Tops GMs are already barely capable of understanding top engine play at this point, so there's no real danger of engines "ruining" chess except in a completely perceptual way.
quantum computer. will it hurt chess theory?
Sort:
btickler Really? You know how well top GMs understand or don't understand top engine play?
Where did you get this knowledge?
Jun 25, 2021
0
#64
Yeah btickler, Ponz is the only one allowed to claim special divine insight into the nature of chess and what GMs know and don't know.
llama47 I was asking where he got his knowledge of what top GMs understand ?
However you are very incorrect in your statement that I claim special divine insight into the nature of chess and what GMs know and don't know?
Since you know your statement is untrue--it is ,just an ad hominem attack.
However I do have some insight as to how some GMs think because I have played some of them and know some of them and have given advice to some of them.
Jun 25, 2021
0
#66
Who the heck will know all the computer solutions to every position? Maybe a person with photographic memory who plays in the top 10 in the world. And that’s a BIG maybe. The rest of us, for sure, don’t have to worry about any of that.
Jun 25, 2021
0
#67
ponz111 wrote:
llama47 I was asking where he got his knowledge of what top GMs understand ?
Why did you use a question mark? That sentence is not a question.
Jun 25, 2021
0
#68
ponz111 wrote:
However you are very incorrect in your statement that I claim special divine insight into the nature of chess
Is taking everything literally part of dementia?
llama47 In English - question marks can be quite ambiguous.
llama47 you should know that taking everything literally is Not part of dementia.
Let's get back to chess and the forum question?
Jun 25, 2021
0
#72
Is this going to turn into the forum about chess being a draw or not between you guys again?
Jun 25, 2021
0
#74
btickler wrote:
tygxc wrote:
Sesse used 48 processors 2 GHz Skylake-SP and 18 processors 2.2 GHz Broadwell-EP to reach 49,582,886 nodes per second.
A 128 qbit quantum computer is equivalent to 1000 billion processors and hence would go much deeper in the same time.
Hardware is commercially available e.g. from IBM
Quantum programming languages are available e.g. QCL or Silq
Open source chess engines are available e.g. Stockfish
A translation of an open source chess engine into a quantum programming language is not yet available (StoQfish ?)
Referring to the earlier analogy to airplanes quantum computing is past the 1903 Kitty Hawk stage and rather in the 1927 Pan Am stage.
This is what QCL provides you:
- Data types
- Quantum - qureg, quvoid, quconst, quscratch, qucond
- Classical - int, real, complex, boolean, string, vector, matrix, tensor
- Function types
- qufunct - Pseudo-classic operators. Can only change the permutation of basic states.
- operator - General unitary operators. Can change the amplitude.
- procedure - Can call measure, print, and dump inside this function. This function is non-invertible.
- Built-in functions
- Quantum
- qufunct - Fanout, Swap, Perm2, Perm4, Perm8, Not, CNot
- operator - Matrix2x2, Matrix4x4, Matrix8x8, Rot, Mix, H, CPhase, SqrtNot, X, Y, Z, S, T
- procedure - measure, dump, reset
- Classical
- Arithmetic - sin, cos, tan, log, sqrt, ...
- Complex - Re, Im, conj
- Quantum
...now go ahead and write a subroutine that evaluates a chess position from that matrix-heavy instruction set. We'll be waiting. Don't forget that the evaluation must be completely done *inside* the subroutine, and callouts or temporary measurements or even variable storage for retrieval later actually result in decoherence and loss of the whole matrix.
P.S. I would have given you an instruction set for Silq, but there isn't one published that I could find, just a PDF abstract about it that compares it to Q# (Microsoft's quantum programming language) and says it's better without showing why. The abstract *does* talk about "uncomputation" and the massive difficulty of working with quantum computers where taking any intermediate measurement of any kind during processing destroys the whole operation being performed.
Quantum computers are going to be excellent at solving massively parallel discrete problem sets. Tablebases and position evaluations are not going to qualify without some currently non-existent breakthrough. Even if you could "evaluate" a chess position by simply creating an 8x8 matrix and rotating and flipping bits, you cannot store the results in any intermediate way to compile your overall answer, so you would need to do the entire computation of 10^46.7 positions in one parallel "shot", with no decoherence. Compared to that criteria, current quantum computer capability is far, far less than the Kittyhawk analogy . It's like comparing a single cell organism to a blue whale.
