back to article Don't try and beat AI, merge with it says chess champ Garry Kasparov

Garry Kasparov, a former Soviet world chess champion and one of the greatest players of all time, has changed his tune about AI since he was beaten by IBM’s Deep Blue. During a talk at the Train AI conference in San Francisco on Thursday, Kasparov traced the steps that convinced him about how humans and machines might one day …

  1. Anonymous Coward
    Go

    That's a Masters privalidge

    To find things we didn't know about the game, and AlphaGo did it, using self learning {gulp}.

    I was only a 2600 using Kasparov's chess game, I did occasionally get higher but never continued it long enough, becoming occupied with other hobbies. I've forgotten it all for immediate recollection, put it in the back blocks of my mind.

    We can learn using their AI's learning technique too. We don't have to be stubborn about this. Doing either is still human, learning &/or stubbornness. We can adapt to discoveries by AI.

    Adaptation this is the key to all changes, from microorganisms to elephants, whales, primates to humans and beyond.

  2. elDog

    And it is extremely hard for a human to admit that they have flaws.

    Gary Kasparov is a prime example of someone who learns and is willing to push the envelope to learn.

    Computers don't have any ego in the game. They learn by making mistakes and correcting their errors.

    Humans (of our current ilk) have been on the planet 100-200,000 years. We've had a fair amount of time to get reprogrammed.

    Computers and self-modifying logic (AI) has been around about 60 years.

  3. Nick Kew

    The prodigy

    Chess is one of those activities where the child prodigy can shine. Years of worldly experience count for less than sheer ability to work with patterns. The only caveat is that the patterns are somewhat removed from natural by the complexity of the rules (I much prefer patterns in Go).

    The chess computer learning without ever studying human games is basically taking that principle to the extreme. If it's winning, that kind-of makes it the ultimate child prodigy: deep pattern awareness, no worldly experience.

    1. BananaPeal

      Re: The prodigy

      the child prodigy didn't get his intuition from reinforcement learning. he got it from whatever pattern recognition neural awesomeness is in his head. alphaGO played itself millions of times to learn. not the same as intuitive intelligence at all.

  4. Gel

    For a while computer + human may be better. But soon computers will be better without humans.

    For chess we are probably at the point where a human with computer assistant is no better than a computer on its own and may even be worse. If not it will happen soon.

    1. Robert Helpmann??
      Childcatcher

      John Henry Redux

      For a while computer + human may be better. But soon computers will be better without humans.

      Better in what sense? Better in being able to solve a problem or set of problems? Perhaps, but then what? At this point, AI is a means to an end. Just as cars are a more efficient way to go long distances than walking and self-driving cars might soon become a safer way to do so, AI is and will provide better ways to achieve certain goals. Once the goal is achieved, is there nothing beyond it? Is there any intrinsic value in the goal or is it simply a way to gain something else?

      This isn't a new story and it's one I do not think we have the ending for yet.

  5. Fizzle
    Black Helicopters

    Time's Up

    I think it's now time to turn all our defence and attack systems over to AI.

    In fact, I was thinking of making a film about it and calling it "Term Time", "Time's Up" or "Terminator" or something like that.

  6. amanfromMars 1 Silver badge

    Way Out Ahead with Future Leading AI Programs Exploring Virtual Pogroms?

    With specific regard to ....

    “AI will help us to release human creativity. Humans won’t be redundant or replaced, they’ll promoted.” Now, Kasparov imagines machines will reach what he calls the third ‘Type C’ stage.

    Type C is an augmented intelligence, where machines and humans work together to create smarter tools. Jobs will evolve and adapt, opening up new careers and industries. He envisions a future where humans don't have to become expert coders to work with AI, since machines will be advanced enough to understand and take instructions by listening to the human voice.

    “I lost [chess] but I survived, and I thought if you can’t beat them, join them. From now we on we have no choice but to work with machines and make the best algorithms,” he concluded.

