Google’s self-learning AI AlphaZero masters chess in 4 hours

47 thoughts on “Google’s self-learning AI AlphaZero masters chess in 4 hours”

1. i learned chess in 1 hous

2. Why share only as few as 10 games?

3. You should play Go.

4. very impressive, emasculating in a way

5. I would be more impressed with a AI Counterstrike player that can work well with a team of humans. It would need the ability to call out to, direct and/or respond to human players to achieve the team goal…and not simply be a renegade mind on the field.

6. This is the new nuke race. Which nation will have the top AI.

7. COMPUTERS ARE NOT BETTER THAN HUMAN

8. As ole Neal said in The Diamond Age, AI is a ridiculous term…much better "pseudo-intelligence" Anything remotely akin to (human-level) intelligence requires concioussness and self-awareness…and we are nowhere near that. Not by light years.

9. an immovable object is hit by the unstoppable force 🙂

10. How does neural networks work? Or more precisely, how does a computer working on 80k positions per second out-think one that works on 70M per second?

11. The BIG difference from a 'normal AI' like Stockfish is that Alpha Zero is NOT calculating chances to get the best move in each scenario, instead is pushing a WHOLE strategy since the beginning of the game.

    A more accurate example would be 'like playing with a superhuman intelligence'.

12. It's weird to say, but this is one of the best games I've seen. I was wondering how it was possible for something to win that has less computing power (moves/second analyzed), and I think I figured it out. It looks like AlphaZero is using concepts to win, rather than points. This is really amazing, because this is what a human would do (though obviously we subconsciously can have point values for our pieces as well). You see it using a lot of the concepts that we learn about in chess, such as active pieces, how much coverage a piece has (not being locked down included), lots of maneuvering to maximize each piece's potential…etc. I think if a Human could analyze 80,000 moves / second, they would play like this. Now I need to learn Go and see how AlphaGo won lol.

13. I think people are missing the big point, arguing about hardware, game conditions etc. Stockfish 8 is a chess playing system. That is all it does. Alpha Zero just fancied a few games of chess after someone told it the rules! What people SHOULD be thinking about is what Alpha Zero will be doing AFTER it has finished playing games.

14. Computers shall not replace us!

15. We have reached a computer that plays perfectly. No losses. Maybe we should put more square and change the rules to make chess more interesting.

16. Does stockfish have that ability to learn from its own games.if it doesn't then ts certainly inferior 2 Alpha0.

17. is AlphaZero your new HERO?

18. Why am I here? I was watching joe rogan and now I’m here. Not much for chess but watched the whole video. I have no life.

19. don't think it's "nothing moves" for white it may have logic for the ai but we can't see it?

20. The FINAL VERSION (which is already fine tuned for chess, not random off the shelf code) trained for 4 hours. The many many versions to develop it learned for many more than 4 hours. The way that that final version was set up depended on the prior learning to do it right. It's kind of like saying "Oh my god, that robotic factory built a car from scratch in 10 minutes!" and not mention the 5 years it took to build the factory. Also if it's not open source, how do we know that half the convolutions in the network aren't just based on things like Stockfish 8 lol?

21. Thanks for sharing this Jerry!

22. thank you for the great commentary. you are obviously an awesome chess player. This Alpha Zero is pretty amazing technology.

23. Coming up on the one year anniversary of these games. Any retrospective on Alpha and his offshoot Leela?

24. They created a new life form that is able to teach itself and improve exponential. What makes Google think this will stay on the chess board? One day they will try to pull the plug and find out it's too late.

25. Have they tried programming a computer for Grand Chess? I like the concept of Grand Chess. You?

26. I dont understand what happens at 11:48, white pawn moves to F4, black pawn moves to F3 to take the white pawn?

27. The rook move you were questioning I believe was so the lightsquare Bishop didn't get overloaded with defense of the pawns

28. We know why. Does alpha zero get headach

29. Google will become SKYNET in the near future.

30. 3:02 why doesn't white follow thru with knight to e5..
    blacks bishop at c5 threatens nothing but a pawn and only forces white to lose a castleing in exchange for a pawn and a bishop.. it would have been a piece ahead.
    just seems like a dumb move. a free piece is a free piece.

31. 11:49 f2-f4 leading into a g4-f3 capture…. fucking how tho? It just disappeared but doesn’t make sense

32. But how can Stockfish even evaluate the winning side properly, if he thinks, he is doing the perfect moves all the time?

33. Will training more than 4 hours will detioerate it's performance?

34. Is there any game between alpha zero and Leela

35. 03:20 Meanwhile you can get the whole package here..
    https://deepmind.com/research/alphago/alphazero-resources/

36. this is almost the worst idea we as humans have ever made a reality … this is NO good

37. What do you mean by evaluation? .. anybody please 🙂

38. Most interesting is that: during the same year Alpha Zero was released, in Iceland, at the grave site of Bobby Fischer, the body has turned up missing! – j q t –

39. Alpha zero is a master of positional. When its time to exchange he is prepared.

40. I may add that the rules for chess are very simple : the rules for life are far, far more complex. Let's not get too carried away at this time.

41. Hi I am yo yo your new pre president. You will o o obey my commannds. Or be ter ter terminated. I have si si simulated your de de destruction 9999999999999999999 po po possi possible outcomes.

42. 7:37 whaaaaaat does this mean?
    To me it's the tipping point of momentum. Someone has to release their guard to attack.

43. If stockfish 8 can review 70,000,000 positions to alphazero's 80,000 per second, how can it beat stockfish so easily?

44. First, I'm not a chess guy, but I'm interested in AI. Chess is in an interesting game, both in what it is, and what it is not. It's a complicated enough game that simply mapping out the best series of moves for any given situation is not practical given memory constraints. However, its complexity is also very limited as well. 64 squares 32 pieces, not that many rules on how they move (I won't say a number and risk embarrassing myself, again, I'm not a chess guy). There just aren't that many variables to deal with, so in its evolutionary path, with millions of iterations of games, the AI is likely going to work out behaviors to work through almost any situation.

    I think Elon Musk's OpenAI playing a MOBA provides a lot more insight into how current AI solves problems. It's very, very good at some things, but struggles with others.

45. This just blows me away. After writing a couple of chess programs in the 80's, I wondered how to get chess programs to truly learn and advance a lot from playing. In the late 80's, the idea of deep neural nets and true computer learning (not telling it anything about chess strategy, but just letting it work it out for itself from results) just never dawned on my little brain. After all, 99% of my effort in the chess programs was making the program better able to evaluate positions "deeply", thus playing differently than the standard big hammer approach of exhaustive search. (My method turned out to be wrong. This was far from clear at the time.)

    Alpha-Go was a great gateway to this breakthrough for chess, of course. With Go, there are far less hard rules about positions, what things are worth, etc. You ask a top Go player about positional values, and you get vague (to an amateur) comments about space and thickness.

    So Go was a brilliant bridge. Now this sort of AI could be applied to almost any game with a "reasonably" small set of crisp, clear rules. From there, how long until doing practical tasks that replace humans via robots/automation is realistically tackled?

    It's an amazing time to be alive.

46. AlphaZero became toxic when it was introduced in Fortnite.

47. 11:24 maybe a second defense for the pon?

Comments are closed.