This is your first mistake: the traditional banking system does not have a security model predicated on the ability to waste power. Bitcoin does, and it’s dynamic so there’s no way to waste less power which isn’t explicitly ceding control.
The second error is treat the two as comparable without recognizing that one of them is used daily by millions of people making billions of transactions, and the other has almost no real world adoption. That matters in two ways because it’s not just that bitcoin uses more power to do so much less but also that the real financial system has higher power draw for work in addition to processing transactions. For example, Visa can do on the order of tens of thousands of transactions per second versus Bitcoin’s 7 but if you are looking at how much power they use, the figures will include running a ton of anti-fraud and other support systems which Bitcoin is missing.
Again, if you think all cryptocurrency has 0 value, then all usage is obviously bad. But the reality is that it has some value, and much of that value is in the form of scams (most types of coins, by number, are scams), and much of the value is just speculative investment, but it is a real thing that people spend real money on. Focusing on miners makes it seem like all the value of crypto comes directly from energy - but nobody would spend energy mining it if others wouldn't buy it.
Yea this is a constant issue in almost all journalism. Numbers are provided mostly without context. E g. "$100k of drugs seized in bust". But was that a single day's re-up? Does it meaningfully impact the local drug supply at all?
---
In this case we can dig into numbers somewhat. Bitcoin.com claims the traditional banking sector uses about 2x as much energy. But cryptocurrency handles a tiny, tiny fraction of worldwide banking stuff, so if it was operating at the scale of e.g. 20,000 qps of credit card transactions, for example, it might have substantially higher energy use compared to now.
Crypto still needs POS systems and accounts and large staffs in large offices and all the trappings of finance outside of mining.
That's not intrinsic to cryptocurrency, but it is part of the reality we're living in, which is what I want people to be looking at.
Bank of America alone has 69m customers and real people have real needs - things like going to a branch occasionally, calling customer service, taking out cash, etc. Not only is BoA itself larger in terms of users than all of crypto (I've done the analysis, checked block explorers, etc) it's actually usable by real people in the real world for real activity that contributes to the real economy and actually provides value and utility other than pancake swapping your doodle coins for poodle coins.
The second error is implicitly assuming that the number of people Bitcoin serves is correlated with it’s power usage. You could serve the same number of people as the banking sector does now without increased power consumption when you bring layer 2 or 3 solutions into the picture.
Bitcoin is like that times 1,000,000 when it comes to energy usage.
The energy and resources expended mining for diamonds is wasteful also.
For the second point, are you talking about Layer 2 solutions that track cryptocurrency in a centralized database, like an exchange, where most transactions happen? Basically just traditional banking but with crypto as the unit of account. Or are you talking about decentralized layer 2 solutions like Lightning Network? I don't think Lightning Network can scale to match what the banking sector does, even if you ignore all the services banks provide other than facilitating transactions. For example you could not pay the US workforce with Lightning Network because it would take several months worth of of blocks just to open a channel to each person, and quickly those channels would run out of inbound capacity and you'd need to open more on top of that, so the Layer 1 capacity still limits the ability to use LN at that scale.
Already now you can get a VISA/Mastercard and use that to spend Bitcoin. But of course, every layer on top of Bitcoin presents its own set of trade-offs in terms of trust and security.
What constitutes waste is completely subjective.
Except, in good old libertarian fashion, they tend not to pay for it. First, the vast majority of the around $100 it costs to process a BTC transaction comes not from explicit fees, but from (invisible) money supply increase (the mining reward). That still neglects the environmental externalities, which are probably in the $30+ region.
This is a logic error: securing a financial system is not waste but paying more than you need to is. If my bank secured my money by paying an army of dudes with guns to sit around watching cash and did transfers by putting a check on the corporate jet, I’d say that was wasteful, too.
Similarly, the argument that the current system could be matched by L2 systems is both speculative and conceding defeat: even if that worked as well as the sales pitch claims it’d be using more power, and we know that there’s no plausible scenario where usage goes up while power consumption goes down.
The argument that L2 systems is the answer is also directly undercutting your earlier marketing pitch. If the justification for the inefficiency is that it’s needed for security, telling people that they should switch to your I Can’t Believe It’s Not A Bank is either admitting that the security benefits are either not real or necessary, or that they can be provided more cost effectively.
The only people who are committed to using Bitcoin are the people who’ve already bought in: everyone else is going to look for advantages relative to what they’re already using. It’s not just enough to handwave about how the system might at some point be less distant from parity, you need a serious plan for being better at something before you’ll see any significant adoption. Parity might seem like a far off goal, and it is, but it’s not enough to get most people to switch.
I’m saying also that L2 (and L3, … Ln) is a way to scale the number of users without increasing power consumption. Every time you add a layer, there are other trade-offs for the benefits gained. But at the base layer you still have the benefits of not having a central authority censoring and controlling exchange of an economic good.
No, it’s simply recognizing that competing systems match or exceed on security without the same cost. You’ve effectively acknowledged this by saying that the system will be more usable and affordable by using something better designed which eventually clears in Bitcoin to reduce the number of expensive transactions.
You leave out other characteristics that make Bitcoin an attractive economic good, when you think that matching or exceeding the security at a lower cost is a valid argument against Bitcoin. If it was all about power consumption, Bitcoin wouldn’t keep increasing in demand and price.
Look, you can theorize all you want about the inefficency of Bitcoin, but the market has spoken and continues to speak.
