Deep Thinking | Garry Kasparov | Talks at Google
Вставка
- Опубліковано 14 жов 2024
- Garry Kasparov and DeepMind’s CEO Demis Hassabis discuss Garry’s new book “Deep Thinking”, his match with Deep Blue and his thoughts on the future of AI in the world of chess.
Get the book here: goo.gl/OwuOcW
Event moderated by Demis Hassabis, CEO, DeepMind.
** About the book, Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins **
In May 1997, the world watched as Garry Kasparov, the greatest chess player in the world, was defeated for the first time by the IBM supercomputer Deep Blue. It was a watershed moment in the history of technology: machine intelligence had arrived at the point where it could best human intellect.
It wasn't a coincidence that Kasparov became the symbol of man's fight against the machines. Chess has long been the fulcrum in development of machine intelligence; the hoax automaton 'The Turk' in the 18th century and Alan Turing's first chess program in 1952 were two early examples of the quest for machines to think like humans -- a talent we measured by their ability to beat their creators at chess. As the pre-eminent chessmaster of the 80s and 90s, it was Kasparov's blessing and his curse to play against each generation's strongest computer champions, contributing to their development and advancing the field.
Like all passionate competitors, Kasparov has taken his defeat and learned from it. He has devoted much energy to devising ways in which humans can partner with machines in order to produce results better than either can achieve alone. During the twenty years since playing Deep Blue, he's played both with and against machines, learning a great deal about our vital relationship with our most remarkable creations. Ultimately, he's become convinced that by embracing the competition between human and machine intelligence, we can spend less time worrying about being replaced and more thinking of new challenges to conquer.
In this breakthrough book, Kasparov tells his side of the story of Deep Blue for the first time -- what it was like to strategize against an implacable, untiring opponent -- the mistakes he made and the reasons the odds were against him. But more than that, he tells his story of AI more generally, and how he's evolved to embrace it, taking part in an urgent debate with philosophers worried about human values, programmers creating self-learning neural networks, and engineers of cutting edge robotics.
![Lp. Сердце Вселенной #30 ДВА В ОДНОМ [Один человек]• Майнкрафт](http://i.ytimg.com/vi/nrXy15orKhY/mqdefault.jpg)
Kasparov can even answer your question before you finish asking.
That's how far he can calculate.
I've never heard him speak before, so that's why I clicked on this video. I read your comment before 0:36, and then laughed a little.
He can even fart at 27:56
@@PornobrillenAli lol
@@PornobrillenAli hahah you are crazy how did you pick it up!!?
@@PornobrillenAli lmaooooo he calculated before the interview exactly when he was gonna fart thats why he didnt give a shit he just let that diarrhea squirt out his ass
13:50 "Computers are useless because they can only give you answers, but everything begins with a question" Brilliant
Its a pablo picasso quote.
@@a.s.7160 Pablo picasso did not say everything begins wit a question in his quote.
42
Humans always have one thing that computers can never possess, and that is WISDOM! ;)
@@antoniobreaux1584 its implied in the statement
There's so much energy and passion in the way he talks. So fascinating and inspiring!
yeah i was thinking of how passionate he is
@@ChessMasters236 he talks in *bold*
@@versatilegeniuses9374 Do you like bold talkers or cute talker?
"It's not about solving the game, it's about winning the game." That is tremendously insightful.
Mike Headley how so?
Solving the game IS winning tho. It's like tic tac toe. The second player only has two options if he's playing against someone who knows the solution. Either take the draw or lose the game.
But that's just when you take the quote at face value.
@@stuartfleming To be able to know what the best possible move is in any given position. In games that are fair, if both players play "perfectly" (ie. know the solution and follow it) the game should be a draw.
@@stuartfleming I don't know since it hasn't been solved lol. Probably, since depending on your opponents next move different moves would be optimal, it would more likely be a move that gives you the highest possible chance of not losing in the long run. Maybe something to that extent can be proven mathematically. There has been a lot of work done on the math behind chess but I've never come across a paper like that.
