Bitcoin is unsustainable
July 2, 2015 5:38 PM   Subscribe

Bitcoin is unsustainable Bitcoin's power usage per transaction isn't remotely sustainable as a wholesale replacement for the conventional financial system.

Computer cooling firm Allied Control estimates the total power consumption of the Bitcoin network at 250 to 500 Megawatts. Looking at the total hashrate, which is the number of calculations the network can perform per second, and applying a generous miner efficiency of 0.6 watts per gigahash, we can estimate our own back-of-the-envelope Bitcoin network constant power draw at just under 215 MW, although this figure is always in flux (it’s important to note that many of the variables in my calculation are constantly changing slightly). That’s around enough zap to power 173,000 average American households’ daily electricity usage.

With about 110,000 transactions per day, that works out to 1.57 households daily usage of electricity per Bitcoin transaction. Yes, every time you buy something in Bitcoin, you could be using as much electricity as 1.57 American families do in a day.
posted by modernnomad (82 comments total) 23 users marked this as a favorite
 
Of course, it wouldn’t be fair to knock Bitcoin’s electricity consumption without comparing it to payment systems most people use today. Let’s take VISA as an example.

According to Network Computing, the VISA network can process more than 80 billion transactions per year or 2,537 transactions per second, using two mirrored data centers, each capable of running the entire network. The larger data center is currently pulling enough power for 25,000 households’ daily electricity, so we’ll double that to account for VISA’s total draw. In 2013, VISA’s investor reports say the company processed 58.5 billion transactions.

Working off these (admittedly imperfect) figures, each VISA transaction consumes around 0.0003 household’s daily electricity use. That makes Bitcoin about 5,033 times more energy intensive, per transaction, than VISA, at current usage levels.

posted by modernnomad at 5:39 PM on July 2, 2015 [15 favorites]


Unsustainable in fiat electricity maybe, but have you heard about my new BitErgs?
posted by klangklangston at 5:49 PM on July 2, 2015 [55 favorites]


215 MW * 24 hours / 110,000 transactions works out to about 46 kWh per transaction. Which, holy shit. Couldn't you do something more economical with that energy, like, I dunno, run an aluminum smelting operation?
posted by indubitable at 5:52 PM on July 2, 2015 [36 favorites]


Go ahead and be cynical. 5 years from now any one of us will be able to buy a flying car from our Linux desktop using a Bitcoin network powered by cold fusion.
posted by skymt at 5:56 PM on July 2, 2015 [28 favorites]


I'm a little confused. Are these energy expenditures including the costs of mining bitcoins? Because that's not the same thing as the costs of transaction, correct? Perhaps I'm missing something here.

Even so, that does sound very energy-intensive for "mining" a purely virtual resource.
posted by Edgewise at 5:57 PM on July 2, 2015 [6 favorites]


Bitcoin also suffers as a currency that is designed to have investment value. If everyone saves, then the value goes up, and nobody wants to spend it, making it problematic as a currency and an investment. If the price is steadily going down, then few would want to accept it as payment, making it problematic as a currency and an investment.
posted by Brian B. at 5:58 PM on July 2, 2015 [4 favorites]


All of that power for 7 transactions/second across the entire network, which is maybe enough to run a gas station chain?

Bitcoin: there is always more and it is always worse. Though I fear we've already passed Peak Hilarity.
posted by fifteen schnitzengruben is my limit at 6:00 PM on July 2, 2015 [6 favorites]


I find it hard to believe that it's even that low or that efficient. I feel like most setups are like this or worse. Those aren't tiny ASIC boxes running on little power bricks, those are plugged in to 6 or 8 pin PCIe cables going in to big desktop power supplies. That's a LOT of current.

I always wonder if the goal of this was to bring on the post oil apocalypse as quickly as possible.

I'm a little confused. Are these energy expenditures including the costs of mining bitcoins? Because that's not the same thing as the costs of transaction, correct? Perhaps I'm missing something here.

The systems mining verify the transactions. I might be wrong, but i'm pretty sure it takes there being systems online and mining to do the verifying.
posted by emptythought at 6:01 PM on July 2, 2015 [3 favorites]


It's not an apples-to-apples comparison, but I was interested in pricing out the cost of making real money (notes) and the cost of making bitcoin-scrip. We can do some back-of-the-envelope calculations to compare on the basis of cost — and then perhaps consider some less tangible details:

1. What the money is spent on in creating paper notes vs cryptoscrip (materials and energy).
2. What pollution and other waste products are generated with each result.
3. What value is generated from each product.

From CNN:

The government produced 6.4 billion new currency notes [in 2010]. Each one cost 9.6 cents to produce, including the cost of paper and printing.

The cost of printing bills in 2010 was 6.4B * 0.096 dollars / 1 year. The average lifespan of common paper note denominations is approximately 6 years. So we get a rough, amortized estimate of $102M / year, give or take fluctuations in material cost etc.

We can perhaps use some stats to calculate the equivalent cost of mining cryptoscrip, using 2010 energy prices and 2015 hardware costs and performance specs.

The current hash rate for one popular form of c.c. is 335M GH/s (cite).