Also, when talking about extra service payment networks like Visa offering, like anti fraud, gotta remember that at least part of problems this service solving is caused by the design of the network itself. And while it is not obviously energy hungry service, it is part of what everyone is paying 2-3% in credit card processing fees.
I would love to see a solid comparison of traditional banking total cost vs Bitcoin (if it would replace traditional fi), but it is certainly not as simple as the above, if even possible...
The design of credit cards does suck for fraud. I go to a website to buy a product, I put in my card number, exp date, CVV, and it turns out that site was actually compromised and some attacker got my card info. Now I have to cancel it and get a new card. It would be way better if what I entered on that vendor's site was a one-time authorization code, not the full information with which the attacker can spend from my card at other places.
Crypto is better in this regard since I never transmit my secret information to the vendor! But crypto is also worse, in that, if I do send a transaction to the wrong place (for example, because the vendor's website was compromised and told me the attacker's address to send to), I'm screwed. And if my electronic device is compromised and my private key stolen I'm really screwed.
Ultimately even though crypto has a technically better design, since it also has no support department and no dispute resolution I'm at higher risk. I've never lost money from a stolen credit card. You're right that I pay for that in the 2-3% processing fees... but it's not like Bitcoin is feeless, either.
Oh, no argument there but that’s not the only part of fraud I’m thinking about. I’ve had a card stolen but it cost me exactly zero because their anti fraud system was able to detect that the person with the credential wasn’t buying the kinds of things I buy at places I normally shop and blocked it. That’s a LOT better than the countless stories of people making a mistake in where they sent cryptocurrency or losing control of their wallet, and not getting help other than people saying it was their fault.
My point definitely isn’t that the current system is perfect but rather that what seems to happen a lot is that people focus on one technical problem like operating a ledger and forget the surroundings. I’d definitely like to see a detailed comparison, too, but my preferred solution to all of this would be an increasingly steep carbon tax to incentivize everyone to do better.
This doesn't make sense. The power is used to secure the network. Your first mistake here is you must think of Bitcoin as useless and any use of power is a waste. Just to help you, if BTC was useless it would be worth $0 traditional banking dollars, as this is not the case your whole premise is a waste.
It’s a great software engineering case study: the algorithm is a neat solution to a problem that almost nobody has, so while it’s cool as an intellectual exercise it’s failed to see much adoption even as it passes the middle of its second decade because it’s not cost effective for most people. It’s kind of like the idea of a supersonic airliner: a neat engineering challenge but a failure in the marketplace because 99.9% of people don’t need the problem it solves desperately enough to pay so much more.
MIC DROP
The more billions it costs to run bitcoin, the more it costs to 51% attack it. If you can outwaste energy, you can hijack the system.
It would be useful it some state, eg., russia or china, would just cut off networks and forge bitcoin network traffic -- and hence bring the whole thing down. Or, do a 51% attack by controlling 51% of the hashing power.
The whole thing is a joke, one only tolerated by states because it's supports joke levels of economic transactions. Were any % of an economy to run on this, states would just seal the network in their borders and end the whole charade.
It's proportional to the amount of money it "burns", in the form of capital costs (miners) and electricity costs (for those miners). If electricity prices would rise tenfold globally, electricity use from crypto miners would drop by about the same factor.
This only applies to Proof of Work blockchain.
Largest market cap, yes, but also lots of probably lost coin.
One source shows Tether having the most volume, and it's very energy efficient to make coin out of thin air.
Take the peer-to-peer traffic of the cryptocurrency network of any area, prevent it from getting outside that area (ie., just block it talking to IPs outside some range) -- then to that network, you can trivially control the total hashing power.
So split any cryptocurrency network into small segements, then add your machine to that network with a false history, design the network to be small enough, and your machine will out-hash the rest, and so it's history will win. Rinse-and-repeat.
It's trivial for any state to take down a cryptocurrency. There's nothing magic about it; it's an incredibly fragile system whose 'safety' relies on no one owning the network, and if that's the case, no one being able to russle-up huge amounts of electricity.
In both cases, this is false for states. So any state, if it wishes, can really do anything it likes.
The global bitcoin system is well within rearch of a hostile state 51%'ing it, even at the global level -- though the cost would be non-trivial. It would be trivial to do it in its own borders though.
As any state, of course will, if you can ever go into a shop and buy somehting with it. At that point you're imperilling a state's ability to use monetary policy to manage its economy, and that's an existential security threat. So bye bye your monopoly money tokens.
You cannot forge history, even with 100% of the hash power. Firstly, each transaction is cryptographically signed with the keys of the sender address. Secondly, each full node within the segmented network will have the full history.
The only thing you can do is publish different blocks to the segmented network than the blocks the outside network has. You cannot create arbitrary transactions. You can only censor others’ transactions within the segmented network. Since the mining difficulty will not adjust instantaneously your segmented network will fall behind the outside network in block height unless you control more hashing power than everyone else mining Bitcoin on either network combined. So as soon as anyone in your segmented network re-establishes connection with the outside network (and they will, they could receive a physical hard drive with the blockchain on it and rebroadcast to segmented nodes) all your work is for nothing.
You might say “yeah well I’ll publish an entire fake history that is MUCH longer than the outside network with lower difficulty so I can stay ahead”. Well, actually you can’t, because if you wanted vastly more blocks between Bitcoins inception and the present then the difficulty will necessarily be much higher because that’s how difficulty gets set, by how quickly blocks are produced. You would have to change the difficulty adjustment algorithm, creating a fork between your own malicious node(s) and the other nodes in your segmented network. Oh and by the way, you _still_ can’t create arbitrary transactions.
