4 ms·

Since the difficulty is periodically changed based on the hashing rate, it makes an interesting case. Everybody would generally be better off to throttle the r

by yason 13y ago

Since the difficulty is periodically changed based on the hashing rate, it makes an interesting case. Everybody would generally be better off to throttle the rate because the difficulty would tone down as well and you could still make the same money with regular, much cheaper hardware. But there would be more people doing that which would spread the gains more evenly. On the other hand, that's not how it will ever work because everyone is locally optimizing for their own profit and buying larger racks -- first big GPUs and now ASIC miners -- which ironically doesn't result in profits proportionally equal to the larger investments. In the end it shouldn't even matter: mining bitcoins is just a way to distribute the 21 million coins to people and the cpu power required to do that hasn't got anything to do with the final value of a single block. Your cost to acquire one bitcoin just varies based on how much competition you have.

vog 13y ago
> mining bitcoins is just a way to distribute the 21 million coins to people No. This is a common misconception. Mining Bitcoins is necessary to keep the network running and secure. Distributing Bitcoins is just a side effect to encourage people putting number-crunching power at the network in the early stage, as this is what makes the network secure in the first place. The additional coins will become tiny and vanish over time. They are continually replaced by transaction fees. Over time, the Bitcoin users will pay the "mining people" directly, via transaction fees, for doing all that hard number crunching with their specialized hardware.
- yason 13y ago
  I basically know how it works, that was just my interpretation of the scheme and I stick to it. But I think I used the term mining in an unorthodox way. What I meant by mining is getting reward coins from doing the number-crunching required to find good hashes for blocks. I think the current semantics is that mining means doing that hash-finding number-crunching but it's also being used to refer to mining bitcoins which is maybe a bit misleading. Thus, I rephrase my statement: I maintain that bitcoin rewards themselves aren't necessary —— the rewards are just an incentive and indeed a convenient way to distribute bitcoins in a rather stable way. When we hit the 21 million mark and have no new bitcoins left, the number-crunching part would still have to continue. This is no problem because of transaction fees will provide a sufficient incentive for that. And this brings us back to my point that theoretically we could've started with just that. Theoretically, we could've started with distributing 21 million bitcoins to 21 million random people and let the peer-to-peer network start with doing only transaction verification, and incentivize that with transaction fees. (Probably wouldn't have worked in reality but that's only because of the human factor.) That's what I meant.
humbledrone 13y ago
> Everybody would generally be better off to throttle the rate [...] You're wrong about this. The Bitcoin network would be _much_ worse off (verging on useless) if the global hashrate was throttled. The purpose of mining is not, as you said, to distribute Bitcoins. The purpose is to make it extremely difficult to append a forged block to the blockchain. The higher the hashrate, the more difficult forgery is. The reason that so-called miners receive Bitcoins is because the hashing takes resources (e.g. machines and electricity), so there needs to be some incentive to perform this service. The fact that it happens to distribute Bitcoins is convenient, but ancillary. (There are other ways distribution could have been done.) Read this link for further information on why you want the global hashrate to be as high as possible. Basically, if a single attacker could match the hashrate of the network, they could do very bad things: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_...
- sktrdie 13y ago
  Good point. It's a known misconception that the act of mining serves only to create new BTCs. Since Bitcoin clients only accept history branches that have the longest proof-of-work, the mining is essential to prevent a forge of the history branch, not just a single forged block since most users have established good practice to wait at least 5 confirmations before approving a payment.
- yason 13y ago
  You're wrong about this. The Bitcoin network would be _much_ worse off (verging on useless) if the global hashrate was throttled. No, it doesn't make a difference because what matters is that the target difficulty is relative to the current number-crunching capacity. A lower hashrate is just fine if nobody is able to do better than that. Suppose that Bitcoin happened in the mid-90's when Pentium would be the fastest processor, running at 100MHz. The network would run 1000 or 1000000 times slower than today, the difficulty would basically adjust to that, nobody could overtake the network no more than today, and everybody would be as happy as today. What keeps that from happening is the fact that the hashrate can't be throttled because of the incentives that make it profitable to maximize your hashrate and your number-crunching capacity. THus, the resulting hashrate and, consequently, the required hashing difficulty follow the bleeding edge of the technology. The difficulty is an arbitrary factor: it's just a community decision that in order to find a good hash for a block, the hash must be lower than X, and that decision is based on the current capacity of the network to prevent some single party from "taking over the blockchain" and to keep the rate of new blocks steady. But from blocks-per-Joule perspective there's no point requiring such a huge computational capacity. Bitcoin itself would work just as fine if the difficulty was so easy that a Pentium I could find suitable hashes in a reasonable time: the problem is that we have better hardware than Pentium I so we're forced to raise the difficulty to match the most computationally capable parties.
micaeked 13y ago
Blocks will continue to be mined after the 21 million is up. The incentive from that point forward is txn fees, which the miner gets to keep.

