you last? [laughter] >>Udi Manber: I actually lasted 23 moves,
which was a lot more than I expected. >>Jonathan Rosenberg: Woohoo! [clapping] Where did you play? >>Udi Manber: This was in San Francisco. >>Jonathan Rosenberg: In San Francis-- >>Udi Manber: In San Francisco, yes, the,
it was right after the-- >>Jonathan Rosenberg: So, I know Udi has some
questions, which I thought maybe we could start with -- >>Garry Kasparov: About the game? [Udi and Garry laugh] >>Udi Manber: Actually, I can tell you a story
about this game. >>Jonathan Rosenberg: Ok. >>Udi Manber: I think you might enjoy this-- >>Garryi: Go ahead, [ ]. >>Udi Manber: So, when somebody asked me just
before we played-- this was a simultaneous game-- somebody asked me what do I think my
odds are. I thought about it for a second, I said, "Well, I think about a billion to
one." And he said, "No, you got it wrong. You'll have to make every move exactly perfect
for about 50 moves and even if you know chess, it'll be about five choices per move, so it's
five to the fifty. It's way over a billion." [laughter] And I said, "No, no, no, no. You got it all
wrong. That's not how I figured it out." I was trying to figure out in my head, what
are the odds of the chandelier falling on his head in the middle of the game? [laughter] The odds are, the odds are much better that
way. So then, the game was played in The Palace Hotel in San Francisco and we enter the game.
You can see there -- it's a really big room, huge, high ceiling with a huge chandelier
right in the middle. [presenter laughs] So, during the game I had to-- >>Garry Kasparov: There was no earthquake
that day. >>Udi Manber: That's right. So, luckily, the
odds were right and the chandelier is still there, but I had to avoid from laughing and
looking up through the game. So, that's my story; I'm sticking to it. >>Garry Kasparov: Just to follow this, the
subject you touched is briefly about the numbers. >>Udi Manber: Ok. >>Garry Kasparov: And the odds. Say, this,
this, when we talk about chess and computers, the general crowd always gets confused, saying,
"Wow. It's a chess, with the game, 64 squares, 32 pieces. So, how long does it take for machine
to, to finish the game off like it happen with checkers to solve the game?" So, just
people don't understand that if, I don't know whether we can calculate the number, but according
to the so-called Shannon paper, published in 1950, the number of all legal moves in
the game of chess is ten to the power of 120. It's quite a big number. It says, it's considered
is the another tentative number of ten to the power 78, which is the number of atoms
in the observable universe. So, that's why chess so far is safe from being cracked by
the computers. But, of course, it's the, so the Moore's Law was good enough to guarantee
that in the, from the practical terms, chess was solved by computers in less than 50 years
because machines now are probably beating every player; capable of beating every player. >>Udi Manber: I was about to say, when it
come to Google, they say things like this, you just present a challenge. [laughter] >>Garry Kasparov: Challenge at what? >>Udi Manber: To do that. >>Garry Kasparov: To do that? So, which number
you have in mind? >>Udi Manber: The ten to the 120. [Garry laughs] >>Garry Kasparov: It's good, you know? >>Jonathan Rosenberg: Maybe we'll go with
both questions from the audience and if you guys can give us the Dory questions, we'll
try to incorporate as many of those. Udi and I also have questions, but if we could flip
here to the Dory, that would-- >>Udi Manber: Lemme, lemme start with -- >>Jonathan Rosenberg: but I know Udi's got
some questions to start with. >>Udi Manber: My questions. >>Garry Kasparov: Go, Udi. >>Udi Manber: So, obviously in chess-- >>Garry Kasparov: Nice start. >>Udi Manber: Wanna do this? >Garry: Nice start. [Garry laughs] Yeah, by the way-- >>Udi Manber: Ok. >>Garry Kasparov: that's a very good question
because it's-- >>Jonathan Rosenberg: Our tradition is that
we read the question-- >>Garry Kasparov: Ok, fine. >>Jonathan Rosenberg: so that they can all
see it. The first question, by the way, the first thing Garry said just to show you Udi
and I have worked hard to prepare some brilliant questions, he's a fan of James Serwicki and
when we told him about Dory, he immediately said, "Well, that must be better than whatever
the two of you think." So, he likes the first question better than ours in any case. The
first question is, "Can you explain what made you claim that game two move 37, presumably-- >>Garry Kasparov: Yeah. We shift to E4. >>Jonathan Rosenberg: "was shift to E4, was
a move that a computer couldn't have made? Do you still believe IBM cheated?" >>Garry Kasparov: No, it's-- [laughter] It's -- I say, they say, but we can probably
try to reach a scientific conclusion in measuring our statements. The only problem is that at
the end of the match, they dismantled Deep Blue and I said, "They killed the only partial
witness." [laughter] It's actually related to the subject that
you touched, it's from in my article, about man versus machine; an actual man plus machine.
And when I, when we talked about this game--game two-- and I made a claim and I'm still behind
this claim that there was some kind of artificial interference that time human interference.
People ask me, "What did you mean that, was it Anatoly Karpov, a very strong player, behind
the computer?" Which is, it just showed us the crowd didn't actually understand the whole
nature of possible interference. Because when we have a game, say, game between two computers,
one or two advices from humans could change everything because you don't a very strong
player actually to help machine, to guide machine. And just to give this human judgment,
intuitive judgment, strategic judgment, to make sure that machine doesn't go wrong way.
And other way around. Say you have two humans playing; one or two advices--quote/unquote--
from the computer, could be, could be decisive. Even a hint that there is the combination
there, so because very often, one would play, we don't know whether the, it, something that
is tempting is going to work. I remember I played it in, long time ago in 1996, at a
very big tournament. I played against Viswanathan, the current World Champion, and I thought
I was winning and I saw the combination, but it was too long and I just calculated for
about half an hour; just move seven, move eight. I smell that there was a blood there,
but I, eventually I decided to go for a safer route, also keeping advantage and then after
the game, my coach asked me why did I do that? I said, "Look, I couldn't see it." Then he
said, "Wow. The machine, even in '96, already showed that it was winning." So, just a hint
that it was the right way to go, could it force me to go deeper and to find a solution?