> The global bitcoin system is well within rearch of a hostile state 51%'ing it
I’ll believe it when I see it. They honestly have a better chance outlawing it and imprisoning anyone who’s ever used it.
> I’ll believe it when I see it. They honestly have a better chance outlawing it and imprisoning anyone who’s ever used it.
First of all I don't think any state is currently motivated to do this. However, if I were a state agency trying to attack Bitcoin, I would start by creating my own mining pool, which of course would purport to be privately run. I would be the most efficient mining pool in the business, offering miners a slightly better cut than other mining pools since while most pool operators are trying to extract a low-margin profit, I'm willing to break even or, if necessary, run at a small loss. It would be ideally to gradually create several sock-puppet pools that appear to be in competition with one another, while in fact I control all of them.
Even with competitive payouts, it may take several years to build up my pools reputation and gain a significant share of miners. And when I start having my pools mine blocks that I'm not actually submitting to the chain (to support my double spend), pretty soon miners will notice and switch. But I only need a couple hours to cause chaos, and I may benefit from miners confusedly switching to other pools that are also under my control. If I look at the regions where I have the most miners and time to the attack to occur overnight in those areas, I may succeed. And unlike trying to 51% the network myself by throwing hardware at the problem, I won't be left with worthless SHA256 hashing machines at the end of the attempt, nor will I have to pay for power. And I don't have to outmine the entire network -- I've enlisted half of it to be on my side. The only cost is the minimal pool operating expenses (not mining, just issuing work to miners, checking their work and arranging payouts), spread over how ever many years it takes me to gain dominance.
If you own the machines and own the network you can do anything you want. Anything at all.
As far as assuming that the state hasnt taken control of the miners (unlikely, this is the easiest thing to do), by dropping communication, delaying it, observing it, etc. much can be done following the protocol, including replaying transactions etc. -- the future can be forged.
There are so many assumptions about the realworld, that do not hold up, behind cryto protocols, they're laughable. Assuming that the system will follow the protocol is itself disconnected from reality, quite literally.
The initial paper's realworld assumptions was that mining would be an at-home affair, ie., decentralised; everyone would run their own. And that networks were not own or controlled by centralised actors.
Neither is true. Mininig is incredibly centralised, as is network control. This makes it trivial for a state to pull an off switch.
Even talking about sophisticated denials of service, transaction replays, forging future transactions... all this takes place in a silly imagined scenario in which the state wants to hide what it's doing. If it didnt care, bang goes the whole thing.
Countries have this neat tool called violence that tends to override all security whenever humans are a linchpin
That doesn’t work with distributed systems. No single country is able to shut the whole Internet down, for example.
(b) "seal the network in their borders" is much easier than that, lots of networks basically act like that already due to either voluntary compliance with local laws, or due to direct government interference with the networks
iii. the miners aren't really all that distributed, they group together for the same reason everyone else groups together instead of being free-range anarchists
[δ] all you need to do to shut down a currency within a country is arrest people using it, which is very easy and has a long history
Countries do this all the time , just yesterday we were talking about internet shutdown during elections and exams , then there are restricted countries like say North Korea.
Invalid blocks are rejected trivially by any node as soon as it sees them. You can't "force" an invalid block to be accepted by throwing hashrate at it and miners don't need to do anything to make an invalid block be recognized as such.
The issue is on malicious actors trying to rewrite/replace recent blocks (to double spend, usually) which was the core innovation, but that also includes trying to censor txs or generally DoS the network by refusing to allow txs to be accepted.
Oops, nope sorry. Most don't use PoW..
Out hashing does not give you any extra powers. Maybe you will be able to rollback a few minutes of transactions, that is all.
You mine quietly, outpacing the regular network, creating your own longer chain. Perhaps you do this for an hour, several hours, or a day. The only transactions on this chain are those in which you send your own coins to yourself -- maybe plus some random transactions from the mempool you throw in to help cover your own tracks. Meanwhile you spend those same coins (UTXOs) on the "honest" chain.
Once you are satisfied, you publish your longer chain. By the rules of the network the longest chain is the true one. Honest miners immediately begin building their blocks on top of your dishonest version of events. Since you are now working together with them, blocks are produced quickly. The honest chain dies.
Chaos would ensue.
Anybody who received coins on the dead chain can try to republish the transaction to the mempool and hope it gets mined again, so they get their money again. But if the sender is fast, they might realize they have a double-spend opportunity here, and submit a higher-fee transaction in which they sent those coins to themselves. It's a race. Of course, for your own UTXOs that were already spent in the dishonest chain, the race has already been won. Those double spends are successful.
As for "a few minutes", it depends on which one we're talking about. BTC, yes -- but disrupting the global BTC network requires only a few minutes of time. Just dump everything into unusable addresses, and watch the reaction.
No, you cannot do that. Transactions would still need to be signed by the corresponding owners. All you could do is reorder (and thus invalidate) some transactions, drop some transactions, or add some that weren't originally included.
> Just dump everything into unusable addresses...
Nope; that's not possible. You wouldn't have the signatures to do that even with 51% hashing power.
In the case of a mere 51% at scale, hijinks are still quite possible from replays, reorderings, etc.
I was more preoccupied by the case where the state's acting in its own borders with control over the network, major miners, most machines on the network -- at this point basically no gaurentees remain
Such a change would be treated as a "monopoly money" fork by those wanting to transact, and ignored. "No I don't want your fake money; send me real BTC or GTFO". This is why miners cannot arbitrarily change the protocol in their favour even today.
