Bitcoin Forum

Bitcoin => Project Development => Topic started by: BenRayfield on March 14, 2014, 06:29:54 PM

Title: Team of a million people rematch with Deep Blue - Humans can take back chess!
Post by: BenRayfield on March 14, 2014, 06:29:54 PM
Trillions of dollars per year are spent to connect pairs of words/phrases in peoples minds. Here's a better way to do it and what to use it for, just to get started.

The plan is a new kind of wiki which spreads our communications in order of priority.

Priority of a communication is on average the amount of computing time, or money spent to rent a computing grid or specialized hardware or to pay someone to figure out how to do it more efficiently, spent looking for a "proof of work", similar to what backs the multi billion dollar bitcoin economy, but this is not money because you can only spend it at time of creating it. It cant be paid as tax because by design it cant move from where its created. So really, its not money, but it is a demonstration of your seriousness about the communication you attach it to. There was an idea for stopping most spam email, to require a "proof of work" calculation be done, to waste 15 second of your computing time to disadvantage those who send huge numbers of spam. Same idea in bitcoin and this wiki.

The protocol for spreading communications through the network by priority is well defined. When asked about any specific short text, for as many texts/boards/objects as fits in the list on screen or as few as are found, for each pair give the 2 communications (connecting the pair) with the hardest-to-create (arbitrarily we set the target as the bigger 256 bit numbers if its SHA256 hashcode), the biggest For (plus) and the biggest Against (minus).

What do the communications do?

They're not disorganized like a bunch of separate paragraphs or sentences by different people. They are each connecting a pair of things. Choose any 2 short texts, like "julianAssange" connect to "isARapist" or connect to "heWasFramedForRape", whichever you want to communicate to whoever may read it. If you want more people to read it, make sure you leave your computer on long enough doing "proof of work" that not too many are above you in the "julianAssange" list. I'm not stupid, not going to give those in power a reason to attack this chess wiki, so in the first version its only for short texts, up to a paragraph, and no document size things will be accepted through the network. You can of course link to it wherever it may be, but the important thing is to demonstrate any huge number of people can work together to do something far smarter than anyone alone, extremely smarter, beat Deep Blue the world chess champion nobody dares challenge anymore. No one person or divided minds could defeat deep blue.

A million, or maybe far less, people can defeat deep blue by connecting things like "the knight is trying to fork our rook and king a few moves after we capture the queen", optionally digitally-sign the message with any private-key, and have your computer put some serious weight on it as much as you think people need to see that. If your slow computer isnt enough, consider asking for investments from some of the bitcoin millionaires who are likely to see a rise in value if we take back the chess crown from deep blue using a "proof of work" calculation as priority, which backs the multi billion dollar bitcoin economy. Theres lots of people who would find motivation in proving a million people can act together to do something far greater than any alone. Google hasn't challenged deep blue. Our world leaders know they can't either. Wall street with their computing grids doesn't want to get embarrassed. But a million people using a proof of work prioritized wiki to talk about and weight on connections between many snapshots of ways the chess board could be (the deep chess tree/network)... Countries have fallen from less.

Its a simple design. I'm just thinking about the details of how it should be seen on screen and how people might expand the system to consider more digital identities (public keys) than you would see in the list of any short text you click, and click some more to navigate as the chess boards and other objects come on screen connected by lines.

For example, maybe people would save large numbers of digitally-signed connections between pairs of things (like 2 consecutive turns with pieces moved, or text written about them) in a bittorrent file and give the link within the network, which gets included in other files, and so on. The chess wiki doesnt have to do everything. Build your own global democracy and talk to billions of people about why or why not we should cut all military budgets in half to use on some other thing we may agree is more important, like refunding the excessive tax or speeding up the Internet or preventing people from starving to death instead of killing them. Lots of things people might connect in the network of short texts with "proof of work" and digital signature.

Normally we expect such debates, even on a wiki, to be quickly forgotten. Consider how that would change if a million people take the chess crown back from deep blue but the world leaders and wall street and google and all the supercomputer grids couldnt. As I see it, that would be a clear demonstration that the people of Earth are capable of more than "mob rule" and should be deciding how the world is run, in the form of online debates similar to the chess wiki. I'm not proposing any political office. That's now how things get done. Use expansions of such a chess wiki to debate with millions of eachother, whoever has some of the relevant skills, what FoldIt/FoldingAtHome did for the science of protein folding, turned it into a puzzle game, and what EyeWire did for brain scans with people telling the computers where the cell boundaries are... Crowdsourcing is a better kind of democracy if it can beat deep blue.

The move we make is whichever connection is strongest from the current board setup to a next move, both seen as small chess boards on screen with a line between them. If the million people work together and talk wiki style, linking whats relevant, weighting it by how long you have computer do "proof of work" or an economy it generates like it did with bitcoin, then the connections For minus the connections Against (you get to choose plus or minus) are highest for the best move. You may want to count digital signatures (who is for or against that connection) like normal voting instead of vote by computing power or money to buy computing power, but the reality is money is any trade of the perception of value, voting has value, so voting is a kind of money.