[askmike]

What article is claiming is based on a misunderstanding of how Bitcoin works, this really is an odd way of thinking about it.

If half of all people stop sending bitcoin around, the amount of electricity used doesn't go down by 50%. So you sending or not sending bitcoin doesn't impact the electricity spend by miners at all.

Miners mine to secure the network, there is not a certain amount of electricity needed per transaction.

[rspeele]

But, the only purpose of mining -- the whole purpose of the blockchain -- is to facilitate transacting.

So if the network can only do X transactions per day, and it costs $Y to run the network, it seems fair to use a "cost per transaction" to describe that inefficiency.

Kind of like how I could represent the total cost of ownership of a car (purchase price, oil changes, tire changes, big repairs, fuel) in terms of $ per mile, even though out of all those costs, only fuel is directly consumed by driving a single mile.

[lowkey]

It is incorrect to assume that the only, or even the most important, purpose of bitcoin is for transactions. Bitcoin users primarily hold Bitcoin as a store of value or speculative savings technology, typically held over long periods of time (years) with an expectation of price appreciation at the expense of short-term volatility.

It is not, as many on HN have correctly pointed out, a viable transactional currency for most use cases due to price volatility, taxation related friction, limited real-world adoption for payments or transaction costs. The only compelling transactional use cases I know of are censorship resistant payments (e.g, Wikileaks) or high value international funds transfers outside G8 countries where wires are slow and risky.

Most Bitcoin is held by savers or speculators over long periods and transactions are infrequent. Therefore the primary purpose of Bitcoin mining is securing the network from bad actors. Bitcoin is a secure vault on the internet. Just because people put money in and take money out of a vault doesn’t mean the purpose of a vault is transactions. It is security against 51% attacks.

Therefore, the appropriate measure is not cost per transaction. It is cost per total value secured.

[rspeele]

I didn't say the only purpose of Bitcoin was to conduct transactions, but that the only purpose of mining (and the blockchain) was to do so.

Specifically, mining exists to ensure that if you send me Bitcoin, I can be quite confident you haven't sent the same Bitcoin to somebody else. All that expenditure is there to guarantee nobody double spends. It facilitates trustworthy transactions.

Now, it's fair to say that in a way, that also protects people who are just sitting on their Bitcoin not transacting it, since their Bitcoin wouldn't be worth anything if they didn't believe that they could transact if they wished to. But the literal, direct purpose of mining is to protect people receiving Bitcoin in trade, not people sitting on it, who are protected from theft by the secrecy of their private keys.

[lowkey]

Mining hash rate is correlated with value in network. It is not correlated with transaction volume. If tx volume dropped by half it would not meaningfully impact the hash rate. If Bitcoin’s price were to drop by half the hash rate would definitely fall.

[rspeele]

For now, that's true, since miners are mostly paid with newly minted BTC. The payout for the job of mining is currently based on BTC's valuation, but the reason for the mining is to secure the transactions. That also aligns with its design.

Why else would Satoshi design Bitcoin's block reward to slowly dwindle towards zero? Apparently in his vision, the eventual steady state for Bitcoin is to be supported entirely by demand for block space (aka, transaction volume), with no correlation to "value in network".

Edit: to put a finer point on my question...

Does the transaction-fee-only mining model work?

If it does work, isn't the current block reward wasteful, funding tens of millions of dollars a day worth of mining, when the fees set by competitive demand for block space are a fraction of that? If fees alone are going to be enough to secure the network adequately, why can't Bitcoin adopt a more aggressive halving cycle and move to fee-only by, say, 2040 instead of 2140?

If it doesn't work, that means there will eventually be a problem with Bitcoin as it slowly moves toward that model. How can Bitcoin change its design to fix this, without compromising the "only 21M coins ever" promise that many of its stakeholders consider a fundamental strength?

[lowkey]

Even in the future when mining rewards shift to tx fees, the purpose of the fees paid to miners will still be to secure the network primarily, not for the txs themselves.

[leereeves]

> It is incorrect to assume that the only, or even the most important, purpose of bitcoin is for transactions.