Bitcoin used 844670Wh per transaction [2].
[1] https://usa.visa.com/dam/VCOM/download/corporate-responsibil...
In theory it’s possible for a Bitcoin-like blockchain to record all of Visa’s transactions in a single giga-block with only a single mining hash attempt.
Transaction throughput is not meaningful here.
"Well, it's supposed to suck like that" doesn't change that Bitcoin is consuming tremendous amounts of energy to serve very, very few customers.
1. Bitcoin is a final settlement system, akin to moving gold bars from one vault to another. It achieves this settlement in minutes and requires no trusted intermediaries.
2. Bitcoin's energy consumption is a function of price, not of transaction throughput, so the two are orthogonal.
You could fly gold bars around in small planes, and it would still be more efficient than Bitcoin.
The banking sector exists for a reason. There are certainly parts of it that are superfluous and excessive, but someone somewhere is willingly paying for that excess, and that won't change under a bitcoin regime. And I personally think that most of the excess is due to excessive speculative activity in our financial system, and if anything, bitcoin seems like it will make that part worse.
Seems I was off by about 5 orders of magnitude, in Blockchains favour!
BTC is about 9 orders of magnitude less efficient than the effective work it does.
Look, you can run a full Bitcoin node on a Macbook that uses 20W. It can easily verify 5 transactions per second, put them into a block, and compute the hash of it, and do some network communications.
The actual network uses 20GW, a billion times more energy, because they have thousands of nodes replicate the same work and then add PoW as a cherry on top.
Bitcoins are just speculative assets . All crypto is, nobody is actually doing real economic transactions on them, I.e buying and selling services/goods for bitcoin. It is not practical to do so with transaction/gas fees .
If you think of it as an Asset transactions then it makes sense, those are always expensive.
A real estate deal for a house will have buyer agent, seller agent government taxes etc easily 5-10% , if you sell an artwork a gallery will probably charge upwards of 10-20% etc.
I don't like it, I don't support this industry, it is still big waste of energy, however if you see it from a lens of asset instead of currency then it kind of make sense that transactions are expensive.
Money has four functions traditionally, Bitcoin does not work for three of them: medium of exchange(too expensive), unit of account(too volatile), standard of deferred payment (too volatile), it kind of works as "store of value", i.e asset despite the volatility as the deflationary pressure due to its by design purchasing power increases.
Smarter economists can explain it better than me, my point is, don't think of it as a currency, it is really not, just an asset class ( still wasteful).
Sorry, but this is a ridiculous statement.
Electricity is one of the most fungible things in existence, and renewables don't yet make up 100% of production, therefore any additional consumption _for the most part_ might as well be 100% generated from fossil fuels.
Even if a miner has a contract with a solar farm for all of their power, that's still electricity that someone else could be using instead of electricity from fossil fuels.
(Not arguing for or against this application, and it may be that this is too rare a phenomenon to matter; just quibbling over the fungibility.)
Not if the infrastructure wouldn't have been built without the miners.
Yet you rolled out the standard crypto talking points flawlessly.
> Energy consumption is frankly not a phenomenon I give two shits about, as an environmentalist
Again, I suspect you aren't an environmentalist either. Please just be straightforward with your motivations here.
POW is the wasteful one.
Actually, there are lots of other ones, but Proof of Work is the only one that is designed to add security by requiring a lot of hashrate (which translates to a bunch of energy) to submit blocks ("mining" in proof of work)
Proof of work was a brilliant innovation at the time, but as it turns out, smart people iterate on technology over time to improve it. The fact that bitcoin didn't sink like a stone after being improved upon (just from a resource-usage perspective) by most other blockchains really highlights the greed and single-mindedness of most crypto speculators.
Nicely put. PoS is still around a million times more wasteful than the effective work it does, though.
Transactions in 2023:
Bitcoin: 174310000
Ethereum: 583530000
Visa: 212600000000
Bitcoin: 167.8 TWh/yr
Ethereum: 0.0067479963 TWh/yr
Visa: 197.57 TWh/yr (2021?)
Bitcoin: 962.652745 kWh
Ethereum: 0.011564 kWh
Visa: 0.929304 kWh
Bitcoin on the other hand is 3 orders of magnitude worse than visa on a per-transaction basis. I suspect both Bitcoin and Ethereum also transact more value in USD per transaction (if you're willing to consider funny crypto money at spot prices) on average, than Visa, by virtue of them having a much higher transaction fee which is also borne by the address making the transaction.
If you don't consider Ethereum transactions to be "effective work" than this isn't a relevant metric of course, but I'm curious how you're arriving at your figure.
Hi, sure. The nodes in the EVM do not collaborate to split the work, rather, they all perform (basically) the same work. So, you could instead run Ethereum with one node that connects to the net and receives transactions, processes them (around 20 per second - it's not exactly a lot of processing), puts them into blocks, and sends out the complete blocks. It would perform exactly the same as the current network, and would consume, say, 4 W (a Raspberry Pi, say).
The one difference is that it would have to be trusted, which is the big no-no. But that's the effective work being done.
But the validators burn around 1 MW (that lines up with your figure approximately) and are paid $5m a day to do that work. And 1 MW / 4 W is a factor of 250k, which I rounded up to 1m considering all the other activity around Ethereum, though feel free to round it down to 1e5.
