"A common misconception is that transactional confidence is based on time"
Transactional confidence is based on the probability of someone compromising the network for enough blocks to undo a spend that you've already honored (IE you gave them what they paid for, then they undo the transaction). At 6 blocks, you are assuming that an attacker won't have the computational power to reverse 6 blocks in a row.
When blocks come out every 2 hours, that means you need to find every single block for 12 hours, or essentially control the blockchain for a long time.
When the block rate is down at 5 minutes, you only need to control the network for 30 minutes - you have more chances to get lucky and undo 6 transactions in a row in your 12 hours of purchased computing time.
A halved block rate does not mean you need less computational power, but it does mean that computational power is probably cheaper and 6 blocks worth of power is going to be more affordable.
Unless your purchased compute time is more than 50% of the network your odds depend mostly on the number of confirmations and not the time they take. The cost of a sub-50% attack increases exponentially with more confirmations but only linearly with longer confirmations, so the number of confirmations is more important.
5 minutes is still a perfectly acceptable block time. It only gets problematic as network latency can be significant in block discovery chances. This happens under a few seconds per block.
There are alt-coins working with 1 minute blocks and they behave just fine. Provided hash power is spread among the network, there is absolutely no problem with this, and it won't be long until we go back to 10 minute-ish blocks.
A successful attack requires more than simply reversing any 6 consecutive blocks. It requires reversing a specific set of blocks containing a transaction one wishes to be reversed.
This is why 5 minute blocks are virtually every bit as secure as 10 minute blocks.
Inaccurate title. The network still sets the difficulty every 2016 blocks ( i.e. 2 weeks) so that it takes 10 minutes on average to create a block.
Right now, however, people are adding so much computational power in that intervening two weeks, that by the end of those two weeks it takes 6 minutes to create a block, not 10.
This is presumably due to custom hardware mining rigs coming online since they just shipped.
The title was accurate. Blocks are being found at a rate of 1 every 5 minutes. There's no way the next difficulty jump will be enough to bring the speed to 10 minutes since it takes the average time over the last 2016 blocks. It's likely in a couple weeks the time to find blocks will be even lower than it is now if the growth rate continues like this.
If only the bitcoin algorithm could have accounted for incredibly various technical and sociological issues.
Good thing Moore's Law somehow exists and has maintained the curve perfectly despite not being a law, despite somehow not making any reference to the consumer activity that seemingly drives it, and a heavily innovation-constraining duopoly for every chip that is used in desktops/laptops.
Some times I feel like the computer industry really is driven by magic smoke.
>If only the bitcoin algorithm could have accounted for incredibly various technical and sociological issues.
I think they did a pretty decent job given that "incredibly various technical and sociological issues" is the definition of a wicked problem (http://en.wikipedia.org/wiki/Wicked_problem).
Some alt-coins have been destroyed by the in-rush of massive hashing power, which caused a difficulty-adjustment up, followed by an exit of that hashing power, leaving block-progress so slow that it would take 'forever' (impractically long) for the next normal compensating adjustment-down to happen.
Since one of the joys of Bitcoin is using it to exercise one's paranoid imagination, consider scenarios where somehow, an enemy of Bitcoin (or promoter of an alternative) can somehow disable or redirect all that ASIC power in a short time frame.
Perhaps, in cahoots with the ASIC makers, some magic input can cause them to begin malfunctioning (either temporarily or permanently). Or, a cartel controls enough SHA256 power to withhold capacity, and switch to a NewCoin needing the same hashing resources, in rapid coordinated succession. Or both.
It might be hard for Bitcoin-prime to adapt in time, while NewCoin's deep-pocketed backers from the USG, CCP, or evil lectroids of Planet-X tempt miners and Bitcoin balance holders to switch to NewCoin, perhaps via an initial endowment mechanism which requires the irreversible destruction/relinquishment of old Bitcoin balances.
Ah, good times. Good Stephensonian-science-fictional times.
"If only you were doing X instead of Y with your money and time."
E.g. helping african children instead of reading HN.
No one yet figured out which "useful" computation has configurable difficulty, uniform success probability and which is easy to verify without using central authority. Complaining about people for hashing blocks is like being disappointed at someone not doing the charity you like.
For me, miners are doing great job providing security to the blockchain and making transactions confirmed faster than usual. And Bitcoin, in my opinion, will help poor people and protect more wealth and health than a potentially better cancer treatment.
I think Primecoin is interesting; instead of mindlessly computing hashes, it computes Cunningham prime number chains, and has already set a few records doing this as I understand. This might not feed hungry children, but perhaps it has some utility in mathematics.