I thought about this a little more. Even in tic tac toe if I go corner, my opponent goes center, I have 2 different corners to choose which would be equally optimal (since the board is symmetrical). But whichever corner I choose my opponent is locked into playing the game on the terms I've chosen, he's locked into the dance I'm leading, otherwise he'll lose on the next move. So his best move is to stop me from winning immediately. Which coincidentally makes me forced to defend against an immediate loss. Only way to lose is a blunder, but with a solved game there is no blunder.
Chess is of course more complex than that. The board stops being symmetrical in the first few moves usually.
@@stuartfleming Yeah. I've never played connect 4 but a friend mentioned it had a known solution. We'll see with chess as computers (or AI) continue to evolve, if it happens in our lifetime at all that is.
I love his point about our brains becoming subordinate to AI, people becoming too reliant on someone or something else to do the math for them. I think it's important we think for ourselves and figure things out for ourselves and not just submit to an established model, in most areas of life, and not just chess. Great talk given by Mr Kasparov!
Now we have the problem of human's cheating using engines. Back then we had the problem of engines cheating using humans.
Hahaha
sort of like when many grandmasters come together to put in the playing style and counters of one single player then after the match take the machine apart so no one else can play it....LOL
Very funny good one go now
@@isaacvongurtberg7341 Nikhil kamat
This comment aged well...
Kasparow has a very entertaining way of explaining his point of view.
Just like you did.
Kasparov is the kind of guy that you hear talking for a minute and you know he is intelligent on many levels, not just chess.
His IQ is enormously high
he is a cheater in every aspect of his miserable life. There well known facts as he cheated young chess players.
Its rare to find "old school" people that can accept the existent of technology like Kasparov
It amazes me how Garry is still so passionate about the game when he talks about it. He is an amazing guy and an inspiration to anyone following his or her own passion.
Wow. That is an incredibly sharp mind at work. It was a real pleasure watching this.
even that sharp mind went blown with a child question
Guy's really fun to talk to. He's not only a great chess player but a good talker as well. Shows how intelligent he is
Nobody sounds to care about the interviewer (Demis Hassabis). Hassabis is the creator of DeepMind (AlphaZero) and is classified as one of the 10 best scientist of the world by the magazine Nature.
He's also one of the best player of chess (2nd best world player of his age at 13 years old) + one of the best world player of Go + one of the best world player of Shogi (japanese chess). 5 times world champion of Pentamind.
Hassabis is probably the smartest guy in the world at this time. He sold DeepMind to Google for 400 Millions $ and now he's working on AlphaFold to predict the proteins structure.
He got a lot of awards and honour all over the world. I have a lot more admiration for Hassabis than for Kasparov : en.wikipedia.org/wiki/Demis_Hassabis
It's nice to see how Gary is still excited about everything.
What we can take away from this discussion is if there ever is a hair regrowth product that actually works, then Kasparov will once again dominate the world of chess.
Underrated haha
If he tries minoxidil, he just might🤔
I'm sure if I could find a hair regrowth product that actually works, that it would improve my game. smile
The thing that struck me most, and what I happened to realize even before viewing this wonderful interview, is that humans can do some amazing things without being able to explain how they do it, and what that means is, that computers will always lag until a human can translate how they do it. Get it? Perhaps I did not express it well enough.
Kasparov, being who he was....turned the 97 match into what it blew up to be. He is a passionate character. And if IBM thought that he's just going to go away with all of his issues during the match, they were fools. Some players might have just shrugged off a thing or two, and might have been able to handle the tensions better, and thus, might have even won the 97 machine. But, it just so happened that the World Champion in 96-97 was Garry Kasparov
Kasparov is sort of a wild man. He seemed nearly as "inflamed" about his 1997 loss to Deep Blue today as he was then.
deep blue beats kasparov anywhere that french is spoken
It's because he's an Aries. Extremely competitive and passionate about things.
I really enjoyed this. Came here expecting a Computer Science talk but got interesting commentary about it's application instead.
*its
For a GM like Kasparov is about finding the best possible move in given position suited to his "style' of play. Chess engines on the other hand get busy polling the ramifications of many moves in seconds. So you have human chess logic versus a checkmate calculator. engine.To the concept of AI you have to add AP Artificial Perception to be more human-like.
