From the referenced document where OP says "No, this is not wasteful" [0]:
"Miners currently use approximately only 0.0012% .. [and] economic modelling predicts miners will not account for more than 0.74% of the energy consumed by the world".
Expending nearly 1% of all energy consumed by the world is an insanely large figure. How much hubris does it take to call this "not wasteful" on its face?
[0] Miners currently use approximately only 0.0012%
Well to make any reasonable comparison we need to see it at a human scale number like dollars per person per day or something otherwise it's just posturing and politics... Big numbers are always manipulated to serve one interest or another
To make a reasonable comparison you would actually need to look at the opportunity costs. This gets even harder as bitcoin mining operations frequently locate at locations which have extremely cheap power. Power which would have huge transmission loss when moving to individual's homes.
The model I used to calculate 0.74% assumes Bitcoin's market cap is $1 trillion(!).
We can all agree that IF the market cap gets this large, it will mean Bitcoin will have had a huge positive impact on society. Therefore it will have been worth it to spend 0.74% of the world's energy.
Edit: I am puzzled by the downvotes. I would like to hear (and respond to) counterpoints if what I am saying is incredibly silly.
Firstly, I think you are underestimating the societal impact that would have led a cryptocurrency to be worth $1 trillion. It would mean it has succeeded on a massive scale: foreign trade made more fluid thanks to Bitcoin's disregard of political borders, hundreds of millions of people using it to escape inflation/financial restrictions/etc (vs. today's few million users), millions of merchants accepting it (vs. today's 130 000 merchants). All Bitcoin would have to accomplish to justify using 0.X% of the world's energy is to increase the world GDP by 0.X%.
Secondly, 0.74% is an upper bound. It assumes power costs of $0.01 per kWh (read section 2 at http://blog.zorinaq.com/?e=82) when actually the average power cost worldwide is 10x higher ($0.10/kWh). Maybe it will be 0.3% of the total energy. Maybe 0.1%. How low is low enough that you would consider it worthwhile?
There is no level of hash calculations that is acceptable.
It is crazy to have a currency that requires this dumb make-work of hashing which results in CO2 emissions. We live in a world of dangerous climate change and Bitcoin users want us to expel CO2 into the atmosphere with every currency transaction. That is dumb and not sustainable.
I actually address this point in my post: "Many/most miners use hydroelectric power because competition compels them to use cheap renewable energy instead of other polluting power sources such as coal power plants" examples:
Question: does the model account for general increase in energy use, within the timespan of (and, possibly, also as a consequence of) Bitcoin growing to reach that market cap, or is 0.74% just the absolute amount of energy as a cut of today's total usage?
If the latter, 0.74% seems less alarming, as it would surely be less when growth is accounted for, especially considering that some of this growth would likely (hypothetically) be driven by an increase in both production and consumption as a direct consequence of the use of Bitcoin, with all the benefits that brings.
But either way, it does still seem moderately alarming, albeit as a kneejerk reaction to the figure without much critical analysis. It's a very interesting economic question as to whether investing this order of energy consumption into Bitcoin would be worth it, both for individuals and for society(ies). I expect many an economics PhD will be minted on this very topic!
Additionally, can only improve with increasing processor efficiencies over time too (unless this is also predicted for in the model) - though saying that increasing processing power is accounted for in Bitcoin itself, with the work getting progressively harder, so this may be moot
1 trillion is the current GDP of Australia, a modest szed industrial nation. 1% of energy consumption for just that much currency is not an all worth it.
Australia consumes 5831 petajoules [1] or 1620 terawatt-hours yearly. In my model miners would consume a comparable amount (actually less): 800 terawatt-hours. Therefore Bitcoin and the Australian economy extract comparable economic value from energy.
IOW if you think Bitcoin is not worth it, then Australia is not worth it :) Thanks for this interesting comparison!
No, Bitcoin creates more than just heat: it created thousands of jobs (see section "Bitcoin already benefits the economy" in http://blog.zorinaq.com/?e=82).
A better argument here is that we must compare it to the current banking system. Sum up the cost of supporting banks, bankers, bank failures, runs on currency, etc. and bitcoin would clearly a good deal efficiency-wise if it was our main currency.