Now, what's happening in game two and why I made this claim is that Deep Blue, again,
from what we know about this computer and the problem is that there's not enough information
to analyze because, as I said, they dismantled it and all scientific tests made later, when
they put the games I played against Deep Blue to be analyzed by other computers--modern
computers--much weaker computers in term of brute force of calculation, but was more chess
understanding, showed that Deep Blue was not exceptional. So, the machines today, they
are doing better by going through, through all these games. And to make a very solid
positional move instead of winning three points, gaining decisive material advantage was highly
unusual for computer of that strength. What I wanted to see, I wanted to see the logs.
So, they never produced the logs and it's actually, same things happen in game five.
But again, ok, it was my mistake. I didn't put enough pressure to secure the integrity
of the game and recently, by the way, there was an interesting interview of one of the
"seconds" – quote, unquote, seconds, assistants of Deep Blue, Spanish Grand Master Miguel
Illescas for, in the chess magazine, New In Chess, the most popular chess magazine, and
some of the stories he told about this match, they were again, could be used as a consequential
evidence to support my, my theory. Even the fact that upon the request of the team, the
IBM team, the IBM corporates changed the, the guards next to my green room, making sure
that these guards speaks Russian to communicate all information they could gather for my conversation
with my coach. Ok, consequential evidence. >>Jonathan Rosenberg: So, just following up
on that, if we assume that IBM did cheat,-- >>Garry Kasparov: No, IBM wanted, let's, let's,
let's, let's, let's, let's, let's use in a diplomatic statement-- >>Jonathan Rosenberg: Assume there was a human,
there was a human-- >>Garry Kasparov: it's for IBM-- >>Jonathan Rosenberg: Oh, you're a, you're
a diplomat now, I forgot. >>Garry Kasparov: we need, we need, I'm not
a diplomat, you know. >>Jonathan Rosenberg: I'm not a diplomat,
so I was gonna posit that they cheated. >>Garry Kasparov: Yeah, but, I don't want
group of lawyers too often, yeah. I, they didn't even hide the fact that for them it
was about winning or losing. So, they did everything to win. And certain thing that
happened in the match, even the fact is that they were very following the story of Miguel
Illescas, they even knew the opening I would play in the last game. Yeah, ok. By accident,
probably, considering the fact that I never played it before in my life and never after. [laughter] Yeah, it's, IBM shares jumped 22 percent during
two weeks, so that was a big-- >>Jonathan Rosenberg: They got what they wanted. >>Udi Manber: They did? >>Garry Kasparov: Ok, that's happened, those
are two. They wanted to win badly and for me, it was more of the scientific or social
experiment and I expected this to be, to continue, and I asked for the rubber match because I
won the first match in Philadelphia. They won the second match and as I said, they immediately
dismantled the computer. That was a really bad sign. >>Jonathan Rosenberg: So, what kind of things
would tell you, if you were engaged in a match, that would be a clear tell that a human had
become involved? Intuitively, what are the types of things that would actually just,
almost be a smoking gun to you that would say that-- >>Garry Kasparov: It's the, yeah, but, again,
let's do the opposite now. In a big chess playing website now, for instance, a playchess.com,
you have this anti-cheating system. So, the computer actually looking for games played
by some players and the moment it sees too many games made by computers or, not made,
but potentially made by the computer, it assumes that it is cheating. So, it's, as for the
game, game two or other games I played with the computer or computers playing each other,
you could, if you know enough about the machine, if you know enough about the decision making
algorithm and about the priorities, because every machine at the end of the day has to
operate with a preordained system of priorities. For instance, it's all about numbers. If you
put sort of the very high numbers on king security--king safety-- so, it will stay the
same way. You cannot, machine cannot regroup these priorities based on the care of the
position. So, if you can understand, more or less, the concept behind this decision
making algorithm, you will feel immediately that there is the, there was an interference.
And ten years ago, I said I would need 15 to 20 games to identify the playing program,
whether it was Fritz or Junior or another one. Now, machines are getting sort of more
sophisticated. There is sometimes some kind of intuitive element in the game because they're
more flexibility in sacrificing the material, actually accepting the fact that other factors
could trump this material advantage, but in, back in 1997, it was impossible. So, machine
turning down very generous offer to win huge material for positional, to, for a move to
increase positional pressure; I didn't believe that. >>Jonathan Rosenberg: Hmm. >>Udi Manber: Actually, Ken Thompson is here
in the audience. >>Garry Kasparov: Yeah. >>Udi Manber: And I remember-- >>Garry Kasparov: Ken was watching the screen. >>Udi Manber: hearing from him-- >>Garry Kasparov: I saw him. >>Udi Manber: that he had some inside information.
I don't know if you want to share with us later about that. [laughter] Ken was one of the judges of that, that game.
I actually wanted to ask you, lemme take one question, a general question about decision
making. Or you rather go with? >>Garry Kasparov: No, no. It's up to you.