Transactions may not be the only purpose of Bitcoin, but they are essential to the existence and security of Bitcoin. The block rewards for mining drop by half every few years and will eventually drop to zero. Soon, miners will be paid by transaction fees.

As Satoshi Nakamoto reportedly said: "In a few decades when the reward gets too small, the transaction fee will become the main compensation for nodes. I’m sure that in 20 years there will either be very large transaction volume or no volume."

[lowkey]

Still, the primary purpose is most likely to be security of the network, not transactions which makes the cost to secure the network the better metric.

[danShumway]

> Therefore, the appropriate measure is not cost per transaction. It is cost per total value secured.

Have you run the math on this? A quick DDG search brings up (https://www.investopedia.com/tech/how-much-worlds-money-bitc...), which estimates about 2.9% of the world's money supply is in Bitcoin.

It's not clear whether they estimated how much of that money supply is actually still accessible (ie, how many dead wallets there are), and it's not a given that all of that money could actually be cashed out anyway (see the recent stablecoin fiascos). But lots of asset classes are vulnerable to runs, so let's assume that it's completely accurate, and Bitcoin is using all this power to meaningfully secure 2.9% of the world's money.

In order to secure that 2.9% of the money, Bitcoin generates more e-waste than a mid-sized country and uses roughly the same amount of energy as the entire country of Sweden every single year. And the problem is that even the most generous estimations of the amount of power that current financial markets use make that energy expenditure look really inefficient. Even pro-Bitcoin articles that I find online (ex. https://news.bitcoin.com/banking-system-uses-significantly-m...) are estimating that gold and banks each use in the neighborhood of 2-4x more power than Bitcoin annually. Which is a little bit embarrassing given that Investopedia above suggests that Bitcoin secures less than 10% the amount of money that gold secures. Similarly, it's tough to estimate how much money is held inside the financial sector (and of course, banks do way more than just secure value), but nobody I can find is giving estimates as low as 6-12%, instead I'm seeing some estimates as high as 25%.

I would not really classify gold as an environmentally amazing asset, but when considering gold we're still looking at a store of value that per-year is basically 2-5x more energy efficient per "dollar-secured" than Bitcoin is.

---

And I feel like this should honestly be kind of intuitive to people, if anything people should be surprised that those numbers aren't worse. Bitcoin's design is such that it uses electricity proportional to the amount of profit available from mining. Until the mining rewards drop to zero, as Bitcoin rises in value the energy/hardware expenditure will also rise to match that value. If it doesn't then the value of the coin will eventually get high enough to make 51% attacks profitable.

So take a step back and think about that: a system that keeps its assets secure via a constant, massive expenditure of energy, that has to grow in energy expenditure as the price of the asset increases, and that has to be maintained in perpetuity in order to win an ever-escalating computing arms race against attackers... well, that's not a system that's exactly setting itself up to be an amazingly efficient store of value. It's not surprising that more traditional methods of running and securing databases and coordinating databases/transactions would be more efficient. It's not surprising that even a mostly physical asset would be more efficient to secure.

[lowkey]

If we are comparing energy consumption of monetary systems I think we also need to look at the incumbent system’s costs with open eyes. The US dollar is the global reserve currency. It is no longer backed by gold. Instead it is backed by energy and military might in the form of the largest military industrial complex in the history of the world, responsible for more death and destruction of humanity and the environment than any other single entity over the last 70 years.

Unlike Bitcoin which currently consumes a greater proportion of green/renewable energy than virtually any other industry, the US military operates almost exclusively on carbon-emitting fossil fuels and has left a trail of dead and wounded, mostly innocent civilians with brown skin, whose only crimes were being born with our oil under their feet.

The US military is by far the single largest fossil fuel consumer in the world. Isn’t it curious that so little attention is focused on reducing the military’s dependence on fossil fuels? Where are the ESG proponents on the topic of the single largest contributor to global greenhouse gases?

Will you stay silent on the subject now that you are aware of the fully-loaded costs of supporting the USD as the world’s reserve currency?

