Nice! Now, next puzzles are solved very quickly, sometimes it takes hours, and sometimes days. But I guess soon it will take weeks, months, or even years, to sweep next addresses (and it will no longer be profitable to fight for thousands of satoshis, because the reward will be worth much less, than the effort needed to get it). Soon, additional incentive will be needed, to keep grinding. I thought for some time, how it could be done, and now I think I know: the solution is a transaction. First, all sponsors prepare their own inputs, signed along with one or more puzzles, and they all use SIGHASH_NONE. Then, the difficulty and amount depends on picked puzzles, and the sum of all inputs. Because everyone signs everything (excluding puzzles) with SIGHASH_NONE, all inputs are signed, so it is impossible to steal the deposit, without solving the puzzle (the signature is invalid outside of signed set of inputs). And also, it doesn't commit to any outputs, so they can be freely chosen by the solver. Then, partially signed transaction is shared anywhere. It can be done on forum, on some mining pool, over some P2P network, or anywhere else. Solvers can validate partially signed transactions, which would contain all needed inputs, and no outputs. Then, they will add any outputs they want, and grind solutions, signed with SIGHASH_ALL, to protect it from being stolen (and their Proof of Work will make it resistant enough to any double spends, if re-mining the solution will take more time, than getting a single confirmation). I think I should make some experiments in test networks first, to make sure, that this model is safe, but it looks promising. And it is definitely much better, than depositing every coin on the same address, as existing puzzles, because then, it requires providing Proof of Work for each coin separately (which raises the difficulty to N*puzzleDiff).
Nice! Now, next puzzles are solved very quickly, sometimes it takes hours, and sometimes days. But I guess soon it will take weeks, months, or even years, to sweep next addresses (and it will no longer be profitable to fight for thousands of satoshis, because the reward will be worth much less, than the effort needed to get it). Soon, additional incentive will be needed, to keep grinding. I thought for some time, how it could be done, and now I think I know: the solution is a transaction. First, all sponsors prepare their own inputs, signed along with one or more puzzles, and they all use SIGHASH_NONE. Then, the difficulty and amount depends on picked puzzles, and the sum of all inputs. Because everyone signs everything (excluding puzzles) with SIGHASH_NONE, all inputs are signed, so it is impossible to steal the deposit, without solving the puzzle (the signature is invalid outside of signed set of inputs). And also, it doesn't commit to any outputs, so they can be freely chosen by the solver. Then, partially signed transaction is shared anywhere. It can be done on forum, on some mining pool, over some P2P network, or anywhere else. Solvers can validate partially signed transactions, which would contain all needed inputs, and no outputs. Then, they will add any outputs they want, and grind solutions, signed with SIGHASH_ALL, to protect it from being stolen (and their Proof of Work will make it resistant enough to any double spends, if re-mining the solution will take more time, than getting a single confirmation). I think I should make some experiments in test networks first, to make sure, that this model is safe, but it looks promising. And it is definitely much better, than depositing every coin on the same address, as existing puzzles, because then, it requires providing Proof of Work for each coin separately (which raises the difficulty to N*puzzleDiff).

Nice! Now, next puzzles are solved very quickly, sometimes it takes hours, and sometimes days. But I guess soon it will take weeks, months, or even years, to sweep next addresses (and it will no longer be profitable to fight for thousands of satoshis, because the reward will be worth much less, than the effort needed to get it). Soon, additional incentive will be needed, to keep grinding. I thought for some time, how it could be done, and now I think I know: the solution is a transaction. First, all sponsors prepare their own inputs, signed along with one or more puzzles, and they all use SIGHASH_NONE. Then, the difficulty and amount depends on picked puzzles, and the sum of all inputs. Because everyone signs everything (excluding puzzles) with SIGHASH_NONE, all inputs are signed, so it is impossible to steal the deposit, without solving the puzzle (the signature is invalid outside of signed set of inputs). And also, it doesn't commit to any outputs, so they can be freely chosen by the solver. Then, partially signed transaction is shared anywhere. It can be done on forum, on some mining pool, over some P2P network, or anywhere else. Solvers can validate partially signed transactions, which would contain all needed inputs, and no outputs. Then, they will add any outputs they want, and grind solutions, signed with SIGHASH_ALL, to protect it from being stolen (and their Proof of Work will make it resistant enough to any double spends, if re-mining the solution will take more time, than getting a single confirmation). I think I should make some experiments in test networks first, to make sure, that this model is safe, but it looks promising. And it is definitely much better, than depositing every coin on the same address, as existing puzzles, because then, it requires providing Proof of Work for each coin separately (which raises the difficulty to N*puzzleDiff).

Nice! Now, next puzzles are solved very quickly, sometimes it takes hours, and sometimes days. But I guess soon it will take weeks, months, or even years, to sweep next addresses (and it will no longer be profitable to fight for thousands of satoshis, because the reward will be worth much less, than the effort needed to get it). Soon, additional incentive will be needed, to keep grinding. I thought for some time, how it could be done, and now I think I know: the solution is a transaction. First, all sponsors prepare their own inputs, signed along with one or more puzzles, and they all use SIGHASH_NONE. Then, the difficulty and amount depends on picked puzzles, and the sum of all inputs. Because everyone signs everything (excluding puzzles) with SIGHASH_NONE, all inputs are signed, so it is impossible to steal the deposit, without solving the puzzle (the signature is invalid outside of signed set of inputs). And also, it doesn't commit to any outputs, so they can be freely chosen by the solver. Then, partially signed transaction is shared anywhere. It can be done on forum, on some mining pool, over some P2P network, or anywhere else. Solvers can validate partially signed transactions, which would contain all needed inputs, and no outputs. Then, they will add any outputs they want, and grind solutions, signed with SIGHASH_ALL, to protect it from being stolen (and their Proof of Work will make it resistant enough to any double spends, if re-mining the solution will take more time, than getting a single confirmation). I think I should make some experiments in test networks first, to make sure, that this model is safe, but it looks promising. And it is definitely much better, than depositing every coin on the same address, as existing puzzles, because then, it requires providing Proof of Work for each coin separately (which raises the difficulty to N*puzzleDiff).