I'm a guest. [laughter] >>Udi Manber: And, by the way, there's very
simple way to add intuition to computer programs and I've learned that a long time ago that
you just introduce bugs. [laughter] I'm very good at that, too. So, my question
really is that part of being a really good player is not just making the right moves,
but knowing how much time to spend thinking about every move. And that's one thing we
don't do in business that much. We usually, the time we spend is proportional to the importance
of the decision. >>Garry Kasparov: How do you know the importance
of the decision? >>Udi Manber: Estimate. But not even that,
I'm sure that's not even been decided yet. >>Garry Kasparov: How much time you spend
to estimate, to make the correct estimate? [laughter] >>Udi Manber: So, my question to you exactly
is what should we do? How much time should we spend and how do we decide how much time
to spend on a decision? Cause that's what you do in every game. >>Garry Kasparov: In chess, we are, I'm retired,
so my chess experience is five years old. So, I don't play professional chess now. But
in chess, you are under time constraints. So, you have whatever, two hours for 40 moves
or half an hour for the whole game, so you know exactly your, the time allowed for you
to make all the decisions. So, you have to adjust. Obviously, if you play rapid chess
or bullet chess-- >>Udi Manber: Sure. >>Garry Kasparov: you have to understand that
you sacrifice the quality of your decisions to make sure that you are not losing on time.
So, at the end of the day, as I already in my book, "How Life Imitates Chess", every
decision, whether it's made at the chess board, in business, at home, in the White House,
its all, every decision contains components of material, time, and quality. And while
material and time is understandable, then the quality is a factor that we all have to
adjust to our decision making system. When one can become comfortable with micromanagement
or just looking at the big picture, at the end of the day, it's all unique. So, I don't
think that there is the universal advice for anybody who wants to make an intelligent decision.
Some people should do it quicker because if they spend more time, they going around. So,
some people should take more time because they need to look at every element. It's absolutely
unique as DNA or fingerprints. >>Jonathan Rosenberg: Ok. So, it looks like
all the Googlers wanna get better at chess, so the next question is, "What is the best
way for somebody who does not quite have the time to devote their life to chess, to improve
their game?" >>Garry Kasparov: Now, the counter question
is if you don't have time to devote your life for chess, why do you want to improve the
game? [laughter] So, it's just to beat your neighbor, to impress
someone or to catch up with your kids. So, now with computers, I don't know how much
time you can allow to play chess while you're here, so but definitely on, with the computer
technologies, you can do a lot, even without reading books. But, I think that's always
the limitations because so many kids now and so many people who have more time to spend
on chess and it will be quite a challenge. >>Jonathan Rosenberg: Ok, let's go to one
in the audience here. >>Udi Manber: Frank? >>member #1: So, you were the youngest World
Chess Champion, too. >>Garry Kasparov: I was. >>member #1: But you couldn't be the World
Chess Champion at 21, right? Can you talk about the politics around what happened with
that match where you didn't become? >>Garry Kasparov: You mean 1984, yeah? It
was long time ago and he's definitely, you, you're somewhat of a chess expert because
you, I guess you read enough books to remember this match. How old are you? >>Jonathan Rosenberg: You weren't born. [laughter] >>member #1: Yes. >>Garry Kasparov: How old are you? >>member #1: 29 >>Garry Kasparov: 29-- >>Jonathan Rosenberg: You were born. >>Garry Kasparov: Just ok, just-- [laughter] Yes, yes, yes. Nursery school. Yeah. I was,
as you remember, I was trailing behind badly; losing five to zero, so Karpov had to win
one more game to retain the title. He couldn't win this game. Eventually, I won three more
games, was catching up, and then the match was stopped and started again seven or eight
months later--eight months later. And there was a clear case of interference by Soviet
officials. I think that they just believed that it would be too much pressure on Karpov
to continue the game. I didn't like it, so the official decision made by International
Chess Federation stated clearly that Karpov agreed; it's part of debate. So, that's, look,
at the end of the day, I won the title and this match was a great lesson for me because
it probably made the ultimate mark on my character. So, trailing bad, so badly, losing beyond
any hope to survive and eventually surviving and beating Karpov eight months later; it
was a proof to me that there's no situation you, in your life where you have to give up.
So, that was a very, very good lesson. So, I'm grateful to Karpov for help me to, to
build up my character. [laughter] >>member #1: Thanks. [applause] >>Udi Manber: Why don't we let-- >>Jonathan Rosenberg: Ok. >>Udi Manber: I wanna take the next question
from the Dory. This is-- >>Garry Kasparov: Nah, but someone is ready
about to go for it. >>Udi Manber: Ok, go for it. >>Jonathan Rosenberg: Please. >>member #2: Great. I'm really curious about
what you would tell us about your learning curve? So, at 23, you're the best player in
the world. Were you better at age 30? Were you better at age 40? When you retired, were
you the best you had ever been? Can you just talk a little bit about, did you ever plateau
and some of the factors around your learning? >>Garry Kasparov: I'm always very cautious
when asked to compare players from different epochs. So, that's – let me start -- I was
-- with general response. Now, you ask me to compare Garry Kasparov with 1985 to 1995
and 2005. I play chess and I, my professional chess career I mean could be divided in probably
in some sections. I'm actually writing now the book about my best games, get Kasparov
and Kasparov, so I'm working now the fourth volume. So, that's probably why I'm too sentimental
to talk about those early days and it ends with the World Championship match in 84-85.
I think that probably I had two peaks in my career. One was the late 80s, 89, when I broke
Fischer's record, the rating, 2785, which was, many people believed it would be sustained
for very long. I crossed 2800 mark and when I, and now not many players did it. I mean,
right now I think there's three or four. But at that time when there were no players in
2700 categories, so only Karpov; that was quite an accomplishment. And then it was probably
inevitable downside in the career because I was, in early 90s, I had to face a younger
generation. And that for me was the biggest challenge because I succeeded in being ahead
of the next generation. That's what probably never happened before and that's what probably
kept Fischer outside of chess because he was not sure he could beat Karpov in 1975. And
so, the second peak was probably 1999-2000, again, that was the, my, year breaking record,
2851, which is still way above the level of the best players. And then I lost the World
Championship match, probably also inevitable because, whether you like it or not, but subconsciously
you become complacent. So, just, even if you understand everything, at certain point, you
lose your determination. So, you want to say, you want to optimize your game, or your business,
and you don't want to make any changes. So, and again, if you don't disrupt your business,
other people do. [laughter] >>Jonathan Rosenberg: Udi, don't get complacent. >>Garry Kasparov: No, no, no, no, and everybody.