But banks provide credit which Bitcoin does not. Most of the problems with banks (bank failures etc) comes from the credit side, not the payments side.
At a glance, it might seem that simple. In more depth, people who try to hold money essentially don't have a place they can easily park and access it other than a bank. People get forced into a fractional reserve system. That fractional reserve system in turn needs a lot of extra work and systems behind it to function. Bitcoin is a way for people to keep and move money without being subject to either money printing or fractional reserve. So it is, in part, a replacement for that system.
Banks and Bitcoin serve radically different purposes. (Try getting a mortgage from Bitcoin!) A more appropriate comparison would be the percentage of bank energy that goes to payment processing, which is probably pretty low.
What about all those air conditioned buildings with all those eating, breathing, commuting human beings everywhere? Bitcoin looks pretty competitive to me against that.
One major thing other than move money that bitcoin does is print it. If you also look into the complexities and inefficiencies where our money/debt come from I think you'll find they've very high indeed.
I agree that Bitcoin's energy costs risk becoming ridiculously high, but general IT energy expenditures are currently 10% of the world's total, eclipsing aviation [1].
Bearing this in mind, using things like PHP, Python and Ruby to build massive web services also seems massively wasteful to me. I'm quite sure that compared to a lower level language like Go or C++, those languages use up an order of magnitude more energy. The energy difference between those two language groups almost certainly exceeds Bitcoin's total energy usage right now, and I don't see people moaning about this.
I mean, build your prototype using one of those languages, but once you starting building a significant audience, switch over to C++, Go, or Rust, and use a fraction of the energy you use with Python or Ruby.
It'd be cheaper to generate the energy cleanly than to spend the cost (development, opportunity, etc...) of developing such things in a lower level language, I'd imagine.
There is a difference between wasting and spending.
It is impossible to get the security guarantees that Bitcoin affords without spending energy. That expense is the end goal of proof of work, not a side effect.
You are right on this one, bitcoin is a large an noneffective resource waste. Our world need to have ideas to make an energy transition, and bitcoin is based on energy waste, mainly produced by carbon resources.
Our future world need energy, and we cannot afford burning real material to create a virtual money with no practical value, this is a nonsense.
As hydropower is clearly the cleanest and cheapest source of electricity and would be used first in any location, bitcoin miners are essentially shifting other power consumers to carbon-based sources.
I don't think they are shifting other consumers. More and more often we have seen that large miners either build their own mini-power plant[1] or use excess hydro power in remote areas that would otherwise be unused[2].
"Although administrative access to the cell still uses single-DES for encryption of network traffic, that single-DES key is time limited and may be resistant to brute-force key recovery prior to its expiration."
This wasn't created until after CloudCracker can break DES, but to be honest OpenAFS is not the only one either. DOCSIS I think do something similar too. I wonder if anyone has created ASICs to break DES.
> COPACOBANA, the Cost-Optimized Parallel COde Breaker, is an FPGA-based machine which is optimized for running cryptanalytical algorithms. COPACOBANA is suitable for parallel computation problems which have low communication requirements. DES cracking is such a parallelizable problem: an exhaustive key search of the Data Encryption Standard (DES) takes no longer than a week on average with COPACOBANA. Other ciphers can be attacked too, and COPACOBANA can also be used for parallel computing problem outside cryptography.
Even with afs3-rxkad-k5 the wire encryption is a watered down variant of DES called fcrypt and the use of encryption is the choice of the client. There is nothing that an administrator can do to prevent AFS3 clients communicating with AFS3 servers from leaking sensitive information onto the wire. The AuriStor File System solves these problems with the yfs-rxgk security class, volume and file server security policies, combined identities, perfect forward secrecy, and much more ....
Well this image proves that the bitcoin network is not only alive and well but it is growing at an impressive rate and anyone saying that bitcoin is dying is lying to you and possibly pushing his own agenda.