1. The banking industry is thousands of times more important that crypto for everyday Americans. It's what literally makes our entire economy run. Crypto is used for speculation. What you should be comparing bitcoin against is the energy it takes to list a few ETFs on an exchange, which is essentially the cost of a few hundred servers, aka nothing. Bitcoin is ungodly inefficient and will only become more inefficient.
2. "It could be less, much less, or much less," This is a filler argument. Study says "around" a number, so you're arguing it could be a really broad definition of "around"? Not a serious take.
3. The absolute number matters because it's showing the economic activity is being concentrated in the pockets of just 137 entities. This is NOT a broad-base economic phenomenon that helps employ people in communities around the country, like say running gas stations (another polluting industry). If you're going to use 2% of the nation's electricity, you'd want the benefits spread across at least 2% of the population. Not 0.00002%.
4. Net energy consumption pulls from our current energy grid. Within a grid, if you use more clean electricity in one area, it'll pull more supply from dirtier sources elsewhere. Saying that bitcoin mining helps "green" the grid is a myth because you have to mine it 24/7/365, which means intermittent sources like wind and solar are at a big disadvantage. Battery storage tech is not there yet, thus gas plants and dirtier sources have to kick-in to keep up with 24/7 consumption.
I also somewhat disagree with (3) because it's ambiguous how many people constitute these mining facilities referenced in the article. You write it as though it means 137 individuals.
But overall I only dislike crypto because it has become a space for charlatans and people who want to get rich quick. Any original principles behind creating a new currency outside of fiat and actually using that for transactions have been forgotten since then and crowded out by people who pretend those principles still mean something but really just want XCOIN to moon.
> "It could be less, much less, or much less," This is a filler argument.
Actually my argument is, "This article claims it could be much less, and that we aren't even sure of the range given, but the title definitively claims the maximum value." I'm saying this article is bad and is FUD, that's why I said the thing about specifically criticizing the article. (Yeah I don't think I was that unclear actually.)
I'm actually ambivalent on the rest of your arguments. They would have made good talking points in TFA, which made none of them.
Also just get some Bitcoin :)
https://www.aceee.org/files/proceedings/2003/data/papers/SS0...
Citation needed .
It is rarely if ever too expensive to transmit electricity anywhere . We have offshore wind solar farms in remote places , there is even cross continent lines between Africa and Spain .
I am deeply skeptical it is too expensive for upstate NY to transmit power when there is probably 50 million people living in metro areas within say a 750 mile radius .
Every other dam is remote, every coal plant is remote they are all perfectly able to transmit to urban and industrial centers
The plant was shut down and nobody made an effort to direct the power elsewhere.
Transmission is insanely expensive at scale. I've seen their power bills and I talked to their head electrician who worked at the plant when it was operational. That's the best citation I can give you, sorry.
Have you read their website? "With 435MW of transformer capacity at its Massena complex" — that is the amount they can draw from the grid, not the generating capacity of any nearby power plant.
The plant is here: https://openinframap.org/#15.08/44.98237/-74.74862 (checked with several sources, e.g. https://www.gem.wiki/Massena_Complex).
The map shows what I'm talking about very clearly... all of the power is either used for the local Alcoa plants (there are several) or gets exported elsewhere.
There is little usage for that power locally (it is in the middle of nowhere, frozen in the winters and otherwise wasteland surrounded by superfund sites thanks to the smelting).
It also isn't like you can just plug in 115kv line into a 230kv line and call it a day. It would require new construction to send that power somewhere else, which is expensive and hard to justify over someone just paying for the power to stay right there.
I never said anything about 'generating capacity'. The dam, where I also went for a tour, is ~2k mw total, part goes to Canada and the rest to the US. The Massena facility has an allocation of that power.
I won't discuss this further.
It is difficult to get a man to understand something, when his salary depends on his not understanding it.
I didn't say it can't be transmitted elsewhere.
I said that it was expensive to do so and I've seen with my own two eyes the proof of that expense. The proof of that expense is that it hasn't been done yet, despite years of opportunities and the fact that it can be done (which I agree with you on).
For the record, my salary does not depend on this at all.
As I said, I no longer mine anything and haven't since ETH went to PoS. Which, by the way, consumes very little power now.
>It is difficult to get a man to understand something, when his salary depends on his not understanding it.
Little sneaky to put that last comment there, useless completely and spiteful? Just because your hypothesis turned out to be wrong doesn't mean you cannot admit it
Cryptocurrency miners rank worse than oil and coal executives in my book.
(I've seen nothing to suggest my hypothesis is wrong.)
That's your opinion.
> and there are other energy intensive industries that would have brought some prosperity to the region.
If that were true, it would have already happened.
https://www.statista.com/statistics/881541/bitcoin-energy-co...
I ran the numbers a few days ago (see https://news.ycombinator.com/item?id=39234766). Bitcoin miners plausibly use more energy than the entire traditional banking sector. Not just on a per-transaction basis (which is blindingly obvious if you think about it), but on a gross energy consumption basis.
So less than 1% of US pop. in banking sector. And I guess banking use less power than e.g. mining, manufacturing per worker ...
Bitcoin probably uses more energy then the banking industry and their workers living their lives.
Digieconomist [1] states that a single Bitcoin transaction is "equivalent to the carbon footprint of 1,044,177 VISA transactions or 78,521 hours of watching Youtube."
Statista [2] estimates a similar energy-cost ratio of about 470,000 Visa transactions to 1 Bitcoin transaction.
https://digiconomist.net/bitcoin-energy-consumption
https://www.statista.com/statistics/881541/bitcoin-energy-co...