That's-- >>Udi Manber: I've learned that. >>Garry Kasparov: That's the problem now.
That's, it happens with everybody, yes. So, and then it was the last challenge for me,
after losing this match, and I don't know if you know this, I haven't lost the match
because I was going down; I lost the match in the middle of my longest winning streak
in a tournament. I won ten tournaments and I was exactly in the middle. So, after losing
the match, I won a few more top tournaments and then, again, probably it's inevitable--maybe
just – it's in the human nature. So, at some point you lose your passion to search
for new horizons. Because for me, chess was not just about winning or losing--so not for
IBM—yes, I wanted to make the difference. I wanted to find something new. I wanted to
make a make a contribution to the game. And at one point I recognized that already I did
more than I could have ever thought of and by 2002, 2003, I was mentally preparing myself
to do something else. >>Jonathan Rosenberg: Well, the next question
relates to how we can do something new and inspirational. "We once believed that if we
could build computers to play chess or pass the Touring test, we would gain valuable insight
into human cognition. We were wrong. Is there a new challenge that can inspire us to create
machines that are truly intelligent?" >>Garry Kasparov: Yeah. Because, because practical
chess was solved by, with simple increase, in brute force of calculation, so there was
no need to look for another path just to make intelligent machines. But I believe that chess
is unique playing field to look for the combination of this, of the machine's power and a humans
abilities. So, it's like to solve the Moravec's paradox because it's exactly what happens.
We are good at something that machines are bad at and vice versa. So, by creating this
the platform for human intuition to be added to machine's brute force of calculation, I
think chess computers, this kind of competition, could offer some assistance to those who are
working on their, on other things. Because I think it's just what's happening with, even
with a search. So, you want to optimize it, but you don't use the visual power of a human.
Let's say you have the-- >>Jonathan Rosenberg: I can't talk about that. [laughter] >>Garry Kasparov: No, no, no, but it's just,
you say it's just the, if you have O'Reilly books--I don't want to use Paris Hilton, which
is, it's too known--but O'Reilly books in the search. So, you will be dealing with two
different elements. So, let's say you have a simple visual above the search where one
is the Fox News logo and one is the computer. So it can immediately, the human takes probably
a split second to separate, so just clicking one button. So, with a machine it takes much
longer. So, in chess that's a good playing field to identify how you can do these things
by adding this, the human insight to the machine's power. >>Jonathan Rosenberg: Right. But if you take
that further and assume humans can bootstrap the algorithm by looking at any one specific
result and saying, "This could be improved by doing X or Y", how do you then scale that
in such a way that you can take those human insights into what the bootstrapping human
would do to the algorithm and then make the algorithm better? >>Garry Kasparov: But it's the, I'm not talking
about the merging it, I'm talking about the interface. So, how you add it because you
know that machines are very good at something, so, but they're not good at recognizing the
visuals. So, and that's one example. And I believe there are many other examples. It's
all; it's more about interface rather than the merge. >>Jonathan Rosenberg: Hmm. Ok. Next question. >>member #3: [Speaks in Russian]. Welcome. >>Garry Kasparov: [Speaks in Russian]. >>member #3: [Speaks in Russian], yes. So,
I have a question about intuition and gut feeling. And so, the question is do you ever
rely on a gut feeling and intuition in making a decision in chess? And do you think such
thing exists? >>Garry Kasparov: Oh, it does exist. >>member #3: Is it valid? Is it valid? >>Garry Kasparov: It's the most valuable quality
of a human being in my view. Yeah, it's probably – we live at a time when we want just to
touch something before we can make our opinion about the subject. I believe that intuition
is like any other muscle. So, like people now, if you go to the gym, you improve your
physical conditions. They know that training memory they're also exercises, but intuition
is the same. And so you have to learn how to trust your intuition. My view is that we
similarly undermine the importance of intuition because intuition means taking too much risk.
And we, whether we like it or not, we live in a risk adverse culture. And intuitive decision
very often cannot be explained into terms that should be required by corporate culture
or by your other family members. So, in my view, by adding this quality of intuition
to the decision making process, we can dramatically improve the results. >>Jonathan Rosenberg: So, I see Steven is
way in the back of the line, but winning the chess championship under 18 should allow you
to cut in the line. [laughter] To be clear, it doesn't mean you get a job
at Google, but you do get to ask a question. >>Steven: How important would you say psychology
is in chess? So, basically like playing against an individual opponent's weaknesses and what
limitations would you say this puts on computers? >>Garry Kasparov: The psychology is an imminent
element of any game between individuals because we know that our opponents -- they have some
strengths and weaknesses and looking at the games, you can analyze these potential weaknesses
and if you have two very strong players are facing each other in the match, then the winner
is more likely not someone who comes up with just a brand new idea, but someone who could
push the game into the direction that is, will suit his strengths and will undermine
his weaknesses. It's like if you use a metaphor from the military metaphor, so in a medieval
battlefield, if you have cavalry, you're looking for a valley. If you're fighting against cavalry,
you'll be looking for the hills. So, it's about creating the environment where your
strengths will manifest and your weaknesses-- and again, all weaknesses are relative. When
I played Anatoly Karpov, we were very good at almost everything, but still I was more
dynamic player, he was more strategic. So, it was very important to push the game in
the direction that could fit my playing style more than his. And that's why, that is the
element of psychology and, by the way, it could actually work in a very strange way
when you're facing the computer because many computers, even today, they still have their
own strengths and own weaknesses. And if you can understand so it may help you to design
the game which will be the most unpleasant for certain computer. [laughter] Because it's actually machine, it might sound
very odd, but machine definitely has a quote, unquote "personality" and it very much depends
on the people behind the computer. So, some of the machines are playing more aggressive
chess; some play less aggressive chess. And again, I don't know whether it's an irony
or not, but the Israeli-made computers are more aggressive than the German-made computer. [laughter] >>Jonathan Rosenberg: It's challenging Silicon
Valley great, Jim Barksdale, who basically said, "Who wins in a fight between an alligator
and a bear? It depends if it's in a swamp or the forest." Thank you. Next, how do you
-- jump to the Dory and then we'll do the next one here. "How do you avoid making mistakes
and maintain your peak performance in a long session of chess games?" >>Garry Kasparov: Let's forget about avoiding
mistakes. We don't make mistakes, we're dead. So, just the making mistakes is a normal decision
making process and again-- >>Udi Manber: Yeah. >>Garry Kasparov: I want to refer again to
this risk adverse culture because what is this in business and life, we want to limit
our ability to make mistakes. It's just, making mistake is like a crime. No. It's a normal
part of the thinking process. So, limiting mistakes I think is just, it's first you have
to kill the fear. So, less you fear of making mistakes, it dramatically reduces the chances
that you will make one. Because, most of the mistakes, they have a psychological root,
so just you can't display all your abilities to make a right decision. So, I believe that
the dominant player at the chessboard in the long match was one who could be more resolute
in the decision making. So, more confident in making decisions. And as for the physical
stamina, when I play chess, I spend quite a lot of time by training myself because I
believe that the physical conditions were quite important to keep your psychology at
the right level. It's just, it always boosted my confidence. >>Jonathan Rosenberg: Now he's channeling
Eric Schmidt, "Don't be afraid to fail." An excuse I've used regularly for the last ten
years. [laughter] >>Garry Kasparov: It's, I would even go further.