One reportedly "at-current-most-efficient" mining device ("ANTMINER S4") mines 2 TH/s per S4 or 2000 GH/s/S4.

Dividing the hash rate by device rate, all the miners in the world are running the equivalent of 0.168M S4s.

The maximal energy consumption of an S4 device power supply is 1.4 kW. We can be conservative (giving more benefit to miners) and estimate a power supply running at 80% efficiency, or 1.12 kW usage per S4.

The total estimated energy cost of running these S4s is: 0.168M S4 * 1.12 kW/S4 = 0.188M kW.

The average cost of electricity in the United States in 2010 was $0.0983 per kWh.

Thus, the cost to run all these S4s in 2010 would be 0.188M kW * $0.0983 / (kW * hr) or $0.0185M/hr.

Over the course of 2010, therefore, the cost to run the S4s is $0.0185M/hr * 24 hr/day * 365 days/hr = $162M/year.

The cost of the S4s themselves is 0.168M S4s * $1749 per S4 or $293.8M (US).

These have to be replaced on a continual basis to reap efficiencies, given the pace of processing speeds. But let's say we only replace 50% of these S4s a year.

So the cost of mining bitcoin scrip in 2010 would be at least an estimated $162M + $147M or $309M.

Putting aside power and miner costs, I'm not sure exactly how we can estimate the cost of administration staff and time, as well as the cost of the cooling, networking and power infrastructure required to support the computational equivalent of 168,000 S4s. But we can maybe try.

Given capital costs that Google spends on one of its data centers, perhaps the facility charges would average out to $30-60M per year, amortized roughly over a decade. Miners do not operate on the scale of a company like Google, so this is again giving their operation the benefit of Google's size and ability to lower operational costs.

Keep in mind that most mining devices are not listed as being as efficient as an S4, so the equipment costs and electricity usage estimates alone are very conservative, or "best case scenario" efforts, where all of humanity is directed to using the most-efficient and lowest-cost available mining gear, and the electricity or energy cost is based on what the "free market" can provide at the best price.

It looks like the conservative, back-of-the-envelope cost of making paper notes at 2010 prices ($102M/yr) is somewhat less than the cost of making bitcoin-scrip ($340-370M) at 2010 prices (with 2015 hardware).

Some thoughts:

- The CNN article suggests most of the year-to-year increase in making paper notes has been in the increased price for cotton. Once printed, real money is a one-time cost for ~5 years, and its continual exchange keeps generating value (based on taxation).

- Money can be widely and directly exchanged for nearly all goods and services. Cryptoscrip cannot, outside of very small, very specialized marketplaces, and cannot be used to pay taxes or debts.

- Notes and coins can be recycled — cotton can be shredded and incorporated into new paper, and metals can be melted. Cotton is grown by photosynthesis and so its energy costs are entirely due to harvesting through normal farming methods, which are increasingly efficient to the point where only 2-9% of the cost of farming production is in energy.

- Conversely, ASICs have to be replaced continually to gain efficiencies. The end-point of obsolete hardware is generally Chinese and African landfills, polluting air and water with toxic heavy metals and PCBs. Recycling might one day put a serious dent into this, but manufacture of new computing hardware is also a notably "dirty" process. Their operation nearly entirely converts electricity into waste heat in the process of generating calculations. The lifespan of an ASIC seems to create comparatively larger amounts of pollution.
posted by a lungful of dragon at 6:01 PM on July 2, 2015 [28 favorites]


This is good for bitcoin.

Really? I said it first in this thread? Me?
posted by Going To Maine at 6:12 PM on July 2, 2015 [7 favorites]


Is Bitcoin...America?
posted by oceanjesse at 6:13 PM on July 2, 2015 [1 favorite]


In the back of my brain I have always felt that bitcoin was a scheme by Enron.....
posted by mrgroweler at 6:16 PM on July 2, 2015 [2 favorites]


If everyone saves, then the value goes up, and nobody wants to spend it, making it problematic as a currency and an investment.

There's been a pretty hilarious drive among the Bitcoin enthusiasts lately for Greece to somehow adopt Bitcoin as its currency. Which will apparently magically get them out of their financial troubles. Not sure how this will work. For a start, how does everyone in Greece get their hands on the Bitcoins? Solar-powered ASICs powered by the hot Mediterranean sunshine? They may have a point though; no-one's ever going to want to offer loans in a deflationary currency, so I guess it will stop Greece getting into that trouble again...
posted by Jimbob at 6:16 PM on July 2, 2015 [4 favorites]


It's good to see the actual energy consumption spelled out like that. Underlines the obvious, I guess, that Bitcoin can't solve money's problems because money is a big part of the problem. Consumptive continuous growth leads to its own set of crises whatever the medium.
posted by Aya Hirano on the Astral Plane at 6:25 PM on July 2, 2015 [3 favorites]


The funniest thing about all this is that having such a huge number of miners doesn't actually serve any purpose: the system would work equally well with a tenth as many miners, or a hundredth. The Bitcoin network as it exists today is fundamentally unscalable, because every miner has to witness every transaction. Having more miners doesn't mean more throughput, and it doesn't even mean more profit — just more people fighting over smaller slices of the same pie.