by yason 13y ago

vog 13y ago

> mining bitcoins is just a way to distribute the 21 million coins to people No. This is a common misconception. Mining Bitcoins is necessary to keep the network running and secure. Distributing Bitcoins is just a side effect to encourage people putting number-crunching power at the network in the early stage, as this is what makes the network secure in the first place. The additional coins will become tiny and vanish over time. They are continually replaced by transaction fees. Over time, the Bitcoin users will pay the "mining people" directly, via transaction fees, for doing all that hard number crunching with their specialized hardware.

yason 13y ago
I basically know how it works, that was just my interpretation of the scheme and I stick to it. But I think I used the term mining in an unorthodox way. What I meant by mining is getting reward coins from doing the number-crunching required to find good hashes for blocks. I think the current semantics is that mining means doing that hash-finding number-crunching but it's also being used to refer to mining bitcoins which is maybe a bit misleading. Thus, I rephrase my statement: I maintain that bitcoin rewards themselves aren't necessary —— the rewards are just an incentive and indeed a convenient way to distribute bitcoins in a rather stable way. When we hit the 21 million mark and have no new bitcoins left, the number-crunching part would still have to continue. This is no problem because of transaction fees will provide a sufficient incentive for that. And this brings us back to my point that theoretically we could've started with just that. Theoretically, we could've started with distributing 21 million bitcoins to 21 million random people and let the peer-to-peer network start with doing only transaction verification, and incentivize that with transaction fees. (Probably wouldn't have worked in reality but that's only because of the human factor.) That's what I meant.

humbledrone 13y ago

> Everybody would generally be better off to throttle the rate [...] You're wrong about this. The Bitcoin network would be _much_ worse off (verging on useless) if the global hashrate was throttled. The purpose of mining is not, as you said, to distribute Bitcoins. The purpose is to make it extremely difficult to append a forged block to the blockchain. The higher the hashrate, the more difficult forgery is. The reason that so-called miners receive Bitcoins is because the hashing takes resources (e.g. machines and electricity), so there needs to be some incentive to perform this service. The fact that it happens to distribute Bitcoins is convenient, but ancillary. (There are other ways distribution could have been done.) Read this link for further information on why you want the global hashrate to be as high as possible. Basically, if a single attacker could match the hashrate of the network, they could do very bad things: https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_computing_power https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_...

sktrdie 13y ago
Good point. It's a known misconception that the act of mining serves only to create new BTCs. Since Bitcoin clients only accept history branches that have the longest proof-of-work, the mining is essential to prevent a forge of the history branch, not just a single forged block since most users have established good practice to wait at least 5 confirmations before approving a payment.
yason 13y ago
You're wrong about this. The Bitcoin network would be _much_ worse off (verging on useless) if the global hashrate was throttled. No, it doesn't make a difference because what matters is that the target difficulty is relative to the current number-crunching capacity. A lower hashrate is just fine if nobody is able to do better than that. Suppose that Bitcoin happened in the mid-90's when Pentium would be the fastest processor, running at 100MHz. The network would run 1000 or 1000000 times slower than today, the difficulty would basically adjust to that, nobody could overtake the network no more than today, and everybody would be as happy as today. What keeps that from happening is the fact that the hashrate can't be throttled because of the incentives that make it profitable to maximize your hashrate and your number-crunching capacity. THus, the resulting hashrate and, consequently, the required hashing difficulty follow the bleeding edge of the technology. The difficulty is an arbitrary factor: it's just a community decision that in order to find a good hash for a block, the hash must be lower than X, and that decision is based on the current capacity of the network to prevent some single party from "taking over the blockchain" and to keep the rate of new blocks steady. But from blocks-per-Joule perspective there's no point requiring such a huge computational capacity. Bitcoin itself would work just as fine if the difficulty was so easy that a Pentium I could find suitable hashes in a reasonable time: the problem is that we have better hardware than Pentium I so we're forced to raise the difficulty to match the most computationally capable parties.

micaeked 13y ago

Blocks will continue to be mined after the 21 million is up. The incentive from that point forward is txn fees, which the miner gets to keep.