It's just, it's -- "afraid" -- it's some kind of, it's defensive. It's more aggressive.
I mean, you have to damn sure you will fail a certain point. So just recognizing that
you, some kind of failure is inevitable, I think helps you to boost your performance. >>Udi Manber: Very good. Actually, lemme,
I have a suggestion cause I wanna, I have an idea for you that might improve chess,
I was thinking before. >>Garry Kasparov: My chess? [laughter] >>Udi Manber: No, chess as a whole. So, the
problem with chess, one problem is that it's hard to make a living out of it. Only the
top, very, very top player in the world can make a living cause there's no-- >>Garry Kasparov: Because, chess is, chess
is not a very professional game. >>Udi Manber: Yes, well because it's hard
for people to follow-- >>Garry Kasparov: Because it's run by people
with very sort of limited reputation. >>Udi Manber: Besides the politics of chess. [laughter] If you look at poker, for example, poker got
to be extremely popular because they figured out a trick that people can watch it by showing
some secret information that's not available to the players. What if you did something
like this with chess where after every game, after every move, the player will go aside
to a soundproof room and say exactly what he wants to do and what he thinks the opponent
will want to do. And then the winner will be not just a winner, but who made the guess
better. >>Garry Kasparov: Who were voted by the public,
yes? [laughter and clapping] >>Udi Manber: But the question is that might
be viewable. People might want to watch it because it would be very interesting to see
who gets better and you better and what was-- >>Garry Kasparov: But one of the, it's this
man plus machine chess what I--introduced in 1998-- I called advanced chess. One of
the ideas was that when you go through this process of -- we have the same, we have access
to the same hardware, we can choose our software, but we goes through some lines and we analyze
it. On the screen that we don't, we don't see the screen. On the screen you actually
could see the whole process; what I'm looking at-- >>Jonathan Rosenberg: Right. >>Garry Kasparov: It's, I was, I had this
kind of an idea that at certain point we'd have to share it with the general public,
but still it's a bit difficult to explain all this combinations and the ideas behind
every move because in poker you just look at the numbers. The machine immediately gives
you those. So, that's it. So, the rest is psychology and you don't have to mobilize
your brain stem to understand what's going on. [laughter] >>Udi Manber: Sure, but psychology is kind
of interesting. In fact, while there is a computer that plays very well in chess, there
is no one that plays poker. Very well. >>Garry Kasparov: Yeah, because machine gives
you odds, but machine, again, same problem. How do you translate into be a bluff into
numbers? So, the element of bluff definitely creates, creates, it's a grey zone. So, like
Twilight Zone for the computer. >>Udi Manber: Absolutely. >>Garry Kasparov: You want idea how to decode
bluff? >>Udi Manber: How to decode bluff? Well, you
do your odds and you sometimes randomly win. >>Garry Kasparov: But, it, the whole idea
of, the best players they base their bluff on reading whatever information they can read
from the face of the opponent, so again, how you can translate it into the numbers? >>Udi Manber: Of course. That's why it's hard. [pause] >>Jonathan Rosenberg: Go ahead. >>member #4: Many people consider great chess
players has talent and intelligent or gifted and you have been trained by, have been trained
by Botvinnik, and you have helped many players to get better and you have improved a lot
during your career. So, how much percentage do you think in the chess ability that is
contributed by the talent and how much percentage by the hard work and practice? [pause] >>Garry Kasparov: It's a very vague question
because are you talking about world champions or just chess players in general? >>member #4: Ok, if I've, if I start off by
just considering yourself we cannot distinguish, but you have seen many, many players. Somebody
might have-- >>Garry Kasparov: Yeah, but it's this, let
me say. To become the World Champion, you need an exceptional talent. It may not be
enough, but it should be there for you to become number one, or at least be top five,
let's say. So, I cannot give you again exact level one can achieve without having unique
talent, but definitely there is a certain level that can be achieved if you have ordinary
talent. You still have to have some talent for chess. But the correlation between these
numbers? I don't know, but undoubtedly you need talent. But some players with unique
talent never became the champions because it also needs character, maybe some element
of luck and ability to work hard. But talent is, is a number one condition for you to become
the best in the world, or one of the best in the world. >>member #4: Thank you. [pause] >>Udi Manber: All right. >>Jonathan Rosenberg: Sure. >>member #5: We've talked a little bit about
the game of chess and as far as studying it with computers. Now, with the advancement
of computers to the point that they can brute force a game of chess, how do you feel this
has affected the game on the higher level, with the newer generations of chess players? >>Garry Kasparov: It's probably again, it
probably was inevitable, but it's some kind of computerized mentality because they look
at certain geometry of the board, so it's this geometrical assumptions they replace
strategic understanding. Because they learn a lot at very early age. Today, 12, 13 year
old kid knows much more about chess than Bobby Fischer did in 1972 because just if he knows
how to use the mouse. So, he goes through the, yeah, through these millions of games
and you absorb the information, but learning this information doesn't mean that you can
understand it and let's not forget that many things that this kid learns today, they were
invented by Bobby Fischer because Bobby Fischer was the first one actually to, not to invent
them, but to bring them to life. And many of these technical ideas not only opening,
but also in the middle game. So, I think that today we have young chess players and the
average age is definitely getting younger, so between these years, 50 years, 60 years
ago, the peak of the chess player was 35, 40 years when--that's why I was the youngest
one to win a world championship at age 22. So, if you look at the modern chess, now Magnus
Carlsen, number one rated player, just, he will be turning 20 at the end of this month.