If an org has a big polluting coal plant powering a bunch of crypto miners and with a wave of a magic wand I give them an equivalent amount of clean renewable energy, they wouldn't turn off the coal plant, rather they would just double the number of hashes they are pumping out.
Proof of work is designed to be as wasteful as possible, it will never ever be efficient or environmentally friendly.
>Energy consumption is frankly not a phenomenon I give two shits about, as an environmentalist.
Despite a few people specifically calling this sentence out as 'ridiculous' or whatever, I think this is a completely valid point. If 100% of the energy powering something comes from renewables, who gives a shit?
We simply cannot currently generate enough renewable energy to cover all our energy needs, so if you are wasting power that could've been used for something more important, it's a waste regardless of the source of your energy.
The more energy society demands, the more likely it becomes that some of it is generated by coal. And does it really matter which electrons are used for what at the end of the day?
I agree that most arguments in this area resembles tribal war more than rational debate, much like any discussion involving either trump or musk, where rationality is also thrown out the window in favour of signalling tribal affiliation
It's rational to conclude the costs are not worth those supposed benefits.
Renewable energy production has externalities, same as everything else. The power production equipment has to be built, which needs materials. It has to be transported, which needs energy. It needs to be maintained and so on. Just saying "well, we can always add a few more renewable power plants, who cares?" would only be valid if we could produce them out of thin air.
1. Even if the energy consumption of the traditional banking industry is comparable or higher, it supports the daily activities of literally 5 or 6 orders of magnitude more entities and people than the 137 outfits mentioned in TFA.
4. The environmental angle is only one aspect - the other is simply that energy costs have never been higher, and crypto mining adds another energy consumer competing against householders, pushing costs up yet further. (Yes, I know, crypto mining can be set up to happen only when there's spare capacity, and I'm sure the miners can be relied upon to act responsibly in that regard...)
We can get a reasonable upper bound: the block reward is 6.25 BTC + fees (the blocks I saw tended to have around 0.5 BTC in fees), so let's be generous and say 7 BTC per block. 144 blocks per day are 52560 blocks or 368k BTC per year. Currently, they're worth 47k each, so $17.3B/year.
The 91 TWh figure (from https://www.eia.gov/todayinenergy/detail.php?id=61364) seems derived directly from a 240 TWh global power consumption multiplied by the assumed % of mining happening in the US, so since we're using global revenue, it only makes sense to use the global number, 240 TWh = 240e9 kWh.
17.3e9 USD / 240e9 kWh => 7.2 ct/kWh
That's the absolute maximum that electricity could cost in order for mining to be profitable.
That's making some very generous assumptions - I used the current price (annual high!) instead of the much lower average over the past months or year, and it completely ignores the very significant cost of the mining hardware. But given that 7.2 ct/kWh is roughly what industrial electricity costs in the US, and miners might be able to get power cheaper (since they can be more flexible than e.g. an aluminium plant that would suffer expensive damage if power is cut unexpectedly), it's not an entirely unreasonable number for an upper bound.
Getting accurate numbers through mandatory reporting isn't the worst idea, if only to prevent such bullshit headlines.
Basically, take the overall hashrate of the entire network which is determined from the block difficulty (this fluctuates and was likely higher when bitcoin was more valuable, though not necessarily), and assume all blocks are mined by operators running state of the art mining hardware (realistically 90-98% of blocks will be mined by people running hardware that's 80-100% of the maximum possible efficiency, and any miners running worse hardware will be consuming more energy, but we can pretend they're all on the best hardware)
Using that, you can get the overall hashrate of the network. Once you have that, you can take the percentage of blocks mined by these 137 miners in a year. Again, be generous and assume they're all running state of the art hardware. If 10% of all blocks are mined by these 137 mining ops, guess what, they're using 10% of the total energy being inhaled by the network.
All commercial buildings in the United States used about 6.8 trillion BTUs, or 2 trillion kWh (2 petawatt-hours), of power in 2018 [1]. That's just over half of the 3.9 petawatt-hours used by the country as a whole [2]. Of that, 16% was consumed by offices [3], so we're down to 8%. That's all office buildings, so what percentage of office building space is devoted to banking? Does banking use more than 1/3 of total office space in the United States? If so, you have a point. That seems highly unlikely.
(Datacenters used less than 0.1 petawatt-hours in 2020 [4], or about 2% of total energy consumption. Again, I doubt banks were using the lion's share of that, so it doesn't move the needle much.)
[1] https://www.eia.gov/energyexplained/use-of-energy/commercial...
[2] https://www.globaldata.com/data-insights/power-and-utilities...
[3] https://www.eia.gov/consumption/commercial/pba/office.php#:~....
[4] https://eta.lbl.gov/publications/united-states-data-center-e....
Great, thank you for doing so. I wish TFA had given even half this effort.
FWIW, that's a projection from 2016, but it should still be in the right ballpark.
Comparing Bitcoin power usage to all other data center power usage is also enlightening. Those data centers are hosting the majority of the world's online services as well as providing services used by physical businesses; they're doing a lot more useful work, for more people, than Bitcoin is.
On the other hand if crypto being on renewables is displacing things that otherwise would have been - is that a win?
1) “No comparison to the banking industry made” - sure, but the banking industry serves the entire world’s financial system, handling billions of transactions a second, while bitcoins transactions take anywhere from 30 minutes to two hours and process, like, 7 transactions a second. There is no reality we live in now where crypto transacts faster, more cheaply, than our current financial infrastructure.
2. Even if the it’s something g much lower, like .5%, point 1 above doesn’t absolve crypto from its share of energy expense for comparatively little work.