By contrast, Bitcoin uses less energy than the world’s hair dryers to secure a considerable amount of value without the need for violence. If we consider the full extent of externalities required to secure the current monetary system, Satoshi’s invention of Nakamoto consensus starts to look like an alternative worth considering for at least some of the world’s wealth.

[rspeele]

It is ridiculous to blame USD for costs that would still exist, arguably not even reduced in size, if we didn't use USD.

[lowkey]

Without having to defend the USD as the global reserve currency, the US military would still have to exist, yes, but not nearly on the scale it does today.

Here is a US general Wesley Clarke explaining in his own words in 2000, before 9/11, how the US was already planning to invade 7 countries, all of whom had the nerve to act outside USD hegemony and some of which had the audacity to run their own central banks outside the approved system.

https://m.youtube.com/watch?v=6Knt3rKTqCk

[danShumway]

> If we are comparing energy consumption of monetary systems I think we also need to look at the incumbent system’s costs with open eyes.

This is kind of goalpost moving. It's fair to ask about the human costs of other monetary systems, but I thought we were just trying to determine whether or not Bitcoin was efficient. It's not. But whatever, I'll follow that goalpost for a little while.

In isolation I do think that criticism of government policy to value/devalue and secure traditional currencies would be a good argument... if Bitcoin was well positioned to be a replacement for traditional currency.

But by your own admission Bitcoin isn't primarily a system for transactions, it's a system for storing value (actually I would argue it's primarily a speculative asset, not a traditional value store, but whatever, it doesn't matter). Bitcoin is not in its current state trying to replace dollars, because dollars need to be good at transactions, and Bitcoin is bad at transactions: it's wildly inefficient and environmentally unfriendly, it's slow and has high fees, it requires you to essentially move off chain to get anything approaching a normal transaction experience.

> Isn’t it curious that so little attention is focused on reducing the military’s dependence on fossil fuels?

If the biggest problem with the US military complex was its environmental cost, I would sleep better at night. I think a big reason why people don't talk a lot about how much carbon the military emits is because they're too busy talking about the massive human cost.

But this is kind of silly; a lot of Bitcoin's critics do criticize the military. I'm not here as part of a conspiracy to prop up government invasions of other countries, I just think your "currency" is bad and has fundamental flaws.

> Will you stay silent on the subject now that you are aware of the fully-loaded costs of supporting the USD as the world’s reserve currency?

Bitcoin is not going to replace USD. It's technologically incapable of doing that; it only supports 7 transactions per second and the last time anybody tried to fix that problem, the community hard-forked and had a giant schism. Because of course they did, Bitcoin isn't optimizing for transaction speed or transaction fees and most of the community doesn't care about any of the high-minded goals that Bitcoin was originally sold on. They just want an asset that goes up in value, so eventually they can convert it back into USD.

This could be a longer conversation, but while Bitcoin was originally based around some ideals like democratic access to currency and reduction of reliance on military/global power, I feel like it's kind of silly now that we can look at how Bitcoin has played out to say that the whole movement is still about raising people up and democratizing finance. Bitcoin is primarily a speculative market, it's not driven by ideals at this point.

Also just as a sidenote, but even in a world without traditional finance, most countries would still probably have a military; so even just the core idea of "tanks use too much power" is a little weird to me given that the US is not going to throw away all of its tanks if it transitions off of USD. Bitcoin does not get rid of the concept of exploitation, the US can still steal another country's oil and then sell it for Bitcoin.

> Bitcoin uses less energy than the world’s hair dryers

Not sure if you intended this to sound like a small amount of power, but that is a heckin large amount of power in order to secure such a small proportion of the world's wealth that it has next to no impact on the current exploitative measures taken to secure other existing currencies.

Bitcoin does not do enough to address the exploitative nature and human cost of securities like gold in order to justify its enormous energy expenditure, and there is little reason to believe that it is capable of scaling to the point where it could address those problems, and there is a ton of reason to believe that if it did manage to scale to that point it would in the process start consuming even more energy.

