Hacker News Reader

Comment by hn_throwaway_99 | Hacker News Reader

Proof-of-work is so obviously, undeniably, mathematically guaranteed to be a crypto dead-end, that I'm gobsmacked we're still having discussions like this.

Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin. It's really just simple arbitrage: if the value of Bitcoin goes up, but the electricity cost doesn't, then you have essentially 2 different prices for 1 good, which should attract more miners to spend more money on electricity. And, indeed, this is exactly what has happened.

The problem, of course, is that if it takes an Argentina's worth of electricity to run Bitcoin now, how does the world manage if Bitcoin has a 10 or 100 times greater market cap? Obviously it can't, and world governments would rightly shut it all down before the majority of our electric output went to mining BTC.

There is simply no way around this issue if you're on a proof-of-work system, and given that, it's so easy to see that BTC is guaranteed to fail at some point. It's like we're on a train rushing towards a bridge that's out, and everyone can see the bridge is out, but somehow we've convinced a large number of people that we'll "jump the gap" when we get there.

View on HN

You are making very bold statements ("of course", "obviously") on a topic that has been deeply analyzed over the years without any decisive conclusions.

For one example, I really appreciate Lyn Alden's analysis [1]

[1] - https://www.lynalden.com/bitcoin-security-modeling/

I'm making bold statements because I've read reams and reams of treatises on the topic, and every single one ignores some basic, simple, fundamental truths that are inherent to a proof-of-work system. I certainly don't believe I'm particularly smarter than average, and there are definitely other people saying the same thing I am, and certainly much smarter people than I have realized this issue and are transitioning other cryptocurrencies to other proof systems, but I have come to believe that many people are content to keep the music going, even if they know it's going to stop at some point.

When it comes to mining BTC, it really doesn't matter much how miners are paid: through block rewards, transaction fees or some other sort of "security fee". Miners spend money on electricity, and get BTC in return. Furthermore, it's, again by design, a competitive system: those who are able to mine faster (i.e. are spending more on electricity, on average) will win the "block contests" faster and get more BTC.

All other arguments are just "yada yada yada" unless you are somehow arguing that electricity spend is no longer proportional to chance of mining a block, at which point I'd argue that's probably no longer proof-of-work.

This is not a hard concept. If I'm an idiot (which is totally a valid proposition), somebody should be able to explain how a proof-of-work system can get around this "arbitrage problem" very simply - it shouldn't take pages and pages and pages of circular argument.

You're making sweeping statements here, when in reality it depends on both the block reward AND the price of bitcoin. Currently, the block reward has decreased as the price of bitcoin has increased. If the market cap of bitcoin stays the same, then the mining reward in terms of energy will decrease.

Argh!!! I get this kind of response all the time, and it's frustrating for one simple reason: miners certainly DO NOT CARE where their reward comes from: block rewards, transaction fees, whatever.

At the end of the day, it's a very simple calculation for them: electricity costs in, Bitcoin value out. No matter where their Bitcoin reward comes from, they will not (for long anyway) spend more on BTC then they are rewarded.

> If the market cap of bitcoin stays the same, then the mining reward in terms of energy will decrease.

NO! The simple rules of arbitrage would ensure that more miners would be incentivized to beef up their electricity spend to mind faster to win the reward (again, block rewards or transaction fees, doesn't matter).

Please just stop and think for a moment FFS.

Miners care about where their bitcoin comes from. Let's imagine the case where 100% of their income comes from the block reward and 0% from fees (seeing as this is nearly the case right now):

1. the block rewards are getting CUT IN HALF every once and a while. If the price of bitcoin stays the constant, their rewards in terms of dollars will be CUT IN HALF. The only way for their income (the upper limit on they can spend on electricity, hardware, etc.) to increase is for the price to increase by more than DOUBLE every time the mining reward gets cut in half.

2. the block rewards increase the supply of bitcoin, decreasing the market cap/price over time. It depends on the emission schedule. Even if the block reward was a constant 1 bitcoin per block, then you would just have a different (logarithmic) inflation/block reward graph with the same characteristics.

Let's say 100% of income comes from block rewards of 1 BTC. If it costs 0.5 BTC to mine a block then miners will beef up their mining capacity so they can mine more. In a vacuum, this would be great, but because of how mining works (winner takes all), the cost of mining would go up as the number of miners increases. This will keep happening until the cost to mine is close to or equal to 1 BTC (and because of currency conversion, the cost might even exceed 1 BTC today but miners are hoping that the price goes up so they can sell at a profit).