3. Again, splitting hairs and taking cover behind technicalities and specific figures doesn’t change the wider picture that crypto is insanely inefficient at what it’s trying to do compared to the financial systems it want to replace. I guess it’s clear I think it IS a waste of energy.
4. Where do you think the lions share of energy generation to feed crypto mining comes from? Even if it was purely renewable energy sources the gross inefficiency of crypto is still notable given its goals. There are better systems to do it.
Production of any energy causes pollution: those wind turbines and solar panels aren't carbon-free, they took a bunch energy/carbon to produce ("embodied carbon").
Increasing energy use for anything causes the need to produce new plants to be built. If you can use less energy to produce the same results and thus reduce demand it may obviate the need for more more supply in the first place.
Re 3. My rule of thumb is that crypto uses about 1% of world electricity production [1]. I'm surprised that the US has such a disproportionate high proportion of mining, indicative of cheap energy & lax regulation?
[1] Note that electricity is roughly a fifth of energy production (the rest being fuel for transportation/heating).
"The results of this survey show that the members of the BMC and participants in the survey are currently utilizing electricity with a 63.1% sustainable power mix."
https://bitcoinminingcouncil.com/
The BMC is mostly composed of US miners.
Now, the above argument is only relevant if we can prove that Blockchain transactions actually are a waste of energy, and you seem to attempt to argue against that as well, by asking for a comparison with the traditional banking system.
I'm not sure that the traditional banking system add a whole uses less energy than Blockchain as a whole, but I AM sure that the traditional banking system handles several orders of magnitude more transactions than all Blockchains does combined, and absent any proof to the contrary, I'm assuming the traditional banking system is not using more than an order of magnitude more energy than Blockchain(since that would mean that traditional banks use around 23% of American energy, which is an absurd amount of energy)
As long as Blockchain uses more energy per transaction than traditional banking, it's a waste of energy, relatively speaking. And while I don't currently have any good sources to prove this point, I see very little reason to believe Blockchain isn't using at least 2 orders of magnitude more energy than traditional banking per transaction.
That’s irrelevant. Banks aren’t going anywhere, end of story. Maybe not in a current form, but they’ll stay, just because of sheer amount of infrastructure built on them.
Sure, but it’s highly unlikely to be as much and, more importantly, the banking industry is actually useful to power the vast, vast majority of real financial activity.
Mining consumes 2.3% of US power and it is used for essentially nothing useful. If it were to end up scaling and becoming a bedrock of US financial infrastructure, imagine how much power we’d need to generate.
2. "I haven't read the study but maybe it's wrong and maybe the real number is way way less"? C'mon. Give me a break. Trusting that an EIA report is generally accurate is not evidence of being "emotional." Handwaving it away without even reading it is.
3. It's interesting and newsworthy that a small number of players are using a tremendous amount of electricity. The article is about bitcoin mining. "Video games" probably use a lot of electricity too, but dispersed across hundreds of millions of users.
4. You're entitled to that opinion, but I assure you that caring about massive energy consumption is not evidence of an anti-bitcoin conspiracy.
That's not what I said. I said the author of the article did read the study and describes it as possibly wrong. In fact the article does actually mention a range given, "0.6% to 2.3%", which isn't orders-of-magnitude huge but nearly a 1:4 ratio. Usually if someone says, "I have this big scary number," and you find out the actual number is 25% of what they told you, you're allowed to be skeptical of the conclusions they draw from their math.
The 137 facilities they are talking about use an 2.3% of US electricity when run at capacity.
Separately they use CEBCI data to come up with a top-down estimate of bitcoin mining at 0.6% to 2.3% of US electricity.
Putting the two together suggests the true number is at the high end of the range.
Despite declaring yourself not one, this (and every other one of your points) comes up _every single time_ crypto proponents are trying to seed FUD in any reporting about the energy usage of crypto. This is especially laughable when you then say
> ~2% of US power is generating over 1/3 of the world's Bitcoin.
It's good to know that it takes 2% of the US power grid to power 2.3 transactions per second.
By your own argument - unless the traditional banking industry tops out at 100% of US power usage for 115 transactions per second - it is definitely not worse than bitcoin.
And it's worth pointing out that it comes up every single time without a shred evidence, just a hunch that the crypto enthusiast ensures us is correct. I'd love to see some actual numbers on the energy consumption of the banking industry vs. crypto.
Or lets be real, let's compare the amount of energy used by the banking sector to provide the same store of value and 3 transactions per second or whatever minor utility BTC provides.
What evidence could I possibly provide for the claim "This article doesn't give any evidence of impact."
It is very clear that crypto transactions, for example, are much less energy-efficient than USD transactions. The article could have easily mentioned that, but it didn't.
> Visa used 189GWh in 2017, worldwide, of which around half was data centres [1]. That handled 111.2 billion transactions, so about 1.7Wh per transaction (half for datacentres). Bitcoin used 844670Wh per transaction.
That's about 500 000x more energy consumed per transaction by BTC than Visa. I guess it goes without saying that it's much less efficient.
For years now I encounter crypto enthusiasts have been suggesting that BTC is an improvement on the energy efficiency of traditional banking, without anything to back up this claim, and ignoring stats like the above over and over again.
You have it backward, it comes up when reporting tries to seed FUD about crypto energy usage. Every point I made is that this article isn't actually saying anything obviously scary, it's just saying some facts in a scary voice.
> By your own argument - unless the traditional banking industry tops out at 100% of US power usage for 115 transactions per second - it is definitely not worse than bitcoin.