I will admit I am not up on block chain tech or the process it uses, but it does seem like a lot of wasted resources to just compute hashes. It seems like someone could make a system where real work units count as credit in an issuing lottery. Kind of like SETI @ Home, medical or genome work where you distribute a problem and then people get tickets for accomplishing some form of work. Seems to me that this would be a better use of such vast computing power and would help with discovery at the same time. Again it's not my area of expertise so I don't understand the details or issues that would arise in ordering the work units around distributed computation to solve large problems. It would be nice if researchers had open and free access to such a system.
The detail you're missing is that these hashes bitcoin computes are designed to be something easy to verify but hard to compute. Relatively few productive problems have this property, and there are some which are doing this already, trying to crack proof RSA keys and such... but that's about it.
These are not "wasted" resources; they create value, which is the trust for the distributed currency system. These are hashes based on real history of transactions, and their computations guarantee that these transactions have been recorded by third party.
(At least, that's how it should work if I understand the theory behind bitcoin correctly).
Exactly. It's like claiming that HTTPS encryption is wasted resources, because all you do is encrypt things at one end, only to decrypt at the other.
Bitcoin's computation does the job of securing the blockchain by forcing an attacker to spend computation cycles to match up the computation done for these hashes.
> HTTPS is the means to an end, bitcoin is just wasting resources.
Bitcoins is also a means to an end, you just don't seem to understand what the end is.
> If the point was just to maintain the blockchain integrity there would be no need to make its difficulty ever increasing.
This is incorrect. The blockchain requires miners to establish an unbiased ledger history. The difficulty parameter is tweaked so that the whole system can be future proofed against an adversary creating a convincing blockchain history (which was biased to double-spending or retracting a transaction).
> This is incorrect. The blockchain requires miners to establish an unbiased ledger history. The difficulty parameter is tweaked so that the whole system can be future proofed against an adversary creating a convincing blockchain history (which was biased to double-spending or retracting a transaction).
What was the excuse in the early days of bitcoin then when the difficulty was orders of magnitude lower?
Difficulty is ever increasing just because of the financial model bitcoin tries to propagate. The author of bitcoin wanted to make it deflational and so it had to be this way.
That part has nothing to do with the integrity of the blockchain and you are mixing two concepts here: Blockchain, which was a great invention and bitcoin, a financial experience (nothing against experiences from my part) which turned out to be a waste of resources.
> > This is incorrect. The blockchain requires miners to establish an unbiased ledger history. The difficulty parameter is tweaked so that the whole system can be future proofed against an adversary creating a convincing blockchain history (which was biased to double-spending or retracting a transaction).
> What was the excuse in the early days of bitcoin then when the difficulty was orders of magnitude lower?
... because there wasn't as much hashing power in the bitcoin network? Isn't that obvious? The difficulty parameters are tweaked so that the proof of work takes 10 minutes for the whole network. As the network gets bigger, you need to make it harder to maintain a 10 minute proof of work.
> Difficulty is ever increasing just because of the financial model bitcoin tries to propagate. The author of bitcoin wanted to make it deflational and so it had to be this way.
Eh. You're full of shit. But that's your prerogative.
> That part has nothing to do with the integrity of the blockchain and you are mixing two concepts here: Blockchain, which was a great invention and bitcoin, a financial experience (nothing against experiences from my part) which turned out to be a waste of resources.
Integrity of the blockchain is maintained by cryptographic hashes. There are other security properties (such as proof of work and thus non-malleability) that are maintained by the difficulty parameter.
The difficulty does not NEED to be ever increasing. In fact there are periods of time where the difficulty is pretty stable, such as mid-2011 to early 2013, or the first half of 2015: http://bitcoin.sipa.be/speed-ever.png
> bitcoin is just wasting resources. If the point was just to maintain the blockchain integrity there would be no need to make its difficulty ever increasing. <
Higher difficulty don't waste any more resources than lower difficulty.
> Bitcoin's computation does the job of securing the blockchain by forcing an attacker to spend computation cycles to match up the computation done for these hashes.