Let's say tomorrow, the income is cut in half. All of a sudden, most of these miners would become unprofitable and stop. However as a result, the cost of mining will go back down to 0.5 BTC. It's a self balancing system but that also means that it's fundamentally flawed for reasons mentions by previous commenters in this thread.

I think you are talking past each other. Leaving out transaction fees for a second (which are currently a very small percentage of the miner reward). The block reward is programmed to half every 4 years, taken alone this together with a fixed market cap would mean that the mining electricity spend would also half every 4 years. So the energy consumption of bitcoin will grow only so long as its market cap also doubles every 4 years. That's a tall order.

Bringing back in transaction fees-- I don't think anyone really knows what they will be, but it will be a long time until the dynamics of that are more important than the dynamics of the block reward halving and the market cap.

> At the end of the day, it's a very simple calculation for them: electricity costs in, Bitcoin value out. No matter where their Bitcoin reward comes from, they will not (for long anyway) spend more on BTC then they are rewarded.

True, but the hypothesis in your original comment, that energy expenditure will equal Bitcoin total market cap, does not follow. As other commenters here note, with each block reward halving, the energy required to mine a new block becomes more and more decoupled from the price of Bitcoin itself. In the distant future when block rewards are minimal and fees make up the majority of miners' rewards, only miners who can utilise the cheapest sources of energy will be able to mine profitably, which in the "negligible-block-reward" era will be for the lowest fees. And people making transactions will not pay more fees than they need to, so the existence of miners able to mine profitably for lower fees will in turn bring down average fees. I think all of this should be uncontroversial; it simply follows from the Bitcoin protocol and the work of Adam Smith.

So, given these incentives and Bitcoin's difficulty adjustment mechanism, in the future the only miners capable of making profit from mining will be those with access to the cheapest forms of energy. Anyone who can mine for marginally lower cost is going to push out other miners. This will likely mean the majority of the energy used for Bitcoin mining will come from otherwise wasted energy - such as flare gas, remote hydro not economical to transmit to where people actually live, excess wind and solar supply when demand don't line up, that kind of thing. (And, as an aside, it's probably not going to involve burning lots of oil and gas, since these are more useful to humans in other ways and will therefore cost more.) Obviously more efficient hardware will have the same effect, so there will also be an incentive there. In the end it will have very little to do with Bitcoin market cap and far more to do with the availability across time and space of cheap energy.

> And people making transactions will not pay more fees than they need to, so the existence of miners able to mine profitably for lower fees will in turn bring down average fees.

Only to a point. Since a limited number of transactions can fit into a block, users competing to get their transactions mined in a timely manner will drive the fees up. That price point can be well above the cost of many forms of energy.

So as long as Bitcoin is still valued and there are enough people wanting to make transactions, there will still be miners burning useful energy.

There is still some coupling over time - the miners' bitcoins depend on transactions happening to retain their value

> depends on the block reward AND the price of bitcoin

The direct block reward goes to zero over the years, but the miner's fee will still be a positive amount.

Each miner will be competing for the miner's fee, and will buy more electricity if they can mine more blocks (and get more miner's fees).

The rewards will be proportional to the market cap of bitcoin. If it goes up, so will the investment in electricity.

> making sweeping statements here

The sweeping statements that poster made are the results of a microeconomics analysis.

There is a large & "efficient" market competing for bitcoin mining rewards by buying more electricity.

This is tapping into laws of economics on the level of supply/demand curves: Well studied scenarios that become more accurate the more efficient the market is.

>The direct block reward goes to zero over the years, but the miner's fee will still be a positive amount.

Yes.

>Each miner will be competing for the miner's fee, and will buy more electricity if they can mine more blocks (and get more miner's fees).

Yes.

>The rewards will be proportional to the market cap of bitcoin. If it goes up, so will the investment in electricity.

No.

You're conflating two different things here: A reward-dominated bitcoin and a fee-dominated bitcoin.

The reward takes money from bitcoin owners in the form of inflation. It decreases the market cap by increasing supply.

The fees take money directly from bitcoin users.

In the reward-dominated case, miners are funded by new demand for bitcoin, which props the market cap up. In the Fee-dominated case, miners are funded by bitcoin users who compete with each other for limited bandwidth. In either case, the money being given to miners by these two parties is always greater than or equal to the money spent on mining (power, hardware).

What we've seen so far is a reward-dominated bitcoin in which demand is very high and the market cap increases as the block reward decreases. I predict the demand will stabilize and the reward will continue to decrease. This means there will be less mining. Eventually the reward will asymptotically trend to zero in which case bitcoin will become fee-dominated.

Fees are dependent on the supply/demand for bitcoin bandwidth (the supply is a static 1MB/10 mins), not market cap or price.