So, chess is getting younger, but they gain this information, they become very, very strong
players, but probably at the expense of less comprehensive approach. It's very natural.
I don't think that, say, if you look somewhere else, you will not see the same phenomenon. >>member #5: Thank you. >>Jonathan Rosenberg: So, if we just shift
to another subject I know you're interested in, after you retired in 2005, you ran for
political office. Or suggested it-- >>Garry Kasparov: It is this; I'm not an American
citizen. I live in Russia, so in Russia we're not trying to win elections, we're trying
to have elections. [laughter] [applause] >>Jonathan Rosenberg: Well, you're also particularly
interested in trying to have change. I think you mentioned in your book that's-- >>Garry Kasparov: Yeah, well that's exactly
the change I want. I want my country to implement, just to have Russian people to speak when
they don't like the course of the country as it happened here 24 hours ago. And so far,
we haven't succeeded but I believe that this is the only way for my country to survive
in this harsh reality of 21st Century. >>Jonathan Rosenberg: So, you mentioned at
a protest, I think a week or so ago, -- >>Garry Kasparov: Well, it's a constant process.
It's not as big as in France, but definitely less violent. >>Jonathan Rosenberg: I think you said, "Putin
and his cabinet should resign." >>Garry Kasparov: Yes, it must resign. >>Jonathan Rosenberg: And you've had some
predictions that you've gone on the record, maybe in 2008, saying they wouldn't make it
through 2012. Do you still believe that? >>Garry Kasparov: I still hope so. [audience
laughter] It depends, it depends very much on the situation of the country, but also
it depends very, it depends a lot on the oil crisis because that what, this influx of money
props up Putin's regime since Russian economy and the-- it's not in good shape-- and the
infrastructure, most the Soviet-built infrastructure, is about to collapse. If you travel outside
of Moscow, not in Skolka, but somewhere else further down, you will just recognize that
this country, it's like travelling in a time machine. You will see that most of the country
still lives in a very dire conditions. >>Jonathan Rosenberg: And if you could achieve
the kind of regime that you think the people would like, what changes would you make to
address these issues? >>Garry Kasparov: Look, I definitely know
this, the wealth created in the country should be used not to buy soccer clubs in England
or basketball clubs in the United States, but to be invested in Russia and we have hundreds
of billions of dollars being invested or taken out of the country by Russian oligarchs, and
I believe that this money can be wisely used to improve conditions of our vast country. >>Jonathan Rosenberg: Should we go back here? [pause] >>Udi Manber: I think that gentleman was first. >>Jonathan Rosenberg: Oh. All right, I'm not
looking to my right. >>Garry Kasparov: It's the second one. We're
outflanked. >>Jonathan Rosenberg: We'll head for a valley. >>member #6: I saw an interview you did last
year about training Magnus Carlsen and they briefly mentioned the idea of the level of
discipline that it takes to compete at the highest levels. I was wondering if you could
elaborate on, as an instructor of such a student, how do you bring about that, those levels
of focus and determination that it takes to get to the top? >>Garry Kasparov: I would very cautious using
words "instructing a student" to describe my relations with Magnus. What I think I did,
and I was quite happy with this, was I helped him to learn how to study the game of chess.
That's the, that was one of the unique things I learned also from Botvinnik is the value
of being part of a Soviet chess school when the experience was transmitted from generation
to generation. And Magnus, as I was talking about this phenomena just a few minutes ago,
he learned a lot by just clicking the mouse. So, he got an appetite for studying the game
of chess the way I did. And I think it was very, very helpful because it helped him to
expand his horizons. Also, what helped me to work with Magnus is that his chess style
is, his natural talent is different from mine, so it's more like Karpov. He's a very quiet,
positional player with a very good, with a phenomenal ability to actually recognize the
positional elements in each position and very sharpened dynamically. So he could learn how
to do other things in chess. So, it was a very good extra value added to his chess education.
So, I enjoyed it immensely and I'm very happy to see that my work produced such impressive
results. >>Jonathan Rosenberg: One more from the neglected
line over here, I guess. >>member #7: So, much of Google's technology
works on the notion of the wisdom of the crowds as displayed by the Web; the connectivity
of webpages, etc., yet the games of, when you played against the world and recently
when Magnus Carlsen played against the world, the world showed a rather fractured approach
to the playing side. Can you say anything more about this, or elaborate upon why you
think that might be? >>Garry Kasparov: Yes, but I think we have
been dealing with, especially the game you mentioned, Magnus Carlsen played with the
world, it's a very special exercise because it's not the, not about generalities. So,
there's the, and you can't install these big numbers to ensure the quality of the decision.