I never imagined Kasparov would be such a jovial, eloquent personality. I thought he'd be a reserved, inarticulate guy
Why?
Bobby Fischer, Anatoly Karpov and Gary Kasparov are shining stars in the universe of chess.
Kasparov is a great speaker. His accent may obscure certain words, but if you listen closely, he makes a lot of sense.
Love his warmth, energy and enthusiasm. Brilliant...and generous.
3 minutes into the video and i realised this guy is not only a chess genius but also very intelligent in other avenues
Great interview, full of passion of energy. Definitely the best spent 40 minutes today :)
Kasparov should play a rematch against Deep Blue for sake of promoting chess and old time's sake. It would be loved by many!!
Deep Blue was dismantled. Doesn't exist anymore.
Well, yes, but in reality its two towers were simply sent to two museums. It would be very easy for IBM to set the computer up good as new.
I doubt it would be. And in any case what would be the point? Kasparov is a retired chess player and Deep Blue is ancient history in terms of computer science.
I mean what was the point of playing Alan Turing's Chess Engine? Play it on the anniversary of the match, for publicity, memories, and fun.
Well, no point in speculating anyway, it'll simply never happen.
What a self-confidence he has! He is so proud of his achievements!! King of Chess for a reason!!!
King of chess for sure.
Garry you are the best player and GM forever
I wish you good look in the next A thousand years
his English is surprisingly good
He knows at least 2 more languages at this lever and 2-3 more with less fluency.
He lives in the US
@@maxsteiner7040 which languages?
@@represiya7035 russian
he's got a very wicked brain, thinks with deep passion and logic at the same time, definitely a skill that he learned from the days living in competitive chess. Nice!
Interesting to see them laughing about setting AlphaGo loose on chess, speculating about the possible outcome. Well, now we know the outcome.
Bababoom Baboom came to comment just after hearing that as well :) actually, I believe the games were already played at the time, only the paper was published some time later. Or they immediately took upon this challenge :)
I suspect Demis was laughing inside at the time, knowing that beating stockfish was only months away.
35:25 if you watch this on 2018, this moment is gold. He knew that 5 months later, stockfish was going to be smashed.
What AlphaZero did to Stockfish is just absolutely amazing and mind boggling! I just pray that we humans will be good stewards of this AI stuff.
“1985, machines were weak and my hair was strong”. 😂
Kasparov is like an intellectual version of Mike Tyson
😀
don't belittle tyson like that, he's pretty intelligent!
Semir Djedovic pretty intelligent and a real intellectual is an insuperable gap
@@vanguard4065 making an assumption of someone else's intellect that you're not certain of is an insurmountable gap
@@djedosemir I would think that it's because Tyson is pretty intelligent that Eric made his comment.
Garry Kasparov is a deservedly household name, both from his chess & political endeavours. I haven't seen any comments about Demis Hassabis here, he also deserves to be a household name after his achievements & contributions (already so far). Hassabis was well on the way to being a great chess champion but gave it up to pursue other things that he felt he could be more useful with. Garry Kasparov being interviewed by Demis Hassibis - wonderful! If you liked this then Demis Hassibis being interviewed by Jim Al Kahlili (another quite impressive bloke!) on The Life Scientific, BBC Radio 4 is worth a listen.
the comment about Magnus human approach to the game is spot on
Who else thinks Garry Kasparov sounds like Gru from Despicable me?
Damn this guy is a genius.
And very down to earth
Andrej Nikolov 194 iq
Yeah, he's the man behind alpha zero
Putin is a genius!!
Losing Is Improving Misinformation... He has 135 iq, he has superior intelligence; however, he is not a genius.
What an incredible man wish I could meet him one time in my life..
I once admired this man when I was learning chess back in the 90s. Then I heard him talk politics and history. This interview reinforces my belief that chess is the only thing admirable about him.
What are his political views?
@@vancedadder In short:
West good, east bad.
@@rikkertbatzback1816 I don't agree. West bad too
@@vancedadderSame. I guess we all choose what we perceive as the lesser evil.