Even taking everything you've said at face value and assuming that Bitcoin is actually just straight-up liberating value from a highly exploitative system (rather than in more than a few ways participating in that same system) -- no; quite frankly, it is not worth using the same amount of energy as Sweden just to liberate a measly 3% of the world's wealth. That is too inefficient, it costs way too much power to do way too little. Come up with a more efficient way to secure that wealth, preferably one that actually scales.

[lowkey]

No moving of goalposts here. OP claimed that Bitcoin’s primary purpose was transactions. I countered that the primary purpose was network security not txs, though pointed out two transaction use cases where Bitcoin beats every other payment method or monetary system.

You compared Bitcoin’s energy use to nation states. I compared it with hair dryers and one nation’s military, while pointing out that the US military’s energy consumption was only a small portion of it’s fully loaded cost in terms of externalities (others being: death, destruction, fossil fuel usage greater than any other entity to ever exist)

Again, no moving of goalposts. Simply fair comparisons of different forms of money.

Bitcoin never claimed to replace all uses of USD or gold, but it does reduce the total external costs in many important ways while providing the world with a powerful new alternative that is peer to peer, can be self-custodied, and has a fixed low inflation rate from now to infinity making it harder money than any that has ever existed.

[danShumway]

If 50% of Bitcoin users went away tomorrow or stopped transacting, almost certainly the amount of electricity used would decrease, because the value of the coin would decrease.

This is like saying that if the US banned gold that gold mines wouldn't be impacted at all. Demand (and thus profitability of the asset) fuels mining.

People mine because mining is profitable, not to secure the network. Securing the network is a side effect.

[steeve]

True, but the network serves a purpose. Without transactions, would there be a network to begin with? Therefore, you can model a per transaction cost within certain bounds.

[lukeschlather]

It's incorrect, but absent Bitcoin having an actual plan to increase transaction volume it's a reasonable way to imagine what a world where Bitcoin actually functioned as a currency might look like. It's possible that the bitcoin design is fundamentally flawed and cannot scale, but in some ways, OP has a very charitable way of imagining what it would look like if Bitcoin could actually function as a global currency.

[paulgb]

It's not so odd. Those people are willing to transact in Bitcoin because it's sufficiently secure against a 51% attack. So in a sense they're consuming the benefit of the overall network hashrate, even though the cost they pay is heavily subsidized.

[danans]

> Miners mine to secure the network

They are ultimately humans, not machines, so they mine for the mining reward, not as a public service of securing the network.

Transactions are the only service provided by Bitcoin (what else is a currency for other than transactions?), so it's totally fair to consider it's energy cost per transaction, especially since the banking system is evaluated in the same way.

[tromp]

> there is not a certain amount of electricity needed per transaction.

Not yet. But that will be increasingly the case in the coming decades as the block subsidy gets repeatedly halved into insignificance. Then the brunt of Bitcoin's security (protection against 51% attacks) will have to be borne by transaction fees.

[danbruc]

How saturated is Bitcoin currently? I stopped following the details years ago but I guess Bitcoin could not handle a ten fold transaction rate increase today, right? Which means at best - if you completely saturate the network - you could get the costs per transaction down to 10% of the given number.

[speedgoose]

If all people stop using bitcoin, the amount of electricity goes to zero.

[yobbo]

Classic confusion about marginal costs.

The price of bitcoin is supported by inflow of "fresh money" from newcomers. If this inflow stopped, for how long would mining continue?

[cookingrobot]

The price of Bitcoin is not supported by the amount of fresh money coming in. It’s based on the lowest price a current Bitcoin holder is willing to sell for.

If all Bitcoin holders decide tomorrow that it’s worth $200k per coin, then that’s the price, even if there are very few buyers at that price.

It’s like the price of a stock: it’s not dependent on trading volume.

[yobbo]

To be accurate, the equilibrium price is the highest price a buyer is willing to pay, and that someone is willing to sell for.

If the highest price is zero, there is no market, and mining would cease.

[leifg]

So?

It’s so frustrating having these discussions about energy consumption.

If you make a statement about the absolute energy consumption of a PoW blockchain, the first response usually is “But this other thing consumes way more energy!!!!”

In order to compare the consumption of the network to anything you will need to calculate it by some unit of utility. Transactions is the only metric that makes sense.