It's weird how so many of the arguments against what I've originally written pretend that it matters whether the reward comes from a block reward or transactions fees. It does not matter. Certainly miners don't care how they get paid - for them they have a simple calculation of "bitcoin value I get out must be greater that electricity cost I spend".

And the fundamental idea behind proof-of-work that they amount of work you're "proving" must be enough to make double-spend attacks infeasible. If the electricity cost to mine a block is low enough compared to the value that could be gained by a double-spend attack (which is of course comparable to total BTC market cap), then the network is not secure.

>It's weird how so many of the arguments against what I've originally written pretend that it matters whether the reward comes from a block reward or transactions fees. It does not matter.

It does matter. Block rewards are a transfer of value from bitcoin owners, which increases as people BUY bitcoin. Block fees increase as people USE bitcoin. Those are two separate things. What we are seeing now is mining that goes far beyond the mining necessary to secure the network, and that's due to (temporary?) speculative demand which increases the value of the block reward.

>And the fundamental idea behind proof-of-work that they amount of work you're "proving" must be enough to make double-spend attacks infeasible.

well, infeasible by a single antagonistic party. There's also the cost associated with the depreciation of the hardware. Even if mining somehow required no electricity, then the network would still be secure because there is some Time-Value cost associated with owning computer hardware.

>If the electricity cost to mine a block is low enough compared to the value that could be gained by a double-spend attack (which is of course comparable to total BTC market cap), then the network is not secure.

I agree. I think bitcoin will eventually succumb to selfish mining attacks.

"It's weird how so many of the arguments against what I've originally written pretend that it matters whether the reward comes from a block reward or transactions fees. It does not matter."

"It does matter."

Miner do not care, so it does not matter.

This is amazing to watch. hn_throwaway_99 complains that he only gets handwaving and circular arguments against this, only to get multiple answers which are exactly that.

Mining will always spend a large chunk (>50%? depends on the complete distribution of efficiency off all miners) of rewards on electricity. A quick search shows this to be 0.5% of all electricity used in the world. If Bitcoin price goes up 10x right now, this would shortly increase to around 5%.

But, there is also a thing that miners still have to pay electricity in dollars or whatever, so they are forced to sell on exchanges. This creates a downwards pressure on Bitcoin price.

Currently this could be somewhere around $billion per month. So, you need a billion a month of fresh sucker money, just for the bitcoin price to stay the same.

Aren't the fees dependent on the velocity of Bitcoin, and not the market cap?

There's some dependence in the fees, since they'll be paid in Bitcoin, and that there's still the odd new Bitcoin being made, but you're still being paid per transaction, unrelated to the value of the bitcoins in the transaction

You're right that there isn't a direct causal relationship between fees and market cap, but I think the ratio between them is important. If miner revenue becomes a tiny fraction of market cap, 51% attacks become a very real threat. Then the community would need to do something to increase the ratio, such as establishing a permanent block reward.

I agree. In fact, some cryptocurrencies like have implemented constant "tail emission" block rewards. The problem is that this scares off investors and speculators because there's not as much scarcity.

When market cap / price goes up, more people enter the market to mine until supernormal profits are gone - This is the relationship being discussed.

People will enter the market which pushes up the difficulty / electricity and hardware consumption until profits are brought back down to a normal level.

Yes, but I predict people will cease entering the market once the hype dies down and people learn how flawed bitcoin is.

People will always enter the market (to produce) if there is a difference between the price of a Bitcoin and the cost to produce a Bitcoin (electricity, equipment etc).

If the price goes down and it costs more electricity to produce a Bitcoin that is economically viable, people will leave because the juice isn’t worth the squeeze.

So the same relationship works inversed too - difficulty to mine and electricity consumption is related to market cap / price from an econ perspective. I agree with your point on Bitcoin hype, but the production side is just economics and people trying to make a buck if there is money on the table.

Upon a second reading of parent poster's comment, there's nothing much to debate the conclusion but mostly debating the degree of certainty of the conclusion, which is a much lower bar.

The way I see it though, unless anyone is shorting cryptocurrency with 100% of their net worth I think they are at some level expressing that the world is not so clear cut. Sometimes meme stocks win before you run out of runway.

Knowledge something will fail eventually, and knowledge of exactly when it will are entirely different propositions.

The market can stay insolvent longer than the investor.

"The stock market can remain irrational a lot longer than you can remain solvent."

https://quoteinvestigator.com/2011/08/09/remain-solvent/

> I've read reams and reams of treatises on the topic

Care to share any of them here? I'm not that knowledgeable about BTC so I have no idea whether or not you're right, but I'd love to learn more.

start with bitcoin whitepaper. but it is a very deep subject, and also full of liars and people with ill motivations

> Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin.