@@rikkertbatzback1816 it's interesting that you agreed. I'd thought you were keen on East being bad too
kasparov is probably one of the few chess masters during the era of soviet dominance that didn't cheat or at least collude with other soviet players
Always great to see good ol Kaspy. If any of you is curious to see how great computer chess is these days, look up the official channel of Magnus Carlsen (current world chess champion) and watch him play against different levels of his own phone app. Again, he plays against a phone... and the engine on that phone is not even remotely close in strength to top chess engines.
a phone can do calls to remote server to get deceisions for chess
you didn't get his point, the point is that nowdays any decent chess engine even running in a phone can crush without much effort the current world champion (Which in opinion of many is even better that kasparov).
True but the real fun is to play against 1990s chess computer. Like a Saitek. I had a Mach III, Elo 2000 that would beat me but at least we had a chance now and then. Mephisto was the king, then Chess Genius. But put Chess Genius against Stockfish and he get annihilated...the level of chess engine is out of this world now although it's awesome for analysis I miss the days where you could tempt it with a pawn and checkmate it even on level 5 or 6:)
"Kaspy" wtf
He's not your pet lol why do you people have to gay everything.
It`s really nice that Google Talks about the book and I cannot buy it from Google Books. Really nice
nice point....
thats is called good customer service 😆
Enlightened by your insight Garry Sir.
Thanks.
I did not know Kasparov had such great language skills! I mean, I do no tthink I could speak so well in my own native language as he does is in a second language.
English is probably his third or fourth language. He didn't grow up in Russia.
You're right Gary. Kids now are super strong
Always a pleasure listening to Kasparov.
I know really!
what a fabulous talk. loved it.
Oh gosh, my favorite guys are here together. I love English so much.
Such a good guy. And so fucking smart. Faster than a warp drive and deeper than the Marianas Trench.
24:53 anyone else hear the sound of a pawn moving forward as he says "Time is moving forward"?
So interesting rewatching these interviews from 2022. Just a few months later DeepMind would publish the first paper on AlphaZero, which could mean that during this interview Demis probably already had the results published in that paper, or the team was in the middle of developing AlphaZero. Makes you see this interview in a different light.
and garry was prophetically correct, alphazero cannot match stockfishs brute strength
This was absolutely phenomenal.
People are only looking at computers to find answers instead of thinking for themselves. That is so true.
Beautiful mind... Russia is sick and Kasparov could be madicine but he left Russia to not being jailed or killed like his friend Boris Nemtsov
We learned today that AlphaZero was able to defeat the strongest chess engine Stockfish, while evaluating a lot less variations per second (80 thousands vs 70 millions). In that context, one can think the AlphaGo team member asking a question @26:50 regarding human intuition vs brute force approach of a traditional chess engine probably already knew the answer :) I believe human "intuition" towards problem solving is merely a manifestation of the energy&time efficiency of our brains. It is more efficient to have a 1st pass that skims the variations pool for most promising candidates, and then evaluate only those in more detail. Our brain is fundamentally lazy :)
AlphaGo's and AlphaZero's intuition were developed by brute force. They were pre-trained on millions of random games. The online version of the engine needed to consider fewer variations, only because of all the work that had been done before.
@@snippletrap the challenge for machine is still there, how can human learn to gain such good intuition with much less "training"?
a machine just give answers but everything begins with a question,the most important in the knowlege is not the answer but the question¡¡¡¡ so true¡¡¡¡
thats the beauty of questions
I wonder if Kasparov was aware that Demis Hassabis was once the second strongest under 14 chess player in the world (Elo rating 2300), behind Judit Polgar...
Dah
vazquezb2011 judit polgar not even close with kasparov but her sister can be a match with him
he is a 2239 rated FIDE Canditate Master
Zoust you are aware judit doesn't fall under the under 14 category, and this statement was referencing years prior?
And she has beaten kasparov previously.
My son beat his brother George who was rated number one at the time in England.
Best wishes from India
His english level is amazing.