I dont know about all of that and im pretty bad with computers honestly but ive been following quantum computers since i joined this website a little over a year every single day, and id bet you all of my money that quantum computers will be able to compete with classical computers fairly soon in terms of our lifetimes. If i were more learned in computers i could safely narrow it but yeah. They already designed computers that can handle decoherence and uh about not being able to store information theres something called quantum memory where they use a beam splitter to store the information of the entangled pair but send one photon dont a fiber optic wire again dont quote me and i have it all bookmarked i think lol but theres no decoherence because they cool the computers to temperatures colder than space. But. theres many competing companies and some have invented ways to operate at close to room temperature i believe. There are alot of fields that are going to unite because of this technology very soon. Including the greedy and probably evil military unfortunately
Jun 25, 2021
0
#75
Dont quote but you mustve missed the latest breakthroughs unless im a gullible idiot i think they are implying that decoherence isnt the issue (?) but actually noise is which is just outside influence like temperature change etc. kind of the same thing. I mean obviously its an issue but its not insurmountable because they are in the process of developing quantum repeaters which uses entanglement to store a photon pair without having to tamper with it (?) I mean just do some research. Scientists arent just developing single computers they are attempting to develop networks, but yeah all im saying is that it doesnt sound like theyve hit a dead-end by anymeans they sound pretty hopeful. Whether or not i can be utilized for chess id literally bet you all of my money lol which is like 300 hundred dollars
I'm not sure why people that have no background always seem to think that they are booked up on some technology because they read some blogs/articles about it. Those blogs and articles are often about sniffing out investors, and they are written by lay people who don't really know anything more about it than the readers do...they are heavily exaggerated, vague on details, and light on actual accomplishments/proof.
When Scientific American comes out with an article about quantum supremacy being a done deal, then we can talk 
. Until then, all the post grad "studies" and "papers" put out by people trying to make a name for themselves are effectively worthless, and the buzzword articles and blogs are even worse. Not because Scientific American is the arbiter of anything, but because they won't pounce until some breakthrough actually happens that is agreed upon by consensus.
I could write up a passable language from the instruction set myself and publish it...it wouldn't mean squat.
Nobody has said that quantum computing is a "dead end". It's a fantastic technology, and has a ton of potential. What they have said it is not really applicable to the problem of solving chess unless there are new and unforeseen breakthroughs.
Hoping for such breakthroughs without any solid and imminent reason to is akin to believing in magic.
Jun 25, 2021
0
#77
btickler wrote:
I'm not sure why people that have no background always seem to think that they are booked up on some technology because they read some blogs/articles about it. Those blogs and articles are often about sniffing out investors, and they are written by lay people who don't really know anything more about it than the readers do...they are heavily exaggerated, vague on details, and light on actual accomplishments/proof.
When Scientific American comes out with an article about quantum supremacy being a done deal, then we can talk . Until then, all the post grad "studies" and "papers" put out by people trying to make a name for themselves are effectively worthless, and the buzzword articles and blogs are even worse. Not because Scientific American is the arbiter of anything, but because they won't pounce until some breakthrough actually happens that is agreed upon by consensus.
I could write up a passable language from the instruction set myself and publish it...it wouldn't mean squat.
Somewhat related:
The Gell-Mann Effect, also called the Gell-Mann Amnesia Effect describes the phenomenon of an expert believing news articles on topics outside of their field of expertise even after acknowledging that articles written in the same publication that are within the expert's field of expertise are error-ridden and full of misunderstanding.
---
I found this interesting for a slightly different reason i.e. the more you know about something the more you're aware of how much misinformation there is on the topic.
ponz111 wrote:
btickler Really? You know how well top GMs understand or don't understand top engine play?
Where did you get this knowledge?
Elo ratings. You should look into them. A 2800 rated chess player cannot fully understand the moves made by a 3500 rated player...and before you even go there, no, a top correspondence chess player is not higher rated, they are effectively usually a master-level player utilizing a handful of 3500 rated engines (whose moves they understand *even less* than the super GMs do).
But yeah, please keep yourself from turning this into yet another "Ponz claims chess is a draw" thread .
Jun 25, 2021
0
#79
I mean i didnt say there was some breakthrough all im doing is bringing up various aspects of quantum computing to see what youll say about them, im not good with computers i wouldnt really have any certainty as to why you couldnt use it to develop chess software or w.e. I thought id like to hear your thoughts on quantum memory and why thats either made up or not applicable, and whether decoherence is actually a problem because i dont think it actually is as big of a problem as you make it out to be
Jun 25, 2021
0
#80
An article from scientific american https://blogs.scientificamerican.com/observations/decoherence-is-a-problem-for-quantum-computing-but/ Which this is probably less informed than the articles i get in my daily news. Many of the articles i read are from the people personally involved but even here they dont claim decoherence is an insurmountable wall.
Exactly. So solving chess would not lead to people giving up on certain openings or playing 1.c3