Proof of work in Bitcoin requires computation of a thing called a Nonce. If a miner gets the right Nonce while hashing, they get that block's reward. The electricity value expended on mining that block is proportional to the total amount of hashes being performed by miners in the network at the time. There is no "design" for requiring the cost of this power be equal to anything directly related to the market capitalization of Bitcoin, which really means how much Bitcoin is worth compared to other markets, such as USD. Are they linked to each other through human action and decision processes? Certainly. Is it coded in the chain? No.

> It's really just simple arbitrage: if the value of Bitcoin goes up, but the electricity cost doesn't, then you have essentially 2 different prices for 1 good, which should attract more miners to spend more money on electricity.

If the electricity cost doesn't go up, then the number of Nonce being computed will only go up if more miners join the effort. If they don't, the network is still operational, and produces the same amount of new coin every so often. What is NOT obvious here is that the technology for increasing the mining rates is always increasing, so the clowns will continue to pay to upgrade their hardware over time to save on electrical costs and engage in competition with the technology. The exchange rate for Bitcoin may be related to the technological "power" that is required to keep the network running and producing new coins, but less directly so related to the amount of power burned.

> Furthermore, it's, again by design, a competitive system

By design, Humans compete. The Bitcoin blockchain could be mined and run by two clowns in a tent if nobody else cared. It's not coded to be competitive, it's coded to be cooperative. This means the the amount of power spent on mining and running the network is a cooperative function. The amount of money spent on developing and producing faster hardware is a competitive function. The two together produce something that is chaotic and interesting.

Bitcoin does have a downside, and that is that it is a viral thing that cannot be taken down by any one person or organization. It would take 51% of those running the network to cause it to fork into two. It would take an impossibility to halt it. Maybe this scares some people and maybe it makes some people excited about the possibilities. Either way, we're going to need Fusion to run it in the future.

Most of what you said I agree with, but not your last point:

> Either way, we're going to need Fusion to run it in the future.

In my understanding, the economics of Bitcoin mean that eventually mining will be performed with predominantly waste energy, which is by definition that which has no other practical use. The existence of a practical use for the energy other than for mining will push its price higher than is economical to use for Bitcoin mining, because other miners using waste energy will be able to pay less to mine the next block. It's likely that energy for Bitcoin mining will come from sources like heat from flare gas, currently just burned at oil refineries and released as photons into the night sky, leaving Earth forever. Or energy from wind turbines or solar panels during periods of low demand. Or a hydroelectric dam 1000s of km from civilisation willing to pay the transmission costs to use it. Or, heck, even heat extracted from sewage in Vegas hotels, if it's enough to run a steam turbine. Basically anywhere where otherwise useless energy can be extracted in quantities sufficient for mining. I don't think, given the costs of the infrastructure likely required for fusion, we'll be seeing it employed for Bitcoin mining long term.

This is one of the aspects of Bitcoin I find most beautiful - it monetises waste.

Ehm.... "waste energy"? I'm not sure there is such a thing. If it can be harvested and used, why should it be used for Bitcoin instead of... well, anything else?

Because energy (at least currently) can not be easily transported, to where it can be used. Factories can not easily move to power plants due perhaps increased supply chain cost or labor cost. Whereas Bitcoin mining can be moved anywhere with a shed and an internet connection.

Aluminum smelters are specifically sited to be near low cost power plants. Refined aluminum is essentially a storable and somewhat portable form of energy.

Good. Now try turning an aluminium smelter on and off when there's surplus/shortage of supply on the grid. Right, you don't.

Here's one key difference between many if not most large scale industrial processes and Bitcoin mining.

I still don't see the prerogative for bitcoin. Yes, most industrial processes are inefficient and we can and should reclaim lost energy. There are many ways to do this, and many purposes. Charging batteries seems the most reasonable: making the energy stable and portable.

Admittedly I don't know much about battery efficiency, so there may be other more efficient uses. Hydroponics lighting?

Where "waste energy" is difficult to access and available only in trickles, it's surely very unsuitable for bitcoin mining, which is energy intensive.

Frankly, there are far more pressing issues than making more bitcoin.

Who exactly do you think is going to be building the expensive infrastructure necessary to capture this low-value waste energy?

Sure, there’s conceivably times of low demand when some energy would otherwise go unused. But a “ hydroelectric dam 1000s of km from civilisation”? Heat extractors in Vegas hotels? You’re imagining a fantasy world.

Further, what you’re ignoring is that none of this is “waste”. It’s already being stored or used. Finding new demand for it simply makes it more expensive for existing users.

