quantum computer. will it hurt chess theory?

Theory wasn't hurt a lot. Computers agree with a lot of stuff humans came up with... of course there are also disagreements, but this is also interesting and good for the game. People learn from computers.

-

As far as I know it's wrong to think of quantum computers as faster versions of classical computers. Quantum is only faster in specific cases. I assume it wont be possible to write a chess program that runs faster on quantum than classical.

-

When computer chess was first becoming very strong, there was a worry that the way super strong computers played would be incomprehensible e.g. a lot of shuffling around that didn't make sense to the human observers, but so far that hasn't been the case, and with the most recent big breakthrough (neural network based engines) the play only became more human, not less.

Yeah i agree. But alphazero learnt chess the human way, just quicker(very more). Quantum computer are supposed to see things different. They will apprehend simultaneous realities at the same time. Thus see simul endgames in theorical variations. I know, that s what gm think of when they have to chose between accepting or refusing an exchange in the middlegame so it s still "human approach". But even in cosmology, or medecine, we don t see what quantum computer will be able of as they obviously will be so powerfull in calculation. So what about a "simple 8x8 game?

AZ didn't learn chess the human way, machine learning is a sophisticated "guess and check" type of method. While it's true quantum effects allow a system to simultaneously consider multiple paths (so to speak) it's not like a GM considering how an early decision affects the endgame... and as I said it's not even clear that it's possible to write a chess playing program that makes use of how quantum computers work (I assume it's not possible).

Computers don't "apprehend" anything. They crunch numbers.

Alpha Zero is a machine learning program...it was given the parameters of chess and then told to play itself millions of times and track the success/failure of each game.  It figured out the best way to play chess by trial and error...the difference being that unlike human beings, AZ doesn't forget or misinterpret the lessons it learned.

It is notable for those who try to say chess is approaching a draw that AZ bootstrapped itself to the level it plays at in under a day, showing pretty clearly that in centuries of human chess playing we have not learned as much as we think we have ...

But they are *quantum* computers!  That also means if you take a quantum computer and send it to Mars, another quantum computer on Earth can play chess with it, because of quantum computer chess entanglement...

I'm not really sure why people take science/technology and turn it into New Age pseudo-science.  Stop reading Blogging sites, that's my best advice.

Won't it be instantaneous if they are quantum entangled? 

Special relativity doesn't rule out travelling faster than light. It just rules out the possibility of reaching faster-than-light speeds by accelerating.

Alcubierre's method of FTL travel (for example) does not violate special relativity.

Raise your hand if you think engines will impact chess at YOUR level?

Any program that can run on a classical computer can be rewritten to run on a quantum computer. The quantum computer 10^14 times faster, as it can perform a a trillion operations at he same time, i.e. simultaneously instead of sequentially. There now exist 128 qubit quantum computers that are equivalent of 2^128 = 1 petabit conventional computers. IBM offers cloud computing service on a quantum computer. There are already desktop quantum computers for sale for $ 5000.
Quantum computers may help to generate 8-men and 9-men table bases.

Almost none of this is currently true in practice. In the current state-of-the-art, quantum decoherence renders existing experimental quantum computers almost useless for practical purposes. 

Some day, perhaps, quantum computing will be a major factor. But not all theorists are even convinced of that much. Dyakonov, for example, is intensely skeptical; estimating that in order to suppress the data degradation due to decoherence, we would have to have some way of manipulating 10^300 simultaneous parameters... an unlikely feat in a universe that contains only ~10^78 particles in total.

As Blueemu said, this is very much not what experts are saying.

Very interesting topic. I would love to be able to play like a computer. But maybe I should focus on just getting a little better before trying to learn anything from computers playing. I am only 800 rapid.

The strongest chess entities are stronger than Alpha Zero.

It is amazing what Alpha Zero did but that does not make it the strongest chess entity.

I would think qubit computation would allow two quantum computers to always make a draw, unless they somehow differ in their evaluation of positions.  Perhaps the real revolution will not be greater computing ability, but rather in evaluations, and new ways of understanding positions beyond plus, minus, and equal, etc. 

There are theorems about this
https://en.wikipedia.org/wiki/Church%E2%80%93Turing_thesis 

https://en.wikipedia.org/wiki/Quantum_Turing_machine 

Bottom line: any program that runs on a conventional computer can be re-written to run on a quantum computer.
E.g. it is possible to re-write a chess engine like Stockfish to run e.g. on the quantum computer of IBM for cloud computation.
The quantum version does not have to look at the possible moves one by one, but it can look at all possible moves at once.
Such an engine running on a quantum computer should be able to win TCEC defeating the conventional programs running on conventional hardware because it could look so much deeper in the same time as it can calculate so much faster with its parrallel instead of sequential calculation.

Those wikipedia entries say nothing of the kind.  One covers a theory from 1930 (clearly not one about quantum computing ...), and the other talks about a quantum Turing machine ala the one in WWII that cracked the Enigma encryption (as seen in The Imitation Game)...which is cryptography, something quantum computing is excellent for.

You're going to need to do more than watch some TED talk on quantum computing if you are going to take this argument any further...

The theory was already developed by Turing, Church, Shannon, von Neumann and others long before feasible hardware was available. You can run any program on any computer. It is possible to run Stockfish on a mechanical or penumatic computer and it will be very slow and you can run a re-written Stockfish on a quantum computer and it will be very fast.
https://en.wikipedia.org/wiki/Quantum_programming