...the depth of Humanistic mistakes in this UNIQUE kind of games called chessboard may be severely but had in later stages conveys deeper meanings to improve further PERHAPS isn't there in the title of this video interview with a great personality we all know and admire because AI algorithm are UNBEATABLE perfections whereas mind flickers like a butterfly hard to understand under such circumstances this personality deserves some " concentration or focus award " be just not money but a citation kind. Thanks.
Gary is analyzing and answering questions like making chess moves.
Kasparov is in. RPA. Accepted ML. ❤
Actually, the true legend here is Demis Hassabis. What he's done with AlphaFold for the proteins is really unbelievable.
Great man that let express his emotions
Magnificent
gary kasparov
international business
machine unbeaten
Well it's curious but i have more admiration for Demis Hassabis than for Garry Kasparov. Hassabis is a real genius in every matter, not only chess. He's a great player of many games AND one of the best scientist. Hassabis is very useful for the future. Kasparov is just a gamer.
kasparov is just a gamer. like einstein is just a scientist.
kasparov doesn't open the door, the outsmarts physics with his 16 move plan that opens the door up for him
In 1997, you needed a supercomputer crunching raw numbers with brute force calculations to even challenge a grandmaster.
In 2017, all you need is an average off-the-shelf laptop with the latest chess engine installed to defeat any grandmaster, simply using clever algorithms instead of raw computational power.
True. But Deep Blue was not only brute force but programming techniques and heuristics have beem greatly improved since then. Also DB could calculate 200million pos/sec which you couldn't get close to with a home computer back then. Now even my old 8 cores/16 threads calculate 15/20 million position/sec. Stockfish on a regular laptop would beat DB no problem.
What a great talk! This guy deserves way more visibility and respect than he gets.
16:35 Increasing the speed of learning is a notable pedagogical issue. I wonder if the general concept he is illustrating there could be applied to other fields.
"it's happening. we just have to adjust."
13:54 Amazing said.
his ego was his problem, the program had no ego, Garry played into a complex position on purpose and paid the price in the last game. Kramnik has a better style for computer matches but my phone could beat kasparov now
even own phone can beat any players .....
this is called science advancement
Kasparov has a very interesting accent
"Energy" is Kasparovs main word. In a computer the energy never drops, unless u pull the plug. The interviewer goes : "Whould you like to explain this?..." Well.... Try to stop him?! Kasparov never stops. He is the Terminator. So... Chess is two energys channeled against one another. And the one that channels the most accurate at the target wins before the other one wins. I wish I knew what my point was. But I don't have the energy to figure it out. I am not Kasparov. I am Karpov.
Good comment, except I don't get why you made that jab at Karpov.
He would make a fantastic football manager
29:53 he could see 15 full moves ahead
The solution to the final question asked by the man in the front about elimination of brute force, is already present in the method of genetic algorithms. Well I guess that's what the interviewer meant when he pointed that self learning systems don't use brute force.
I would have loved to attend this presentation.
you just somehow did :)
"you look at this position and you say, this smells" - Garry Kasparov
Kasparov meets Hassabis. OMG !!! Amazing :D
Finally one who care about Hassabis :) Seriously, it's an honour for Kasparov to meet such a brillant brain ;)
Humans are general purpose machines. Machines, on the other hand, can specialize very deeply in a specific task. There's not point in trying to resist.
And now Alpha Zero crushed Stockfish with some amazing chess.
Very interesting video, I am happy this ended up in my suggestion list.
Fascinating and inspiring! What do the thumbs down mean????
When normal people start to calculate moves, Kasparov already calculated a whole tree of variants.
The question was asked to Gary (paraphrased) if he would be surprised if when AlphaGo (a learning system) was programmed to play chess that it would beat the best chess engine. That already happened 11+ years ago when a relatively weak engine with reinforcement learning was beating the world's best engine Rybka in 100 game matches. What engine was that?
That relatively weak engine was RomiChess. Romi beat Rybka with a far greater score than Alpha Zero beat Stockfish and did it with far less training than AZ. In a 20 game match Romi won against SF's predecessor Glaurung 19/20 with no losses. That of course was after 19 previous matches where Romi only scored 1/20 in the first match with two draws. Romi did that with no prior training and on EQUAL hardware.