The Bitcoin mining activity fluctuates with supply and demand, just like a business. But unlike a business that actually does things or makes things, there's no incremental change in net social benefit when the activity increases, but a massive increase in expenditure. It's just a gigantic demonstration of perverse incentives under the banner of "rational self-interest".
posted by teraflop at 6:25 PM on July 2, 2015 [22 favorites]


Which will apparently magically get them out of their financial troubles. Not sure how this will work.

Probably the same way everything around BitCoin works -- on the surface, wistful thinking, underneath, chicanery.
posted by GenjiandProust at 6:25 PM on July 2, 2015 [10 favorites]


Meanwhile, Dogecoin remains the superior option
posted by the uncomplicated soups of my childhood at 6:26 PM on July 2, 2015 [12 favorites]


I guess we could look at the difference in per-transaction cost between Bitcoin and more traditional money systems as the cost of zero trust. As you can see, it pays handsomely to be able to trust others.
posted by indubitable at 6:31 PM on July 2, 2015 [4 favorites]


I'm a little confused. Are these energy expenditures including the costs of mining bitcoins? Because that's not the same thing as the costs of transaction, correct? Perhaps I'm missing something here.

That is my understanding, though I don't know any of the technical details so someone may correct me on any of this. Back when I was following Bitcoin more closely I saw some people claim that the actual transactions themselves could be handled with something with the processing power of an NES at the usage levels at the time, but it would require more as the number of transactions rose.

The mining was intended to give people an incentive to provide that processing power and create the coins as it went, with an automatic update in difficulty at certain points as it proceeded. The miners blew past expectations and increased the difficulty extremely quickly with all the high powered hardware, and that hardware uses a lot of energy.
posted by Drinky Die at 6:34 PM on July 2, 2015


Out of curiosity, is there any idea of how much wealth inequality there is in Bitcoin? I know the early miners got huge payouts compared to what's getting awarded now.
posted by ckape at 6:38 PM on July 2, 2015


I always thought this was kind of obvious. The energy to create one is absurdly high, but I figured people were just speculating that the cost to do that wouldn't outpace the currency's price.
posted by lownote at 6:44 PM on July 2, 2015


The article already lists the cost of Visa's infrastructure. It's significantly more efficient and powerful than the blockchain can/will ever be, and doesn't require an entire household's worth of electricity to pay for your tab at the bar.
Working off these (admittedly imperfect) figures, each VISA transaction consumes around 0.0003 household’s daily electricity use. That makes Bitcoin about 5,033 times more energy intensive, per transaction, than VISA, at current usage levels.
posted by fifteen schnitzengruben is my limit at 6:55 PM on July 2, 2015


OK, so I'm not a big fan of bitcoin, but maybe if we're comparing costs, add in the fraud losses enabled by Visa's laughable security. There is a real cost to using people's zip codes as passwords.
posted by ryanrs at 7:11 PM on July 2, 2015 [2 favorites]


It's not like the largest bitcoin exchange in the world was brought down by massive fraud or anything.
posted by Justinian at 7:12 PM on July 2, 2015 [28 favorites]


It will be interesting to see what happens to bitcoin once mining ends—or even as the mining reward gets small enough. Without mining, the reward for participating in bitcoin is purely fees, which are offered voluntarily by the payer side of the transaction. If the typical transaction reward is too small (in cost-of-electricity), most of the hashrate will simply leave. But if too much of it leaves, then mounting a 51% attack becomes much more feasible.

.. for instance, the NSA's Utah Data Center is reported "to require 65 megawatts of electricity". If accurate, and the bitcoin electricity expenditure declines to 50MW (10% of the high-end estimate above), then the NSA could own the blockchain just by plugging in commonly-available ASIC rigs in Utah.

and it seems pretty rational to imagine the electricity value of mining dropping 90%. If the value of bitcoin for buying electricity stays constant, but mining fees remain essentially zero and we see four halvings of the mining reward, rationally the pool power would drop to 6.25% of what it is now.

.. and if you don't think it for four halvings, why not 8 (in which rationally the pool power would drop to .4% what it is now)?

Only magic could make the cost of electricity in bitcoins go down by half every 4 years (the approximate rate at which the mining reward is cut in half). Or else it's one hell of a deflationary currency and I really should get myself a few bit-cents now.
posted by jepler at 7:20 PM on July 2, 2015 [2 favorites]


(alternately, bitcoin fees could just go really, really high, like BTCs per transaction, to maintain a high reward level for "miners")
posted by jepler at 7:21 PM on July 2, 2015


Every four years, the mining reward cuts in half, so miners will be incentivized to find more efficient ways to mine bitcoins if they want to make a profit. Miners compete with each other, so the difficulty of mining (which gets harder or easier as the average miner gets better or worse) should average out close to the ability of the most efficient miners.

As the block reward approaches zero, transaction fees become the only way to reward miners. In theory, fees will stabilize near the cost of mining which will be whatever miners can eke out of transaction fees.