> ... none of this is “waste”. It’s already being stored or used.

Turned off wind turbines at times of surplus. They tell us that nope, it is not (yet) being used or stored.

> Finding new demand for it simply makes it more expensive for existing users.

It's not that simple. It could very well be the case that selling a portion of energy that otherwise has no demand to a buyer of last resort enables the business to be profitable. Not selling this portion of energy would mean that the operator would have to charge their other customers more in order to be profitable.

So it could in fact make things less expensive indeed.

I'm not sure you understood my comment correctly.

> mathematically guaranteed to be a crypto dead-end

Your argument is not mathematical at all. It assumes first that bitcoin needs to appreciate 10-100 times current value. And second, that political willpower exists to then shut it down purely because of energy usage. Neither of those conditions are mathematically guaranteed, nor are they even in the realm of mathematics. I won't even address the underlying assumption that energy usage is unilaterally bad.

One possible outcome, for example, is that bitcoin just hovers around its current price for a long time. If feels like every discussion on bitcoin is made of people who either think that it must either take over the world or must go to zero.

OK, of all the comments I've seen trying to refute what I've written, yours is the only one I've seen that at least makes sense.

Yes, I totally agree, the crux of my argument is really just "in a proof-of-work system, it is undeniable that electricity spend is proportional to total network market cap", so if the market cap doesn't change, then things can certainly putter along.

That said, if the idea is that more and more of the world's economy will be represented by BTC transactions, which these days is pretty much what every person who owns BTC believes, then the market cap can't stay static. I also have a very difficult time believing that if the market cap did stay static for 10 years or so that people wouldn't start to realize "the jig is up".

But yes, I totally concede, if the market cap doesn't rise, current electricity spend doesn't need to rise either.

What jig would be up? Mining as profiteering? For sure, and the end of mining is an inevitability. Bitcoin as currency could continue though, no?

By "the jig is up" I mean that the vast majority of reason people invest in Bitcoin is as a speculative investment. If it stays flat for 10 years, it loses its appeal as a speculative investment, and then would collapse.

The other reason some "true believers" (or, as I refer to them, insane cultists) invest in Bitcoin is because they believe it will be "the world's currency", i.e. that it will replace major currencies like the dollar. But, were that to occur, it's market cap would skyrocket, at which point it would require a planet-and-a-half's worth of energy to stay secure, which is also not happening.

With a price collapse, the energy consumption would also collapse, no?

> require a planet-and-a-half's worth of energy to stay secure, which is also not happening

But what are the implications if it does?

> at which point it would require a planet-and-a-half's worth of energy to stay secure

Ehm, either that or the price of energy increases or the percentage of reward per block mined to mcap decreases.

No, total dollar amount of electricity expenses spent on mining does not depend on the price of electricity.

Today, miners are getting paid about $1.5 billion per month. If they like money, they will keep adding new miners and run them until they all in total spend less than, but close to that amount.

So, $1.5 billion per month of fresh money needed to offset this pressure on Bitcoin price.

There goes the argument that Bitcoin is a good "store of value."

> total dollar amount of electricity expenses spent on mining does not depend on the price of electricity

I agree. As far as I understood the issue was not with the money spent on energy but with the energy spent, no?

> So, $1.5 billion per month of fresh money needed to offset this pressure on Bitcoin price.

If I understand correctly, you're saying that BTC inflation is a problem and I believe this is true, both in the economic and ecological sense.

But there was no way in 2009 to foresee the optimal inflation rate (halving every 4 years isn't too bad actually) and modifying it now defeats the founding principles of BTC, so it'd be a big no no. So it was kind of a trade off and will continue to be, every 4 years a bit less though unless we hit hyperbitcoinization.

From an economic point of view the investment into bitcoin is only ever relative to other assets and most of them inflate far more (relative to their mcap).

The end of mining is the end of bitcoin. If there are no miners contributing to create the network consensus, then anyone can come in with a 51% attack and destroy the integrity of the block chain.

> Your argument is not mathematical at all. It assumes first that bitcoin needs to appreciate 10-100 times current value

Satoshi consensus doesn't care about the numeric value of Bitcoin at all. It cares about the value of the transactions conveyed by bitcoin.

Satoshi consensus is the observation that if the energy required to fork the network and roll back a transaction costs $X, then it's financially non-viable for someone to roll back a set of transactions that is worth less than $X. That's all. If the value transacted on bitcoin increases above $X, then it is potentially financially worth it to attack the network, unless the amount of energy expended increases accordingly. So the value of the transactions is directly tied to the (value of the) energy expenditure.