When Magnus played the world, there was three Grand Masters, relatively strong Grand Masters,
but still weaker than Magnus, offering the moves. So, and what's happened is that this
moves, very often they were contradicting each other because they played their own games.
So, adding players to the process didn't help to improve the quality of the decisions. When
actually I played a Microsoft match when we had not two minutes per move, but two days
to make a move, that game was probably one of the best games ever played because the
group of the strong chess players, they created a very powerful algorithm by using computers--it's
about the process, also-- and picking up the best moves. And at certain point, the game
was simple phenomenal and I just still don't understand how they could make this terrible
blunder at the very end, probably this complacency, and they lost. But the game had to actually
end up in a draw and it was a very, very high quality game. >>member #7: Thank you and thank you for joining
us. >>member #8: [speaking in Russian] I have
a bit of a philosophical question for you. Earlier you said that-- >>Garry Kasparov: You mean the others were
not intelligent, yes? That's the question? [laughter] >>member #8: They were very intelligent; this
one probably is not intelligent. Earlier you said that any computer can beat any player
today by analyzing-- >>Garry Kasparov: No, I didn't say any computer
or just any player; I said just the strongest computer, ok but-- >>member #8: A strong computer can beat any
chess player today by analyzing any one the ten to 120th possibilities of moves that are-- >>Garry Kasparov: No, no, no, no, I didn't,
you, Udi, please help. Did I say that? Ten to the power 120? No. >>Udi Manber: Ten to 120 is all the possible
moves ever possible in chess. >>member #8: Ok, but if a computer can analyze
all those moves, it can potentially beat any-- >>Garry Kasparov: Can you imagine the number?
Ten to the power 120? [laughter] >>member #8: I can. It's a big number. >>Garry Kasparov: Big? >>Udi Manber: It's more money than we make. [laughter] >>Garry Kasparov: It's even more than the
amount of princes, yes? >>member #8: Ok, the question that I have
is if a computer can beat a human, it seems to me like chess is a game that's deterministic;
there's no chance, there's no free will. Do you think that your mind works the same way
that it's deterministic? Do you think that-- >>Garry Kasparov: But again, let's move back
to the limitations of the game of chess. So, probably the number of all legal moves probably
was a wrong example, but right now if you look at the end games analyzed by the computers,
the computer solved all four pieces end games. So, including kings. So, that's fine. Then
solved all five pieces end games. I think they solved most, probably almost all six
pieces. Now, they are just approaching seven pieces. I'm not sure it will happen soon,
if it happens at all. Maybe the limits will be eight pieces, now, may be because the numbers
since you start looking at it, this data, they grow exponentially. Now, the chess game
is an ultimate end game of 32 pieces. So, that's why there's no chance a game of chess
can be solved. Theoretically, it cannot be solved. But what happens is that the game
played by the humans is not perfect, so one of the biggest problem, the human, strongest
human player experiences while playing the strong or relatively strong computer, is the
different level of resilience. When we play, two humans play the game. So, and I'm getting,
so I have a superior position. Very often, this psychology works now in my favor because
my opponent is losing his will to defend. And very, in chess we say that the mistakes
do not walk alone, so you always have more mistakes because it's a pressure. So, winning
the game of chess between two humans doesn't require the same energy as to beat a machine
because machine doesn't care. So, it's the, when you look what is a good position or bad
position, starts from each, every move it starts from the scratch. So, when you play
the machine, you have to make sure that the quality of all of your moves, all, literally
all the moves, will be almost perfect. It doesn't happen in the game of chess. When
you look at the best games as I played against Karpov or Kramnik played against Anand, I
bet you you'll find many not mistakes, but inaccuracies. So, inaccuracy is a natural,
some kind of a small mistake, is a natural part of any decision making process; against
the machine, it's deadly. Because small inaccuracy could kill the result of your very hard work
that you, all you accomplish in previous two or three or four hours of a game. >>member #8: Don't you think those mistakes
are also deterministic in your mind, though? That they're based on some sort of decision
or some past experience or-- >>Garry Kasparov: No, it's just again, it's
the mistakes. It's part of whatever weaknesses we are performing at a certain moment. You
could be tired. Your mind can be blind by whatever, by family problems or crash of the
stock market, so whatever. So, you have a lot of problems so we have fundamental problem
of concentrating all our abilities to this specific problem and I believe that if we,
to continue this experiment, man playing versus machine, it should, we should change the rules
and a key alteration should be related to the overall result. It could be four games
match, six games match, eight games match, but we understand that the longer we play,
so the better chance for the computer because it will play almost steady game while humans
could, the quality will fluctuate. I believe that what we have to find out is whether the
best human player at his peak can beat the best computer. So, if human player wins one
game, that's it. Because we don't have to look for the longer, for the long competition
to prove that we are better. It's very important to find out that at the peak of our performance,
we're still capable of competing or even beating a strong computer, which is very, very difficult
now. >>member #8: Thank you. [pause] >>Jonathan Rosenberg: Go ahead. >>member #9: Hi, my name is Mila. >>Garry Kasparov: Hi. >>member #9: First of all, Garry, thanks so
much for coming over to Google to visit us. And let me shift the subject back to politics
again if you don't mind. I have two questions. What's your take on the Khodorkovskii's case
and do think the justice will prevail in the outcome of his trial in December this year? [pause] >>Garry Kasparov: We do not have elections
in Russia. We do not have justice in Russia. So, I'm quite pessimistic about the outcome
of his trial, which is a shame for my country and it's a tragedy for Khodorkovskii and the
employees of YUKOS. Which again, not sure how audience here is, how much people know
about this case, but YUKOS was the most successful oil company in Russia. By the way, the most
transparent oil company in Russia. The company that paid most taxes per barrel of oil and
because of the success of Khodor--, Mikhail Khodorkovskii and his team and his willingness