(One thing I wonder about is if legitimate miners will be able to make a decent profit at this point. If not, bitcoin-mining malware will be profitable again, which has no incentive to save energy. There's a post here that talks about other difficulties at this point)

But the reason VISA is more efficient is volume. There are about 1-2 bitcoin transactions/sec, and ~2500 VISA transactions/sec. I understand bitcoin would need to make massive changes to handle this rate, but the difficulty of mining is largely independent of transaction volume. If it is able to scale up, it could be power-competitive with VISA. (Of course, they are two different systems with different properties, yadda yadda)
posted by RobotVoodooPower at 7:24 PM on July 2, 2015


Couldn't you do something more economical with that energy, like, I dunno, run an aluminum smelting operation?

Oddly enough, I've seen quite a few posts/claims of mining operations specifically locating themselves where aluminum smelting used to happen, specifically *for* the infrastructure already in place.
posted by CrystalDave at 7:24 PM on July 2, 2015 [2 favorites]


I continue to think Bitcoin is a terrible. Yet it seems to be highly resistant to going away or collapsing. The imminent death of Bitcoin seems to have replaced the imminent death of the Internet as a recurring story that never pans out.
posted by humanfont at 8:03 PM on July 2, 2015


This is BS. It pretty plainly includes all the energy going into mining Bitcoin as part of the transaction cost. That doesn't make any sense. Almost all of the energy involved in Bitcoin is for mining and the energy used for transactions is many orders of magnitude lower.

This doesn't even pass the smell test! 46kWh per transaction? Even where electricity is very cheap, that's a few dollars of electricity per transaction. And yet, no one is paying a few dollars for a Bitcoin transaction. In fact, they are paying a few pennies per transaction, if anything at all.

I totally agree that the Bitcoin mining system is an environmental mess, but that doesn't apply to the transaction system at all.

VICE should be embarrassed to have published this.
posted by ssg at 8:03 PM on July 2, 2015 [1 favorite]


It pretty plainly includes all the energy going into mining Bitcoin as part of the transaction cost. That doesn't make any sense.

The entire point of mining is to support transactions. If we switch off all those miners, can people continue to perform transactions?
posted by Jimbob at 8:28 PM on July 2, 2015 [5 favorites]


Has anyone said buttcoin yet?
posted by Ghostride The Whip at 8:57 PM on July 2, 2015 [3 favorites]


The entire point of mining is to support transactions. If we switch off all those miners, can people continue to perform transactions?

The comparison with Visa only makes sense if you throw in the cost of running the US mint on Visa's side.
posted by pwnguin at 8:59 PM on July 2, 2015


The entire point of mining is to support transactions.

No, the point of mining is mostly to create new Bitcoin. The required transaction part is a very small fraction of the total mining power.
posted by ssg at 9:12 PM on July 2, 2015 [1 favorite]


If mining was not necessary to support transactions, then mining would not be part of bitcoin.
posted by ckape at 9:50 PM on July 2, 2015


The reward for mining is presently the block reward. The point of mining is to commit transactions to the blockchain. Mining has to go on for transactions to happen. The amount of mining that is presently being done is unrelated to the number of transactions processed. If half the miners dropped out, the difficulty factor would decrease and all would continue as it has.

That said, it is perfectly reasonable to count the cost of mining as part of the total cost of each transaction. The person spending the bitcoin isn't paying that cost, but that's because the reward is temporarily coming from somewhere else. Mining is an absolute requirement for the whole thing to work, even after the last coin has been mined and no more bitcoins can be created. At that point mining will become quite unprofitable unless a lot of people turn off their ASIC boxes or people start paying a significant transaction fee. I'd expect that the remaining miners will not include transactions with low fees in their blocks.
posted by wierdo at 10:11 PM on July 2, 2015 [2 favorites]


This is BS. It pretty plainly includes all the energy going into mining Bitcoin as part of the transaction cost. That doesn't make any sense. Almost all of the energy involved in Bitcoin is for mining and the energy used for transactions is many orders of magnitude lower.

For somebody who's such a relentless booster of Bitcoin you don't seem to actually understand how it works.
posted by Pope Guilty at 10:27 PM on July 2, 2015 [12 favorites]


For "mining" defined as seigniorage, and "transactions" defined as the fee for transaction processing, then yes, it does. All of those bitcoins are fungible, so implicitly the ratio of energy used for the two is the same as the ratio of their amounts. Longer comment on the way.
posted by topynate at 10:35 PM on July 2, 2015


"yes, it does make sense", missed the edit window, sorry.
posted by topynate at 10:41 PM on July 2, 2015


"The point of mining" can be interpreted in multiple ways.

For miners, the point of mining is overwhelmingly the production of new bitcoins (25 every ten minutes right now, though this rate will decrease automatically over time to limit the maximum quantity to 21 million). These bitcoins can be sold for money to people who want to have and/or use them. To the miners, this money justifies the cost of energy, network speed, and hardware.

For non-miner users, the point of mining is to confirm transactions. It generally doesn't matter to these users how many miners there are, except inasmuch as more miners make the system more resistant to disruption by a hostile group. To first order, however, the number of miners (and the energy costs) are not strongly coupled to the transaction rate. The users pay a fee of something like ~0-3 cents per transaction, which is meant to establish that the transaction is not spam trying to flood the network. That's all they pay. (Well, you might say that they and we suffer the worldwide consequences of arguably wasteful energy use.)