As a side note, there is no automatic mechanism built into bitcoin that ties these two values together. The assumption is that the value increases and the rewards decrease and it all sort of works out (especially if the network transitions to transaction fees instead of inflation-based block rewards). But there is no mechanism that it must, and if the value dropped really hard all of a sudden, for example, it might suddenly be more viable to attack the network. Attacker energy cost will significantly decrease as rational actor miners realize they're taking a loss and turn off their mining hardware, miner energy expenditure will decrease, which decreases the energy (and cost) required for attackers to fork and roll back transactions.

It's an interesting variation of the "frisbee on the roof" attack, like a "fire all the janitors and then shit on the floor" attack. Nobody ever wants to go back into that business because there's shit all over the floor, but without any customers there's no money to hire janitors either. The network now enters a terminally unrecoverable state, without manual intervention from outside actors.

> if the value dropped really hard all of a sudden, for example, it might suddenly be more viable to attack the network

In practice, users of bitcoin would simply wait a longer time for more block confirmations, namely if you estimated the cost of a double spend attack was $X you would just wait for block confirmations with a total of more than $X of fees + block rewards such that the transaction you are confirming is deep enough in the blockchain to not be economical to include in the double spend attack. Note that the slowness of difficulty adjustments is also a defense here against the further problem of $X decreasing very quickly such that a transaction you had already confirmed becomes viable to attack. Because difficulty changes only every 2016 blocks, if all mining stopped and $X theoretically became 0, in practice you won't get to the next difficulty as no new blocks would be found, so $X in practice wouldn't move.

It would be very strange for the current situation to be a long term stable equilibrium for Bitcoin.

Bitcoin is a zero sum ecosystem, minus mining fees. So it's a net negative sum by some $20-30m per day in electricity and hardware costs.

Why would it stabilize around this particular value?

The inflation due to bitcoin mining is already smaller than other systems that are widely considered stable. Note that I don't think bitcoin will actually be stable, but I think inflationary pressures are almost insignificant as a factor that would drive it to collapsing.

For bitcoin, 90% of all bitcoin has already been mined. So the remaining dilution for all time is in total just about 11%, and the current annual rate is 1.8% inflation. Compare this to gold, which has 2% inflation and no max supply. There are also technological innovations that could inflate gold much faster, e.g. asteroid mining. Compared to USD as measured by CPI is 8%, and the Fed's own target for CPI inflation is 2% annually, so this is much higher than bitcoin and also unbounded total dilution.

Don't confuse money supply growth and inflation. (Unless you're a monetarist, of course, but their doctrine has been quite convincingly refuted over the last decade and a half.) Inflation is change in the price of real goods.

BTC experiences 1.7% money supply growth p.a. until the next halving, and it had deflation on the way up to $60k. BTC had 100% inflation since then, obviously, way worse than fiat.

Fiat had massive money supply growth since the GFC, but virtually no inflation until now (which can be explained by supply chain problems due to COVID, war, and too loose fiscal policy (Biden's stimulus arguably too large, after Obama's was too small)).

You're wrong. Block rewards halve every 4 years. Next halving is in 2 years. This means that electricity use funded by block rewards trends to zero. Miners earn also transaction fees, which will gradually replace block rewards in the long term. Transaction fees are revenue from utility.

It's a well designed incentive system; it will keep Bitcoin running as long as people find it useful, and it won't eat the world.

The issue is total transaction fees must be kept high enough to avoid miners defecting into double spend attacks. That would be easy if people where using it as a high turnover currency, but if people want to use it as a long term store of value transaction fees can quickly become many orders of magnitude smaller than the value stored in Bitcoins.

It need not be illegal. Short Bitcoin, preform double spend attack between addresses you own, profit.

This is the same conclusion I came to. Bitcoin's incentive structure makes sense if the users of it gain value by being able to use it as a common medium of exchange. If you look at the money that e.g. mastercard and visa pull in, it makes sense that you could sustain bitcoin off of that indefinitely. It could also conceivable result in a relatively lower market cap if people do not in fact save money in bitcoin, reducing the threat of 51% attacks. As a store of value, the incentives no longer make sense: for the 'store of value', it must be deflationary, thus mining cannot produce new coins indefinitely. But it also cannot be too expensive: if just keeping your coins requires paying a significant percentage of the value of them to secure them on an ongoing basis, you're ultimately losing in just the same way as someone in an inflationary coin does. There's also no particular mechanism which makes sure said price is actually correct: ultimately you need to figure out how much mining is 'enough', and also realistically have some mechanism to spread those costs out evenly amongst holders of bitcoin, lest it become a tragedy of the commons where no-one wants to contribute.