So we could call this approach to scaling communications, "brute-force messaging"? (..since it also forces anyone who wants to enlarge its available message space to pre-compute hashes ...with an ongoing race condition, so "raced, brute-force messaging"?) Scaling seems a bit off though, since its design enforces "expensive, frantic scarcity". That seems rather limited for a growing population's needs for communications around an increasingly valuable resource base.
> ... anyone who wants to enlarge its available message space ...
Bear in mind that the proof-of-work algorithm actually aims to reduce chatter on the network. Every ten minutes, a single Bitcoin full node broadcasts: "Hey, this is the latest state of the ledger, and you can trust it because [insert proof-of-work that was calculated out-of-band]." So, as the network's processing power scales up, the computation difficulty is also scaled up, so that the "message space" required for consensus remains constant.
> That seems rather limited for a growing population's needs for communications around an increasingly valuable resource base.
While the way distributed consensus is achieved (see "Byzantine Generals' Problem") consumes most of the network's processing power, it's what makes the resource valuable. However, since this is done in parallel, it hardly consumes any of the network's bandwidth, freeing up that "message space" for important things like sending/receiving/propagating transactions, which are much less demanding computationally. All that is required there is some elliptic curve cryptography (ie. a valid public/private key pair + any low-powered device with correct software and an internet connection.)
The problem with basing a money on something that has value, is that changes in perception of the value of the thing, result in changes in prices across the economy.
Say gold is the currency base. Now we discover Alzheimer's can be cured with a medicine that requires a lot of gold in the processing, and the price of gold skyrockets. That causes the price of everything else (priced in gold) to plummet.
However, I think a compelling case can be made that the amount of hashpower being used to secure Bitcoin is overkill for the measly 3tps that the network is currently capable of. If the same hashpower was used to secure 100X more transactions, it wouldn't seem like such a waste of resources.
> I will admit I am not up on block chain tech or the process it uses, but it does seem like a lot of wasted resources to just compute hashes.
To be fair, the blockchain is the largest distributed rainbow table in history, so it does have some value in that sense.
But arguing that "it is a waste" to use hashing is to not understand the security properties offered by bitcoin and why it's actually vital to use hashes.
I'm not a math type, but it looks to me that mrb is saying that the number of computed bitcoin hashes outnumber an estimated number of all of the grains of sand on the world's beaches, and in the world's deserts, by a factor of 10^6th
70 comments
[ 0.18 ms ] story [ 145 ms ] thread"Miners currently use approximately only 0.0012% .. [and] economic modelling predicts miners will not account for more than 0.74% of the energy consumed by the world".
Expending nearly 1% of all energy consumed by the world is an insanely large figure. How much hubris does it take to call this "not wasteful" on its face?
[0] Miners currently use approximately only 0.0012%
We can all agree that IF the market cap gets this large, it will mean Bitcoin will have had a huge positive impact on society. Therefore it will have been worth it to spend 0.74% of the world's energy.
Edit: I am puzzled by the downvotes. I would like to hear (and respond to) counterpoints if what I am saying is incredibly silly.
What, why?! Even if that were the case, it doesn't follow that the good it does outweighs using 1% of all energy.
Firstly, I think you are underestimating the societal impact that would have led a cryptocurrency to be worth $1 trillion. It would mean it has succeeded on a massive scale: foreign trade made more fluid thanks to Bitcoin's disregard of political borders, hundreds of millions of people using it to escape inflation/financial restrictions/etc (vs. today's few million users), millions of merchants accepting it (vs. today's 130 000 merchants). All Bitcoin would have to accomplish to justify using 0.X% of the world's energy is to increase the world GDP by 0.X%.
Secondly, 0.74% is an upper bound. It assumes power costs of $0.01 per kWh (read section 2 at http://blog.zorinaq.com/?e=82) when actually the average power cost worldwide is 10x higher ($0.10/kWh). Maybe it will be 0.3% of the total energy. Maybe 0.1%. How low is low enough that you would consider it worthwhile?
It is crazy to have a currency that requires this dumb make-work of hashing which results in CO2 emissions. We live in a world of dangerous climate change and Bitcoin users want us to expel CO2 into the atmosphere with every currency transaction. That is dumb and not sustainable.
BitFury: https://www.businesswire.com/news/home/20151211005837/en/Bit...