The difference between these two ways of calculating "cost per transaction" has generated a lot of discussion.

I wonder if it might be useful to use the analogy of mining a certain mineral that might be crucial to a ubiquitous technology in 2050 OR might be utterly useless. Some people mock any discussion of the mineral as ridiculously speculative. Some are stuck on the failure of early (corrupt or legitimate) mineral trading companies. Some object to tremendous energy use in refining a compound that may end up having no practical purpose. And so on.
posted by Mapes at 10:50 PM on July 2, 2015 [3 favorites]


"It's not like the largest bitcoin exchange in the world was brought down by massive fraud or anything"

Well, it's looking now like the fraud may have been perpetrated by rogue DEA agents.

(and like somebody said in another FPP thread: We seem to be living in a William Gibson story.)
posted by I-baLL at 11:10 PM on July 2, 2015 [2 favorites]


Wow, why did I ever stop following Bitcoin news? The crazy train never stops.
posted by Drinky Die at 11:23 PM on July 2, 2015 [8 favorites]


One of the problems facing Greece is that the central bank which is supposed to have the job of providing liquidity and preventing bank runs, has decided to actually encourage bank runs and threaten to stop liquidity, in order to force the government to take specific fiscal measures.

If Greece had a currency and transaction system with no central bank, that would not be possible.

Bitcoin has plenty of potential problems. But one thing it definitely offers is a weakening of the power of international banks and institutions. It's not just people on the political right who are concerned about that power.
posted by TheophileEscargot at 12:33 AM on July 3, 2015


Or, you know, they could have their own central bank that would have more mutual self interest with the nation.
posted by Drinky Die at 12:34 AM on July 3, 2015 [1 favorite]


The USofA tried making our own central bank that had more mutual self interest with the nation... that's why we called it the "Federal Reserve" instead of the "Bank of the United States". (Then somebody else set up the "Bank of America" and confused everybody.) It worked for a while... not so much anymore.
posted by oneswellfoop at 1:43 AM on July 3, 2015


For somebody who's such a relentless booster of Bitcoin you don't seem to actually understand how it works.

For someone who is so eager to disagree with people, you don't seem to be able to actually read what they write.

I'm not a booster of Bitcoin, relentless or otherwise. Obviously, Bitcoin is a divisive subject and plenty of folks here are eager to point and laugh, but perhaps we'd be better of to use a little critical thinking.

I think the mining system is incredibly stupid, shortsighted, and damaging. The computing power that goes into expansion of the Bitcoin monetary supply is a huge and pointless waste.

But to include all that in the transaction cost just doesn't make sense. Is anyone really arguing that the same amount of power will be used for transactions once new Bitcoins aren't being created anymore? Should we expect transaction costs to jump to a few dollars then?
posted by ssg at 5:55 AM on July 3, 2015 [1 favorite]


I'm not a booster of Bitcoin, relentless or otherwise.

Aren't you the guy who shows up in every Bitcoin thread to post endless arguments about how Bitcoin is the best thing ever, over and over again? Am I thinking of somebody else?
posted by Pope Guilty at 6:02 AM on July 3, 2015


Looking at his posting history, you're way off the mark.
posted by daniel_charms at 6:28 AM on July 3, 2015 [1 favorite]



The USofA tried making our own central bank that had more mutual self interest with the nation...It worked for a while... not so much anymore.

Which is why we should have a Jubilee every fifty years. You have to re boot the system now and again to make it work
posted by eustatic at 6:32 AM on July 3, 2015 [1 favorite]


Back on topic, though, the analysis provided in the article seems convincing, but I have a massive issue with the writing style. Instead of an analytical piece, which it should be, it reads like a press release, to the point that I'm having trouble reading it due to the nagging feeling that someone's trying to sell me something.
posted by daniel_charms at 6:44 AM on July 3, 2015


Well, it's looking now like the fraud may have been perpetrated by rogue DEA agents.

They were talking about Mt. Gox. Silk Road wasn't a BitCoin exchange, it was a market that used BitCoins.

And it wasn't brought down by that DEA agent, that DEA ran a unit that was an ancillary part of the investigation that ultimately netted Ulbricht, but the evidence used in the trial was from another unit in New York.
posted by Sangermaine at 8:12 AM on July 3, 2015 [1 favorite]


The comparison with Visa only makes sense if you throw in the cost of running the US mint on Visa's side.

Done.

Seriously, the cost of the U.S. Mint is *nothing*. A few dozen stamping machines, long since paid off, a few die makers, feed stock, mostly nickel. Even if we also throw in the Bureau of Engraving and Printing, the people who make the printed notes, as opposed to the coins that the U.S. Mint does, the cost is still basically trivial compared to the usage factor of that money. There's no continuing transaction cost to a quarter or a $20. You spend them until they're worn out, and while $1 bills do wear out too quickly, because we are stupid and won't do the smart thing* because AMERICA FUCK YEAH.

Sigh. Seriously. If you're going to argue money, at least learn who makes it -- and far more importantly, learn the difference between "making physical money" and "creating money" which are two very different things and someone confusing them means that talking to that someone about monetary policy is a waste of time so I'll stop now.