We don't know what will happen. These problems and potential solutions have been hashed for years, but we'll have to just wait and see. You'll have to just believe that the economic incentive the preserve all the value will result in a solution.

There's a simple solution to this... just wait longer

51% attacks work on arbitrarily long timescales. Though obviously you can still do short double spend attacks at some probability with less hashing power.

An attempt to rewrite the ledger with a 51% attack would be detected almost immediately. The Bitcoin in the rewritten ledger would quickly become worthless, including that belonging to the keys owned by the attacker. All that mining gear used to perform the attack would then be worthless too. This doesn't seem to me like a profitable attack vector once you consider the complete picture.

Shorting Bitcoin doesn’t require you to ever own any Bitcoin. It does require a counter party(s) that would pay out.

The value of the gear is just a question of the future profits it can generate. If that future revenue is less than what you get from the great short then it’s profitable. And that’s assuming you can’t reuse the same equipment for some other purpose.

> doesn’t require you to ever own any Bitcoin

It requires you to borrow bitcoins from someone -> sell them -> do the attack -> buy them back -> and then return them.

If you could borrow enough bitcoin, and gain enough hash, and crash the price enough for it to be profitable, it could possibly kill Bitcoin. Seems like an insane gamble to me though.

I agree.

That said if I where writing a story I would have the protagonist gain the hashing power first, something strands their assets such as increased electricity prices making mining unprofitable or a next generation of hardware coming out etc. Then they come up with the idea of a big short.

The cool thing about that also is the hash power can be accumulated secretly and dumped on the network all at once

If you can maintain it, sure.

But you can only double spend coins you already control, so at this point you would control 51% of global SHA2 hash and have a bunch of bitcoin that you would be double spending?

Who would accept that as payment? What incentive would you have to do something like that since you would be taking a massive loss?

Maybe as a government attack?

The assumption was the point of the attack would be to destroy Bitcoin’s value to make money from a short position not to sell the same coins twice.

I think you are the first person I have read to say this, but I totally agree.

The cost to secure the Bitcoin network is a function of the market capitalization. If the market cap increases so will the cost of securing the network.

If Bitcoin market cap continues to increase, the resources required to secure the network will start to crowd out other economic activity.

I feel like we are seeing the first glimpse of this with the chip shortages.

>Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin.

This is mostly true, but I think you're conflating a couple different things here. At market equilibrium, you would think it would look like the cost of power + hw depreciation = crypto price * (block reward + fees).

>if it takes an Argentina's worth of electricity to run Bitcoin now

It might be justified if bitcoin could replace the traditional financial system (think about how inefficient that is), but I do not think it can.

>how does the world manage if Bitcoin has a 10 or 100 times greater market cap?

I don't think bitcoin's market cap can multiply by a factor of 10 or 100

> It might be justified if bitcoin could replace the traditional financial system (think about how inefficient that is), but I do not think it can.

Bitcoin is already way more inefficient than the current financial system. Just compare cost per transaction between Bitcoin and Visa, for example. People seem to think that Bitcoin will somehow replace the entirety of the financial system when it's not even designed to do that and only really handles transactions.

Yes. The fees are mostly due to bitcoin's unreasonable 1MB blocksize constraint. Most other cryptocurrencies do not have this problem because they make a more reasonable throughput to bandwidth/storage tradeoff

That constraint is perfectly reasonable and rational, not a thumbsuck.

It's neither reasonable or rational. As technology improves, the cost of bandwidth and storage decrease, and latency decreases. Rationally block size and block time should reflect that.

Without some way to squash history, I don’t see how it’s possible to fix this. The bitcoin chain is currently 400gb. If we bumped it up to 10mb blocks we would be looking at several terabytes while still not being anywhere near big enough to handle real transactions.

It does not really matter, as only miners (and large buisnesses accepting bitcoin) would need to run a full node.

All these non-mining nodes have no real effect over decentralization, as they can't counter a 51% attack.

Fully agree with your analysis. Other's have nitpicked your exact words but your point is totally correct: There is a cap of how much Bitcoin price can increase, at least before multiple halvenings. If it increases above this cap, the mining power usage will become so absurd that external forces will come into play (regulations, and very negative perception once people realize that it's absurd).

That's why I don't believe in any crypto that doesn't at least have a path to switching to PoS (or some non PoW system).

> Proof-of-work is so obviously, undeniably, mathematically guaranteed to be a crypto dead-end

Bold claim.

> Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin

False statement (as it ignores the bitcoin emission schedule and transaction fees).