to merge with multinational corporations and to bring this transparency to Russia and to
make it a standard. I believe that's why the company was ruined and it ended up in the
hand of Mr. Putin's cronies. I made this very green prediction in 2004; as long as Putin
stays in power, Khodorkovskii stays in jail. And now it's a second trial, which is, it's,
it makes even Kafka pale by comparison because first time Khodorkovskii was convicted for
not paying taxes. Now, he's convicted for stealing the, everything, the entire oil production
that YUKOS extracted. And, in fact, if he didn't pay the taxes from that, how he could
steal it? So, that's the, but it, the whole process is a terrible display of a lack of
rule of law in Russia and cruelty of Putin's regime and also corruption because I already
said YUKOS and its most precious part of the company, now owned by Putin's closest, closest
friends. I hope that Khodorkovskii can leave jail alive because this is one of the most
brilliant executives in Russia and hopefully, one day, his talent could help the country
to come back to the civilized world. >>Jonathan Rosenberg: So, is the Internet
and online tools and access to information and information transparency and the kind
of things that we tend to preach here as being the lifeblood of democracy -- are these things
helping increase the odds that there's gonna be change in Russia, or are they simply allowing
the existing regime to reinforce its hold on power? >>Garry Kasparov: Now, Internet plays a very
important role in beefing up the opposition in Russia. We don't have the same phenomenon
as in China. So, the Russian stories this day, they are not as aggressive in limiting
the Internet access because it's very much part of this Putin's, not a deal, but his
philosophy in dealing with the, with the population. Private life is some kind of a guaranteed
unless you are interfering with politics. So, as long as you don't mind Putin's business
by robbing the country, your personal life will not be under any threats. So, that's
why Internet has been rapidly growing in Russia. I think now the number is 41 or 42 million
people somehow connected to the Internet in Russia, but obviously not more than ten percent
of this crowd and those numbers probably they, they are similar to the rest of the world.
Ten percent are interested one way or another in politics. And out of this ten percent,
maybe one-third to a half could be interested in reading alternative source of information.
So, that's why the audience which is related to our website, and we have very few websites
in Russia because the media, the mainstream media is under a hundred percent control of
Kremlin, so we have about a million and a half, two million people following what we
are doing and it's, but this number grows. So, that's why you see more and more people
on the Moscow streets or St. Petersburg, and what is very important the information exchange.
Now, anything that happens in Vladivostok or in Siberia or in Urals or in other parts
of Russia, very, very soon makes the news. But on the Internet asides in Moscow. >>Jonathan Rosenberg: Ok. So, we're almost
out of time, but maybe we have time for one more question here. >>member #10: Hi. You wrote a book called
"How Life Imitates Chess" and you've talked about peak performance in other areas, whether
it's business, politics, etc. Is it constructive to think about life as a game? And if so,
can you share your secret for winning the game of life? [laughter] >>Garry Kasparov: No, I, winning the game
of life. Wow. [laughter] >>member #10: What does that, what does that
mean? Is that what peak performance is, to view it in that way? Can you talk about how
somebody would go about achieving peak performance in other areas other than chess? >>Garry Kasparov: I, when I wrote the book
I had an idea of just starting a dialogue with an individual who had a desire to improve
his or her performance in decision making. What I always didn't like is just to provide
universal advice. Somehow it goes against the conventional wisdom that you can have
an advice how to improve your performance. I believe that we're all different, we're
all unique. And something that may work for me will not work for you. So, it's very important
we start analyzing our own performance and looking for, as we talked about today already
about strengths and weaknesses. And then you create your own formula, which might be not
working for me and when people trying to analyze their own performance, sometimes they get
upset if they look at their qualities, they say, "Oh, I'm too arrogant." Or "I'm just
very defensive." It doesn't make any difference. You could be very good in attack; you could
be very good in defense. At the end of the day, you have to understand what you're good
at and what you're not good at. So, you can have a tennis player who has a very powerful
serve, who's rushing the net and someone who plays very defensively. Both could be number
one. It's very important that you look inside first. That's the, that's my advice. I, when
I hear this -- some of the big names in the industry providing the audience in a massive,
the huge audience with this tip and advice that can work for everybody, I don't think
it's – they're doing a good service because we are different. And by, without recognizing
what's inside of us, we cannot come up with the right decision making formula. It may
not be winning at the end of the day, but I believe that it's all; it's a lot to ask
to make all the difference. So, and I think that we, almost all of us, probably all of
us, we underperform because there's a huge potential and what keeps us, what keeps us
behind is a fear of to make mistakes. So, it's about courage. So, and again, we're going
back to these, to the modern culture. So, I, if you look at back in the history timelines,
so there's the great explorers, I mean, today I don't think Magellan or Columbus could get
venture capital. [laughter] Yeah, it's very important to understand it,
why the, I mean, you can't expect let's say a ten percent annual return without taking
risk. So, otherwise, we move into the Bernie Madoff's world. So, we have to take more risk.
We have to recognize that risk and mistakes, they're inevitable for those who want to make
progress. So, it's another, it's a very long story, actually. I'm working another book
now with my friends, Peter Thiel and Max Levchin, and its tentative title is The World of Fake
Values, and we talk about what we believe is the, I'm afraid to say, technological stagnation
of, in modern days. And it's very much related to complacency and the risk adverse culture.
So, and I'm used to take risk and I did it all my life and I believe that that what can
help us to start pushing things forward. >>Jonathan Rosenberg: Ok. If you're looking
for more answers, please buy all of his books. [laughter] So, let's all thank Garry for coming. >>Garry Kasparov: Thank you. >>Jonathan Rosenberg: Thank you very much. [applause]