* Which is to do what everyone else does to retire a paper note in favor of a coin, which is to release a coin and withdraw the paper note.. The moment you pull the $1 bill, vending machines across America will be, almost as if by magic, refitted to accept $1 coins, as will cash drawers.

Ask England, or Australia, or Canada, or any other country that got rid of a note for a coin. The only reason our $1 coins are a joke is we keep printing $1 bills that last less than 18 months, which is why, by far, our biggest note expense is printing $1 bills. We have to churn them out, because we use them all the time and thus, they don't last.
posted by eriko at 8:24 AM on July 3, 2015 [2 favorites]


What is the cost of having a currency, though? It's definitely a positive return on investment for USG, but there must be some expense involved. At minimum, stopping counterfeiting and imprisoning counterfeiters should be counted...
posted by topynate at 8:31 AM on July 3, 2015 [1 favorite]


...and I point out anti-counterfeiting as a cost because the same function is fulfilled by the energy intensive proof-of-work system used by Bitcoin.
posted by topynate at 8:34 AM on July 3, 2015


Is anyone really arguing that the same amount of power will be used for transactions once new Bitcoins aren't being created anymore? Should we expect transaction costs to jump to a few dollars then?

For bitcoin to work, the amount of power has to remain high. As indubitable said, that's the network's replacement for trust. The cost of corrupting the network is the cost of contributing a majority of processing to the network.

Put another way, the total value of transactions that bitcoin can protect is directly proportional to the total amount being spent on power, regardless of whether that power is repaid in "mining fees" or "transaction fees." The network couldn't (eventually) move $1 billion a day while spending $10,000 a day on power, just like you can't keep a $1 million diamond in a bus-station locker that costs $100 to break into.

So what should we expect to happen when all the power has to be paid for by "transaction" fees (which are paid by the parties) rather than "mining" fees (which are paid by ... I don't know, borrowing against future value of the network)?

In the success version of this story, the cost of mounting a 51% attack always stays higher than the value of corrupting the network. Say the network ends up moving $100 billion per year, and users can tolerate transaction costs of 1% before switching to alternatives. So the network can afford to spend $1 billion per year on computation as a wall to protect transaction integrity. What's the value_of_corrupting_the_network? Considerably less than $100 billion, since you can't just steal all the value without destroying it. What's the cost_of_corrupting_the_network? Considerably more than $1 billion, since the first $1 billion bought up all of the convenient aluminum smelting plants near power generators, excess chip fab space, etc.

Bitcoin "works" as long as value_of_corrupting_the_network < cost_of_corrupting_the_network, at a transaction cost that users are willing to pay.

But again, this means we're replacing trust in some institution (like Visa or the banking system) with the act of burning more value as a group than any individual can justify burning.

One interesting escape valve is to use bitcoin processors as space heaters. Bitcoin is only "wasting" energy in the sense that it's turning it into heat. But we turn lots of energy into heat on purpose. Replace all those baseboard heaters with bitcoin ASICs, and get all the secure distributed transactions you want for free -- at least in the winter.
posted by jhc at 9:53 AM on July 3, 2015 [3 favorites]


So in the future, we may be literally warming ourselves with trust. You couldn't make this stuff up.
posted by RobotVoodooPower at 9:59 AM on July 3, 2015 [1 favorite]


One interesting escape valve is to use bitcoin processors as space heaters. Bitcoin is only "wasting" energy in the sense that it's turning it into heat. But we turn lots of energy into heat on purpose. Replace all those baseboard heaters with bitcoin ASICs, and get all the secure distributed transactions you want for free -- at least in the winter.

Is this for real being floated as a solution to the energy use problem? If so, why not go whole hog and turn processors into pizza ovens?
posted by Aya Hirano on the Astral Plane at 10:40 AM on July 3, 2015 [3 favorites]


Bitcoin and the like are clearly environmental disasters in their own unique way, but I wonder if a successful second-generation cryptoscrip could come out of a collaboration between a cryptographer and a physicist with a basic understanding of thermodynamic laws. At least one hopes we get there before we cook ourselves to death — all that waste heat has to go somewhere, in the meantime.
posted by a lungful of dragon at 10:56 AM on July 3, 2015


Perhaps a smarter place to do cryptoscrip is in outer space. Set up solar-powered satellites that do nothing but run their calculations up where there is limitless energy and a heat sink the size of the universe.
posted by a lungful of dragon at 11:01 AM on July 3, 2015


If so, why not go whole hog and turn processors into pizza ovens?

We shall call them BitPies, and the value will rise after purchase (at least until the cheese goes cold). What is more trusted than pizza?
posted by GenjiandProust at 11:35 AM on July 3, 2015 [2 favorites]


Unfortunately, the near-vacuum of space makes a terrible heatsink.
posted by ckape at 11:37 AM on July 3, 2015 [3 favorites]


Well, whether in space or not, solar is a better way to go. Bitcoin isn't just wasting energy through heat; unless all these processors are running on solar, wind, tides, hamsters on treadmills or whatever, all that power is coming from some place that costs us all in terms of resources. I mean if we're gonna run this thing down to its inevitable conclusion, we may do well to use it up a little more efficiently.
posted by Aya Hirano on the Astral Plane at 12:13 PM on July 3, 2015


Well, add huge radiators.
posted by a lungful of dragon at 12:17 PM on July 3, 2015


A lot of Bitcoin mining is powered by hydro, now. Like these guys in China. Note they probably pay about 3 cents per kilowatt-hour; they're not driven by any high-minded commitment to low emissions.
posted by topynate at 12:53 PM on July 3, 2015


Looking at his posting history, you're way off the mark.