KnC: http://thenodepole.com/2015/06/04/kncminer-to-build-new-20mw...
Therefore your worry about CO₂ emissions is misplaced.
Well that's just unreasonable. Do you have any earthly idea how much CO2 the current banking system expends?
If the latter, 0.74% seems less alarming, as it would surely be less when growth is accounted for, especially considering that some of this growth would likely (hypothetically) be driven by an increase in both production and consumption as a direct consequence of the use of Bitcoin, with all the benefits that brings.
But either way, it does still seem moderately alarming, albeit as a kneejerk reaction to the figure without much critical analysis. It's a very interesting economic question as to whether investing this order of energy consumption into Bitcoin would be worth it, both for individuals and for society(ies). I expect many an economics PhD will be minted on this very topic!
Additionally, can only improve with increasing processor efficiencies over time too (unless this is also predicted for in the model) - though saying that increasing processing power is accounted for in Bitcoin itself, with the work getting progressively harder, so this may be moot
IOW if you think Bitcoin is not worth it, then Australia is not worth it :) Thanks for this interesting comparison!
[1] http://www.industry.gov.au/Office-of-the-Chief-Economist/Pub...
Bearing this in mind, using things like PHP, Python and Ruby to build massive web services also seems massively wasteful to me. I'm quite sure that compared to a lower level language like Go or C++, those languages use up an order of magnitude more energy. The energy difference between those two language groups almost certainly exceeds Bitcoin's total energy usage right now, and I don't see people moaning about this.
I mean, build your prototype using one of those languages, but once you starting building a significant audience, switch over to C++, Go, or Rust, and use a fraction of the energy you use with Python or Ruby.
1. http://www.theregister.co.uk/2013/08/16/it_electricity_use_w...
It is impossible to get the security guarantees that Bitcoin affords without spending energy. That expense is the end goal of proof of work, not a side effect.
Our future world need energy, and we cannot afford burning real material to create a virtual money with no practical value, this is a nonsense.
No, many/most miners use hydropower. See next-to-last paragraph in section 1.
[1] http://thenodepole.com/2015/06/04/kncminer-to-build-new-20mw...
[2] https://bitcoinmagazine.com/articles/the-unknown-giant-a-fir...
"Although administrative access to the cell still uses single-DES for encryption of network traffic, that single-DES key is time limited and may be resistant to brute-force key recovery prior to its expiration."
This wasn't created until after CloudCracker can break DES, but to be honest OpenAFS is not the only one either. DOCSIS I think do something similar too. I wonder if anyone has created ASICs to break DES.
https://en.wikipedia.org/wiki/EFF_DES_cracker
http://www.copacobana.org/
> COPACOBANA, the Cost-Optimized Parallel COde Breaker, is an FPGA-based machine which is optimized for running cryptanalytical algorithms. COPACOBANA is suitable for parallel computation problems which have low communication requirements. DES cracking is such a parallelizable problem: an exhaustive key search of the Data Encryption Standard (DES) takes no longer than a week on average with COPACOBANA. Other ciphers can be attacked too, and COPACOBANA can also be used for parallel computing problem outside cryptography.
https://www.auristor.com/openafs/migrate-to-auristor/
http://bitcoin.sipa.be/speed-lin-ever.png
It's a cryptocurrency where you have to prove that you are retaining a copy of a file. Whitepaper included.
These people are also behind IPFS, the distributed alternative to HTTP (ambitious, huh?)
But yes, that could totally be done! Only problem is everyone would have the result, which I don't think they'd like too much.
(At least, that's how it should work if I understand the theory behind bitcoin correctly).
Bitcoin's computation does the job of securing the blockchain by forcing an attacker to spend computation cycles to match up the computation done for these hashes.
HTTPS is the means to an end, bitcoin is just wasting resources.
If the point was just to maintain the blockchain integrity there would be no need to make its difficulty ever increasing.
Bitcoins is also a means to an end, you just don't seem to understand what the end is.
> If the point was just to maintain the blockchain integrity there would be no need to make its difficulty ever increasing.