Anyone have an opinion on any of this?
Romi's learning algorithm is very well known in a very small group of individuals and I am 100% certain that the Alpha Zero team knows about RomiChess's learning algorithm. Here is a quote from the Chess Programming Wiki.
"RomiChess is famous for its learning approach"
Greats Chess player of all time.
Wonderful and worth watching by every student
Sheer Genius in action. Excellent speech what an energy and language
English not even his mother tongue.
Deep Blue was a fish. But Libratus seems the shark to be feared.
Yeah, I think him in his peak, could beat any GM at their peak, (all time). I.e. Fischer, Tal, Anand, Karpov, Morphy, but Carlson hasn't yet reached his peak yet. Although I still think Kasparov would beat him, at his peak.
It would make sense for him to be without a doubt one of the greatest chess players in history at his peak. But, let's not forget how unheard of Morphy's peak was, or how remarkable Fischer was at his peak. Also, I feel Magnus is close to his peak. He has lost badly, he didn't crush the WCS, and his rating has been 2850s-2880s forever. Gary's peak is 2851, right? That is amazing if you think about it being pre-computer. Again, Paul's and Bobby's peaks were the greatest in history and pre-computer with the greatest rating gap. Bobby most likely is the first super-GM, no? In Paul Morphy's case, he had pretty much no books, can you imagine, most can not.
So, what I mean to say is, how do you define peak? Yes, Gary, by definition, would beat Paul every time, peak or otherwise, (Gary 1980, Paul 1850) but does that mean Gary is a better Chess player? I do not think so. Gary was truly great. Morphy was the first modern Chess player and the only true Chess genius and *the* first Chess world champion. Everybody else happens to be a genius and great at Chess due to a lifetime of training/computers or a kind of self-made genius like Bobby Fischer (Gary and Bobby having IQs of 180s). But Morphy was *the* Chess genius. Paul Morphy is the greatest Chess player in history. I am utterly convinced of this.
Good post, Thought.
Well, the best peak ever belongs to Capablanca. I read some scientific research where the moves were analysed by a top engine and the players were given points for every move, the most points was for the top engine move. According to this research ( really sorry for no source, forgot the name of the analysis, just google it ) Capablanca had circa 3130 accuracy at his peak, Fischer about 3050, Garry and Magnus about 2900 or 2800. I don't say Magnus, Garry or Morphy was bad, but technically, Kasparov wouldn't beat everyone at his peak, Capablanca would crush him.
" I am utterly convinced of this." -Thought Workshop. It takes conviction to believe in mere conjecture. The evidence simply can not exist, but I appreciate your perspective.
@mangann. Please, if you could, show us your source(s). Extraordinary claims require extraordinary evidence. What you have presented is no evidence at all, although I will admit that I think I've heard of such an experiment, that alone does nothing to prove the projected ratings of those players.
Yeah, I'm perfectly aware of that. Sadly, I couldn't find the one I saw a few years ago, but I found another one. It isn't so convincing, but still, Capablanca makes it to the top spot with his accuracy.
en.chessbase.com/post/computers-choose-who-was-the-strongest-player-
ailab.si/matej/doc/Computer_Analysis_of_World_Chess_Champions.pdf
Expert opinion, and it's constructive.
Alphazero, playing with black, losing to stockfish, does show that in a rule based chess boardgame, its self learned techniques can be defeated much more if greater number of searches per decision is used by stockfish..
that happened because of the opening
it was a losing opening for black if played by perfect chess players
Alphazero won when it played as white against stockfish with the same opening
Take notes Google, he is talking about you and your media machine, and the mass media at large.
Queen of Katwe looks like a good chess movie!
Great interview by Kasparov but what disappointed me was.....he should have lectured on how a club level player can immensely improve his chess game and what are the steps involved to overcome the bad play to excellent play.
Neither the interviewer nor audience asked any question on above.
GO is a board game between 2 players where the object of the game is to control as much territory as possible on the board.
35:28 Prophetic.
The computer that beat him studied his moves. He never studied the computers moves.
even his all generations can not study all, commputers moves has figured it out😆
Brilliant mind and personality.