Well I apologize, then. I dunno who I was thinking of.
posted by Pope Guilty at 1:37 PM on July 3, 2015


I get the feeling the huge Chinese bitcoin operations are money laundering schemes. The Chinese government keeps making it harder to run a bitcoin exchange, but hasn't outlawed mining bitcoin. Therefore, mining bitcoin, even at a loss, that you could then send to another country to cash out in the currency of your choice, is an attractive idea. And even if mining were outlawed, I imagine they could disguise it as another type of energy intensive computing and feed the bitcoin network traffic through a VPN.
posted by mccarty.tim at 3:19 PM on July 3, 2015 [1 favorite]


By the law of conservation of energy, you can make your house toasty warm by refunding all your Bitcoin purchases.
posted by storybored at 6:36 PM on July 3, 2015


I'm glad people are talking about this; some time back when I was experimenting with bitcoin I did some back-of-the-envelope calculations along these lines and my jaw just dropped. That is a ton of energy to burn per verified block. And yeah, that means that's the real-in-the-sense-of-systemic cost per transaction, too.

The situation might become different as the system transitions from paying miners out of block rewards to paying them out of transaction fees. But there would still be a race among miners to be the one to get those block rewards, whatever they were. So the system would still have to have some difficulty rating to slow down however many people were in the race.

But talking about that future ignores the current situation, since block rewards won't just end, they taper logarithmically. In (probably) 2017 the block reward gets halved to 12.5 BTC. meanwhile the best numbers I've found say the recent average fees per block have been, on the high site, 0.3 BTC. Assuming that stays constant (a big assumption) it'll take 23 years before block rewards drop that low. With all the mining rigs burning Joules all those years.

Now they might slow down… in 2017 the block reward drops, but users don't want to raise their fees, so a bunch of rigs become unprofitable and shut down. So the difficulty gets lowered and the mining efficiency increases. Or users are willing to raise their fees to compensate, and no rigs get shut down, and the difficulty stays up as does the energy waste.

Who knows? But there'll still be some energy burn just to keep the miners slowed artificially as long as there's some block reward in there, and that means years and years of carbon emissions. And, I worry about the hidden army of unused miners, waiting for a blip of energy prices to give them a chance to switch on and dominate mining until the difficulty changes or their competitors get the same idea and tragedy-of-the-commons that advantage to nothing. I think it was someone here who wrote something good about certain industries where the market forces inevitably create extremely tight margins and perpetual instability.
posted by traveler_ at 10:58 PM on July 3, 2015


Oh, after all that I forgot to mention: there's an alternative to the energy-intensive proof-of-work method used by bitcoin to slow down miners, called proof-of-stake. I don't understand it much but it uses temporarily not spending coins (staking them on mining) to slow miners down. Apparently it has its own weaknesses but wasting energy isn't one.
posted by traveler_ at 11:07 PM on July 3, 2015


Its the rich get richer. Because if you have a stack of the currency then the algorithm thinks you have a bigger interest in not messing up the system.
posted by Iax at 1:20 AM on July 4, 2015


At minimum, stopping counterfeiting and imprisoning counterfeiters should be counted...

I'm thinking the closest thing to counterfeitting in BTC is double-spending. And Bitcoin doesn't actually protect anybody against that, and certainly doesn't imprison anybody for it.
posted by Orb2069 at 3:16 AM on July 4, 2015


Blockchain.info is as crap as ever, but I see what you're getting at. Those are double-spend attempts (or would be if I could see them), not successful double spends, where I define a "successful attempt" as one that could be used to defraud a reasonably savvy counterparty. To my knowledge this has happened precisely once in Bitcoin's history, and was not actually used to defraud anyone.

The Bitcoin protocol's double spend protection mechanism is proof of work. The current longest chain of blocks is 363,778 blocks long. If you can find two transactions, spending the same bitcoins, in two chains of at least that length, such that the transactions are at a depth of, say, six blocks or more, then by all means produce them and prove me wrong.

The above paragraph translates into less jargony speech as "wait to see six confirmations of a transaction before you trust it". For very, very large transactions (thousands of bitcoins), perhaps wait for a dozen.
posted by topynate at 3:52 AM on July 4, 2015


I'm thinking the closest thing to counterfeitting in BTC is double-spending.

It would also include quantum computers and the ability to bypass the whole game. Bitcoin rests on specific assumptions about cryptography and security which date from a specific time period not grounded in future reality.
posted by Brian B. at 10:10 AM on July 4, 2015 [1 favorite]


Correct, we would need to switch to something like Lamport signatures instead of the current ECDSA. Hashing would still be OK – we might have to switch to longer hashes, but quadratic speed-up isn't that big a deal. It would be pretty strange if there wasn't plenty of advanced warning, though – the development of a large quantum computer would be a major event that would precede general availability by months at least.
posted by topynate at 11:40 AM on July 4, 2015 [1 favorite]


It would be pretty strange if there wasn't plenty of advanced warning, though – the development of a large quantum computer would be a major event that would precede general availability by months at least.