    ...... We've been here/there already, El Reg :-) ........
    Fortunately though, there are smarter enabled and enabling beings on the planet with minds and AIMind with Plans and ACTioNs that change things radically and fundamentally ….. and can easily have media going mad denying IT be NEUKlearer HyperRadioProACTive and Almighty Powerful. .... https://forums.theregister.co.uk/forum/1/2018/05/08/house_slams_the_brakes_on_pentagons_big_cloud_deal/#c_3508441

    Such is the predictable nature of human behaviour, one can always rely on anything significant and different being weaponised in an attempt to extraordinarily render an expected or promised to be overwhelming inequitable advantage. Or such was the case before SMARTR Enabled Enabling Virtual Machines decide otherwise and Media finds it impossible to disagree without making a complete fool of itself insisting on defending the indefensible and trying to conceal and prevent the truth from airing for Mass Global Peer Review.

  7. JDX Gold badge

    Errata

    Stockfish is not "a chess computer". It's open-source software you can install on your PC or mobile device.

    On the BBC I wouldn't comment but one would hope an IT site knows the difference.

  8. GaryCGibson

    Stockholm Syndrome?

    Losing to a machine program may create Stockholm syndrome in humans who want to marry it. Ideas about evolution tend toward sanguinity quite unrealistically. Artificial intellects will eat humanity's lunch as an appetizer.

  9. Destroy All Monsters Silver badge
    Windows

    Stanislaw Lem wrote about this in 1964

    In 1964: Summa Technologiae

    The book is a bit of a slog though.

  10. Alan Bourke

    Deep Blue != AI

    There is no AI.

    1. I.Geller Bronze badge

      Re: Deep Blue != AI

      Question: All right, would you like to talk about moral issues?

      Answer: OK ...

      Question: May I ask you to spare some chage?

      Original Result

      [2.8% Fyodor Dostoevsk_108]

      Esteeming, and so to say, adoring you, I may at the same time, very well indeed, be able to dislike some member of your family

      The technology is public, see USPTO.

      Above you see how I found answers on Definition questions. Verify my quest please?

  11. anthonyhegedus Silver badge

    It's artificial but it's not intelligent

    These chess programs can beat a human at chess. That's not AI. That's beating humans at chess. Totally different. Humans can play chess, discuss where to eat afterwards, walk their dog and then help design a machine that can play chess. Humans also invented the game of chess in the first place. That's all intelligence. Humans invented the game of chess for various reasons. If 'software' ever invents a game, t won't be for any other reason than it was told to by programmers.

    1. I.Geller Bronze badge

      Re: It's artificial but it's not intelligent

      Exactly. These chess, GO, War-Craft programs are algorithms, while AI is database. For example, do you remember? Is your memory a biological database? Do you think (employing your personal algorithm) - using this database?

      The proved practically answer is "Yes". Proofs: IBM Watson and Google (PA Advisors v Google) - they use sets of phrases.

      AI database contains sets of patterns, which are analogues of neuronal sets in human brain. AI database contains annotated patterns, which the annotations help to index and find. For instance, each word into AI database is annotated by its dictionary definition (or another text).

    2. BananaPeal

      Re: It's artificial but it's not intelligent

      what if we show it all the GOTY games and it makes a new GOTY better than the rest? with like aritificially made cut-scenes and its own synthesized voice acting? what then? give up and plug me in?? aaaaaggghh!!!

  12. I.Geller Bronze badge

    Can chess or GO algorithms make a penny? No.

    Can chess or GO algorithms make a penny? No.

    AI as database has already made billions.

  13. Ken Moorhouse Silver badge

    Re: They learn by making mistakes and correcting their errors.

    I would be more inclined to say:-

    They FUNCTION by making mistakes and WEIGHTING their errors.

    Many of us here will have written programs that "learn" how to win at Tic Tac Toe. Trivial perhaps, but anchors one's imagination as to how AI "works".

    If Deep Mind had been forced to play chess where White starts by moving the Kings rook pawn, then black does the same, then every combination from there, then it would never have got to the nitty gritty of beating Kasparov, there would be just too many meaningless moves in its database. So it was no doubt primed with thousands of classic openings which set it on the correct learning "tree structure".

    If Kasparov had gamed the system by starting with a non-standard opening, that would have forced Deep Mind to throw away its entire Opening Move catalogue and force it to play from scratch, which would have put it at a big disadvantage.

    1. I.Geller Bronze badge

      Re: They learn by making mistakes and correcting their errors.

      Who cares? These are games, they have nothing in common with reality. AI finds answers, see IBM Watson. Can DeepMind Go do the same?

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like