1. In less developed countries, unfair taxes and harassment and are one of the reasons for poverty. Bitcoin reduces the risks of outsiders interfering.
2. Irrelevant (edit: and slightly disturbing that you think the condition of poor people can only be improved by the welfare system. There is some evidence that the inverse is true.)
3. Poor people in Africa are already using M-Pesa and other electronic currencies. Bitcoin is a less costly alternative to those payment systems.
4. Many guest workers in rich countries are transferring a part of their income to their relatives in their home country, often for large fees. These fees are much smaller when using Bitcoin.
1. How does Bitcoin make taxation more difficult than paper money? How does it make taxation difficult at all? Even if somehow Bitcoin made it hard for a corrupt government to impose taxes, there are a lot of things the government could do to make Bitcoin hard to use (and they only need to ensure that a large number of people have trouble).
2. Welfare is a reality of life, and has been as far back as we have written records. If you have a better system that might actually work, I'm all ears, but until then we need to deal with the reality of this world.
It is also wrong to assume that transaction fees for Bitcoin will remain low forever. As Bitcoin exchanges and other services start complying with financial regulations, the costs of that compliance will be passed on to their customers. Bitcoin is not useful without exchanges in countries that will not accept Bitcoin for tax payments, which at the moment is all countries.
1. Cash can be easily found and seized, not only by corrupt governments but by organized crime (If you think there is a difference). Bitcoin is decentralized and based on cryptography - you only need to have internet access and private keys to make transactions.
2. Free trade and industrialization have a much better track record of lifting people out of poverty than handouts. Even Bono agrees. Bitcoin helps free trade and wealth aggregation, especially in countries with unstable currencies (Argentina, Zimbabwe)
3. As explained in (1), Bitcoin is much harder to control than a centralized system. There is no need for centralized exchanges, everybody with a smartphone and internet access can trade Bitcoin. The fact that government don't accept taxes in Bitcoin does not imply that it will not help poor people, it implies that Bitcoin will not help the government if it chooses to criminalize virtual currencies.
1. Computers can be easily found, seized, searched, backdoored, confiscated, etc. How is that any different than the situation you face with cash?
2. Free trade does not make welfare unnecessary. The point of welfare is not to lift people out of poverty, it is to ensure that people are not starving to death and that they have a place to live. If everyone had to depend on the free market to make enough money to eat, there would always be people who starve to death -- just like there will always be businesses that go bankrupt. In countries with dysfunctional or non-existence welfare systems people do starve to death when they cannot find enough money to pay for their food.
3. Governments do not need to criminalize Bitcoin to stop it from becoming popular. As long as businesses are required to pay taxes those businesses will need to get the money the government accepts tax payments in. Either those businesses will demand the money up front, or they will accept Bitcoin payments and then run to the nearest exchange when the tax collectors come. Either way Bitcoin is only going to be used as long as people have access to Bitcoin exchanges.
Of course, it is possible that the government is so corrupt and so dysfunctional that it is unable to even enforce the tax code. Such governments tend to not last long.
Here are ways people are made less poor with Bitcoin:
1. Sending money over the border to relatives for 0 fee. There are a lot of stories on how people risk their wealth transferring big amounts of hard earned capital across the border.
2. Do not have to pay hidden costs of using traditional payment systems. Costs are for sellers and buyers: 1) transaction fees 2) fraud risk 3) undersupply of goods because of fraud risk or fees. Tons of business models are simply not possible without Bitcoin. And many are not open to anyone in the world because of regulations, fraud risk etc.
3. No cost of inflation. If you save money for buying something expensive or opening your own shop, central bank steals your wealth by printing money and giving it to privileged organisations (that are supposed to pay you back in a form of "growing economy", but we have yet to see a proof for it). For almost 5 years now, Bitcoin value was significantly growing every single year, comparing to any other asset on the market. And even when/if it stops growing, the value of Bitcoin depends on a decision of every market participant, not a small group of people in the Federal Reserve.
4. Those who wish to save wealth by avoiding some taxation, are able to do so. If some poor teenager learns HTML and makes websites for Bitcoin, he doesn't need to lose 20-50% of his income (depending on your country and how you measure the total tax).
5. If you worry about gangsters from main street or wall street stealing your money by force, you can hide it well extremely cheaply. You can remember a single password to re-create a wallet from it. Or back up your wallet on different cloud services. Or give copies to all your friends. It's much easier to protect bitcoins than gold or paper bills. We only need some easy to use software, but it's entirely possible to do, it just takes some time.