This is incorrect. The blockchain requires miners to establish an unbiased ledger history. The difficulty parameter is tweaked so that the whole system can be future proofed against an adversary creating a convincing blockchain history (which was biased to double-spending or retracting a transaction).
What was the excuse in the early days of bitcoin then when the difficulty was orders of magnitude lower?
Difficulty is ever increasing just because of the financial model bitcoin tries to propagate. The author of bitcoin wanted to make it deflational and so it had to be this way.
That part has nothing to do with the integrity of the blockchain and you are mixing two concepts here: Blockchain, which was a great invention and bitcoin, a financial experience (nothing against experiences from my part) which turned out to be a waste of resources.
> What was the excuse in the early days of bitcoin then when the difficulty was orders of magnitude lower?
... because there wasn't as much hashing power in the bitcoin network? Isn't that obvious? The difficulty parameters are tweaked so that the proof of work takes 10 minutes for the whole network. As the network gets bigger, you need to make it harder to maintain a 10 minute proof of work.
> Difficulty is ever increasing just because of the financial model bitcoin tries to propagate. The author of bitcoin wanted to make it deflational and so it had to be this way.
Eh. You're full of shit. But that's your prerogative.
> That part has nothing to do with the integrity of the blockchain and you are mixing two concepts here: Blockchain, which was a great invention and bitcoin, a financial experience (nothing against experiences from my part) which turned out to be a waste of resources.
Integrity of the blockchain is maintained by cryptographic hashes. There are other security properties (such as proof of work and thus non-malleability) that are maintained by the difficulty parameter.
Higher difficulty don't waste any more resources than lower difficulty.
So we could call this approach to scaling communications, "brute-force messaging"? (..since it also forces anyone who wants to enlarge its available message space to pre-compute hashes ...with an ongoing race condition, so "raced, brute-force messaging"?) Scaling seems a bit off though, since its design enforces "expensive, frantic scarcity". That seems rather limited for a growing population's needs for communications around an increasingly valuable resource base.
Bear in mind that the proof-of-work algorithm actually aims to reduce chatter on the network. Every ten minutes, a single Bitcoin full node broadcasts: "Hey, this is the latest state of the ledger, and you can trust it because [insert proof-of-work that was calculated out-of-band]." So, as the network's processing power scales up, the computation difficulty is also scaled up, so that the "message space" required for consensus remains constant.
> That seems rather limited for a growing population's needs for communications around an increasingly valuable resource base.
While the way distributed consensus is achieved (see "Byzantine Generals' Problem") consumes most of the network's processing power, it's what makes the resource valuable. However, since this is done in parallel, it hardly consumes any of the network's bandwidth, freeing up that "message space" for important things like sending/receiving/propagating transactions, which are much less demanding computationally. All that is required there is some elliptic curve cryptography (ie. a valid public/private key pair + any low-powered device with correct software and an internet connection.)
This is what they posted about 2016: https://ripple.com/insights/2016-will-be-the-year-you-realiz...
Say gold is the currency base. Now we discover Alzheimer's can be cured with a medicine that requires a lot of gold in the processing, and the price of gold skyrockets. That causes the price of everything else (priced in gold) to plummet.
However, I think a compelling case can be made that the amount of hashpower being used to secure Bitcoin is overkill for the measly 3tps that the network is currently capable of. If the same hashpower was used to secure 100X more transactions, it wouldn't seem like such a waste of resources.
> the bottom line is that investing a hundred+ megawatt in a system that creates thousands of jobs is a valuable economic move, not a waste
1. https://bitcoinmagazine.com/articles/putting-the-blockchain-... 2. https://www.gridcoin.us/ 3. http://www.coindesk.com/5-global-problems-bitcoins-proof-wor...
To be fair, the blockchain is the largest distributed rainbow table in history, so it does have some value in that sense.
But arguing that "it is a waste" to use hashing is to not understand the security properties offered by bitcoin and why it's actually vital to use hashes.
Plus many other good arguments in http://blog.zorinaq.com/?e=82 please read it.
http://tinyurl.com/hf9aybm http://tinyurl.com/qc59ora