I think we would know when a public-funded quantum computer might be available, but not the secret ones. I imagine a government adversary using one in a secret lab to end bitcoin in a matter of days before anyone knew what hit them. That would also be their proof of concept to their funding source.
posted by Brian B. at 12:04 PM on July 4, 2015 [3 favorites]


I guess it would be relevant to note here that bitcoin.org has just posted a notice that some miners are generating invalid blocks because well, hell I don't know, BITCOIN!
posted by eriko at 8:38 PM on July 4, 2015


Has anyone said buttcoin yet?

I did some back-of-the-envelope calculations and buttcoins are an ecologically sustainable, organic and free range alternative. So, no.
posted by a lungful of dragon at 7:38 AM on July 5, 2015


I guess it would be relevant to note here that bitcoin.org has just posted a notice that some miners are generating invalid blocks because well, hell I don't know, BITCOIN!

I saw that yesterday. A considerable proportion of miners were not validating the blocks they mined on top of. So when they saw an invalid block - with the correct proof of work, but invalid - they... mined on top of it. Hopefully the $25K revenue they lost will smarten them up a bit. (They were skipping the check because doing so saves them a small amount of time and so decreases their 'orphan' rate. Estimates are that they were earning about 1% more revenue by doing so.)

Re the quantum computer surprise attack, I've thought about it, and while it would be pretty harsh, it might actually be survivable. Any address that's been reused would be vulnerable, but others would be protected both by ECDSA (quantum crackable) and a hash (not so crackable). An attacker would have to break the ECDSA at the moment of seeing a spend transaction from some address, then attempt a double-spend attack. If the attacker can do that reliably, it means it already has majority hash power, and could break Bitcoin anyway. Naturally the risk of outright cryptographically enabled theft of bitcoins goes from negligible to worrying, so the network becomes unusable until technical measures are taken (some sort of zero-knowledge proof system would probably do it, details on request), but it wouldn't forever destroy all confidence in the currency.

To return to the original question of Bitcoin's sustainability: I looked into Visa, and reached the conclusion that they are emitting at minimum about 45 grams of CO2 for every dollar revenue, on a profit margin of a staggering 45.5%. For Bitcoin, assuming a 30% hydro component to the electricity, and using miner income data from blockchain.info's data set, I estimate emissions of about 3900 grams CO2 per dollar revenue. So dollar for dollar, Bitcoin mining is about 87 times more carbon intensive than Visa Inc. That is not pleasant to contemplate, but certainly isn't >5000x. The other factor is the profit margin. This is not public information, but is widely believed to be about 10% on average.

Why is the smallish profit margin of Bitcoin mining important? Well, it means that carbon taxes bite into Bitcoin miners far more than they do into Visa, so they would respond to carbon taxes and green power credits with more alacrity. The environmental externalities of Bitcoin could mostly go away with cheap enough green energy, even if the total amount of electricity used continues to rise.
posted by topynate at 10:58 AM on July 5, 2015 [1 favorite]


VICE should be embarrassed to have published this.

The day VICE encounters the concept of embarrassment its lead story will be "Hell Freezes Over" or maybe "Is Hell Becoming The Ultimate In Cool ?"
posted by devious truculent and unreliable at 1:50 PM on July 5, 2015 [2 favorites]


" They were talking about Mt. Gox. Silk Road wasn't a BitCoin exchange, it was a market that used BitCoins.

And it wasn't brought down by that DEA agent, that DEA ran a unit that was an ancillary part of the investigation that ultimately netted Ulbricht, but the evidence used in the trial was from another unit in New York.
"

I'm talking about DEA agent Mark Force IV and Secret Service agent Shaun Bridges. They were working together.

http://fusion.net/story/112680/silk-road-agent-carl-mark-force-iv/

says:

" Force implied to Mt Gox’s founder that he had helped the U.S. Department of Homeland Security take down the popular Bitcoin exchange."

The warrant for the seizure of MtGox's assets was issued to.... Shaun Bridges:

http://www.scribd.com/doc/162503556/Mt-Gox-Wells-Fargo-Seizure-Warrants

(from the fusion.net link) " After the Mt. Gox seizure, Force e-mailed Karpeles saying, “Told you [you] should have partnered with me!” with a smiley face as a subject line. "
posted by I-baLL at 9:44 PM on July 5, 2015 [1 favorite]




Just fyi, I'm now working for Taler, an RSA blind-signing based electronic transaction system with many nice properties : Actually anonymous, while Bitcoin is only pseudononymous. Bad for tax evasion. Good for distressed economies who need to deploy a new currency quickly, ala Greece. And easily scalable and environmentally friendly, no blockchain, no proof-of-work, etc.
posted by jeffburdges at 2:39 AM on July 29, 2015 [2 favorites]


« Older The Man Who Saw America   |   Reddit going dark. Newer »


This thread has been archived and is closed to new comments