The only cost for you is higher perceived uncertainty (however, our shaky financial ponzi scheme created by government only seems stable, but it can collapse anytime and goes in that direction) and a lack of tools and services. But all of that is coming. Every year it becomes easier to use Bitcoin and more people accept them.
Interestingly, custom-made ASICs[1] to perform Folding@Home calculations have also been developed that perform at unprecedented speeds. Of course, there isn't any reason someone outside of academia would be compelled to buy one.
So, introduce a coin variant where 1/1000 or 1/10000 of cycles is donated to folding@home or seti@home or just a random Boinc project. That way, you still make money, but now you get to donate your cycles to charity as well.
It's not easy to do something like that without sacrificing decentralization, which is a cryptocurrency holy cow. It would easily become a competitive advantage to remove the charity cycles from the mining client.
"It's not easy to do something like that without sacrificing decentralization, which is a cryptocurrency holy cow."
Only among the Bitcoin crowd. In the crypto research community people are fine with having a bank issue the money and have published a mountain of research on how to create secure digital cash.
The people who are fine with centralized bank-issued cryptocurrencies haven't launched anything successful at the scale of Bitcoin.
There may yet be hope for one, especially if it has government backing... but right now "the Bitcoin crowd" has a good existence-based case that decentralized-issue is important for adoption.
And all those systems were never deployed. Perception (f the banks, you can print money at home, and the money will grow in value over time) trumps research apparently.
There was, in fact, an effort to deploy those systems, during a period of time when the government was fighting as hard as possible to stop cryptography from "going mainstream." The efforts failed because:
1. David Chaum did not understand when he had won and kept trying to get more money for his startup.
2. There were (and still are) patents on the math that prevented these systems from just being deployed without waiting around for Chaum.
3. The security guarantees of digital cash are not nearly as important as people thought. It turns out most Internet users are not terribly concerned about privacy and banks are good enough at dealing with fraud that the benefit of digital cash does not outweigh the cost of deploying it.
You may argue that none of these apply to Bitcoin, and that is probably why Bitcoin has been so widely adopted. It remains to be seen if the economics of Bitcoin really work at scale or over long periods of time.
The ASICS for Bitcoin mining only do one thing very well: SHA256. They're not general purpose computers in any sense.
The FPGA versions can be theoretically reprogrammed, but any good design will use the maximum number of gates, leaving nothing left over for Folding type applications.
Once equilibrium is reached, we can expect hashing rates to bottom out. A similar rush will occur over time as the cadence of Bitcoin hashing technologies approaches more regularity, not unlike how new generations of CPUs and GPUs are released on a timely and predictable basis.
This will be the pattern of Bitcoin hashing -- waxing and waning with each new generation of ASICs and the lull between each release.
One has to wonder what will happen to all the surplus ASIC hardware as well-financed miners regularly upgrade their hashing infrastructure.
Perhaps sold for very little money to people who have free power. Soon it will be about power efficiency more than raw hashing power like what happened during the switch form CPU to GPU mining.
In order to significantly affect the overall bitcoin issuing rate long-term, the hash rate would have to continue to grow exponentially for an extended amount of time - years. In reality, the hash rate can only grow exponentially while an older mining technology is being supplanted by a newer one. It happened already with CPU to GPU/FPGA, and now it's happening again with GPU/FPGA to ASIC. It will grow exponentially while everyone hashing upgrades to ASICs, but once the hashing market is saturated with them, it will drop back to modest linear growth, as the only growth will be either new miners coming online or modest, linear improvements in ASIC technology.
Haven't looked at this part, but I have a feeling that the difficulty itself wouldn't be able to go high enough to support long-term exponential growth in hashrate.
It's not an exact science since no one knows future network increases, but you can see a detailed view of the thought process around that question here:
Transactional confidence is based on the probability of someone compromising the network for enough blocks to undo a spend that you've already honored (IE you gave them what they paid for, then they undo the transaction). At 6 blocks, you are assuming that an attacker won't have the computational power to reverse 6 blocks in a row.
When blocks come out every 2 hours, that means you need to find every single block for 12 hours, or essentially control the blockchain for a long time.
When the block rate is down at 5 minutes, you only need to control the network for 30 minutes - you have more chances to get lucky and undo 6 transactions in a row in your 12 hours of purchased computing time.
A halved block rate does not mean you need less computational power, but it does mean that computational power is probably cheaper and 6 blocks worth of power is going to be more affordable.