That's not my argument. They do different things. Despite its name, most cryptocurrency isn't really used as currency. As you say, it's slow and not very useful for that. I'm just making a comparison because most of the thoughtless FUD discussion assumes that USD is free (in energy terms).
Sure, but 2.7% of USA's electricity is almost certainly more than all of traditional banking. Now, I've just tried googling that, but the results are clearly being distorted by the exact things you're complaining about here, with estimates varying from 15.8 GW to 558 GW worldwide and that higher value being about 20% of global electrical demand which is obviously wrong as it's a bit more than the entire commercial/public sector of the global economy combined.
> which I am sure is not a particularly energy-efficient industry.
Compared to what it could be, perhaps. But it would be surprising if it was as bad as a system which specifically regards using as much computational power as possible to be a security measure.
> The point is the amount of crypto being mined.
But that's designed to be independent of the energy use?
Also, mining isn't the value proposition of a currency, transactions are. One of the arguments that I often hear that's supposed[0] to be in support of bitcoin, is that it prevents inflation. But mining is inflation. Unfortunately, Bitcoin specifically has a transaction limit too small for just interbank balancing, let alone for direct personal use — for that, the system-wide limit would only barely cover Berliners taking cash out of ATMs once per week.
> If Bitcoin mining ends up using a higher proportion of renewable energy than, say, traditional banking, it could be seen as a net good for crypto use to rise.
No. To find out why, let's take it to absurd levels: build a Dyson swarm, cover the sun, use all that to mine crypto. What changes, economically or environmentally?
Nothing. Literally nothing.
The rest of the economy doesn't get greener, there isn't any more cryptocurrency than there would have been, the transaction rate hasn't changed.
[0] it doesn't actually support Bitcoin, but why is out of scope for this comment
>The point is the amount of crypto being mined.
Is it? According to who? You? What's the utility of crypto? I've yet to see any materialization other than casino and funding terrorism.
Yes, because cryptocurrency is the thing we're measuring. The number of miners is meaningless in this conversation. It would be like talking about how many companies manufacture wind turbine blades.
It can't even serve every lunch in Vermont without extra layers running on top of it.
Most of traditional banking is about financing, not about money transfer. The risk estimation and risk control are the tough parts of financing, and I don't see how on-chain computation is a good fit for that risk modeling.
So, "What's the energy cost of the combined Visa and Mastercard networks" is a much better benchmark than "What's the combined energy cost of the traditional banking system?"
One big energy use I optimized was daily risk calculation for exotic books. Say an Australian insurance company sold a bunch of life insurance in Japan. They want to cap their losses on $5 billion USD notional (maybe 2,500 policies, each for $2 million) over the next 30 years, and they want to cap their losses at $1.5 billion AUD. This extremely bespoke, not something the bank can offload to another entity, so the bank is pretty much stuck with this re-insurance on its books for the next 30 years. How much does the bank need to keep on-hand in case this goes sour? To answer that question, every day the bank needs to run Monte Carlo simulations to cover shifts in the Japanese life expectancy curve and shifts in the JPY/AUD exchange rate over the next 30 years.
It turns out that for some of these exotic options contracts, just compute cost for calculating risk exposure over the lifetime of the trade can end up being a noticeable percentage of the profit from the trade.
Sure, you could pull some of that simulation compute on-chain in smart contracts, but what does a DAO do when it has insufficient reserves to cover its outstanding obligations? I guess you need a diverse enough DAO so its risk exposures are sufficiently uncorrelated, but that means that it needs to do vastly more computation to calculate its current exposure every time someone asks the DAO to enter into an obligation. Either all miners need to redundantly run these calculations, or you push them off-chain.
If you push the risk calculations off-chain, then you're back to the energy consumption being very different than the traditional banking system.
(edit: while I still can, I'm going to jump in and repeat that I'm criticizing this as a bad article which isn't saying anything very meaningful on its own. Thanks everyone for doing the author's work for them and gathering more data. I won't be upset if it's proven that crypto is intrinsically bad for everything I love - except insofar as I believe it can't be stopped.)
1. There is no comparison made to the energy consumption of the traditional banking industry, which I am sure is not a particularly energy-efficient industry.
2. The title of the article is itself editorializing the content of the article and is speculative according to their own words. From TFA: "[The EIA] released a study suggesting that cryptocurrency mining represents up to 2.3% of U.S. power demand." It could be less, much less, or much less, although I haven't read the actual report yet (I'm criticizing this article only), because "suggesting" could be doing a lot of heavy lifting
3. "Just 137" isn't meaningful AFAIK. Who cares if it's 100 or 10 or 10,000 miners? The point is the amount of crypto being mined. (The number of miners will always be in flux by design anyway.) And the answer to that, from TFA: "The EIA found that the global share of Bitcoin mining that takes place in the U.S. grew from 3.4% in 2020 to a whopping 37.8% in 2022." So roughly ~2% of US power is generating over 1/3 of the world's Bitcoin. This is only obviously bad if you believe Bitcoin is a total waste of energy, which I don't, but I understand that some do.
4. Energy consumption is frankly not a phenomenon I give two shits about, as an environmentalist. Fossil fuel consumption is. From TFA: "We intend to ... quantify the sources of electricity used to meet cryptocurrency mining demand." Meaning this has not been done, and publishing scary warnings is premature. If Bitcoin mining ends up using a higher proportion of renewable energy than, say, traditional banking, it could be seen as a net good for crypto use to rise.