> The problem, of course, is that if it takes an Argentina's worth of electricity to run Bitcoin now, how does the world manage if Bitcoin has a 10 or 100 times greater market cap? Obviously it can't, and world governments would rightly shut it all down before the majority of our electric output went to mining BTC.

Constructing an irrelevant hypothetical based on the initial false statement.

> simply no way around this issue if you're on a proof-of-work system, and given that, it's so easy to see that BTC is guaranteed to fail at some point

More bold declarative statements with nothing to back them up.

Learn more here: https://en.bitcoin.it/wiki/Controlled_supply

Such a good point. I get that it's tempting for people to hold on to the idea of Proof-of-Work because it was 100% necessary to get the idea of crypto off the ground, but this is like people holding on to the idea of burning coal to run trains forever.

Bitcoin was the proof of concept, but it to me seems 100% obvious and inevitable that something else will replace it (if crypto continues to exist, which I believe it will) -- ESPECIALLY since the whole point of money is fungibility-in-practice.

> The problem, of course, is that if it takes an Argentina's worth of electricity to run Bitcoin now, how does the world manage if Bitcoin has a 10 or 100 times greater market cap?

The amount of electricity is not proportional to the market cap. It's proportional to the difficulty. You can create a copy of bitcoin where the difficulty is keep very low and add a very big amount of transactions and be very easy to make fake blocks and forks and other nasty stuff. The difficulty is adjusted to keep the time to find a block somewhat constant.

One problem to handle a bigger marketcap is the block size. One solutions is to increase the block size, another is to use a secondary chain. I'd like to increase the size, but most people prefer the other.

Also, when you add the energy that goes directly from sun light to soy beans, I think that we are in Argentina still using a little more more energy than bitcoin.

>The amount of electricity is not proportional to the market cap This misses the point of the argument, which is to say that if the market cap of bitcoin goes up, the marginal benefit of mining by definition goes up. This in turn drives up demand for electricity, until the marginal cost of additional mining units and associated power equilizes. That's not to say that there isn't a benefit to a more efficient proof-of-work scheme, as in theory if the scheme can support _all_ future and current transactions like the current banking system, then there's a cap on energy consumption. However, what OP points out is that the speculation and monetary utility of Bitcoin both act to increase power consumption due to basic economic forces.

>Also, when you add the energy that goes directly from sun light to soy beans, I think that we are in Argentina still using a little more more energy than bitcoin.

This is a non-sequitur. The point of using Argentina is to illustrate that in verifying transactions, Bitcoin uses as much energy as millions of people lighting their homes, watching television, running air-conditioning, heating water, charging their phones, etc etc etc. Nobody is claiming that Bitcoin consumes more metabolic energy or solar energy, OP's just highlighting that Bitcoin's electricity costs are potentially unsustainable.

> The amount of electricity is not proportional to the market cap. It's proportional to the difficulty.

But the difficulty is proportional to the amount of miners, which is proportional to the profitability of mining, which is proportional to the price, which the market cap is proportional to (i.e. price x supply). So electricity is indirectly proportional to the market cap.

> One solutions is to increase the block size, another is to use a secondary chain.

Those solutions don't address the energy spent on mining, they just allow more people to transact. The former means more transactions (more fees), the latter means fewer but more expensive on-chain transactions (more fees). Either way I believe the block reward (and thus the amount of energy miners will throw at getting that reward) will be roughly the same, it'll be a function of how much people value and want to use bitcoin.

"It's proportional to the difficulty."

You just confused the amount of electricity it takes to mine one block, vs, the amount of total electricity all miners spend.

And the total amount of electricity is indeed proportional to the market cap. (with a constant of proportionality being halved every 4 years).

As the original poster said, it is ridiculous that we still have to debate this.

> the electricity value expended on mining is proportional to the total market cap of Bitcoin

That's not true, it's at most, equal to the printing rate of bitcoins (coinbase transactions)

> the electricity value expended on mining is proportional to the total market cap of Bitcoin

That would be the case if price doesn't vary much over time. But marketcap is current supply * current price, while the large majority of current supply was mined when bitcoin was WAY cheaper and thus much less electricity was spent on it.

Here you are confusing "total amount of electricity Bitcoin used trough history" and "amount of electricity Bitcoin miners are currently using".

The second is proportional to the current Bitcoin price, the first one, who cares.

How is this different from any other good or service? Is it bad only for bitcoin?

What is the problem with it? Do you know why oil has the price it has and why we built all the enormous infra structure to extract it? Because it is useful, and we can kinda measure it by its market cap. If the market cap of bitcoin is huge, that means that...... its usefulness is huge (or you can argue it is a bubble, but then it's a different argument). So the higher the market cap the more useful it is, and its a good thing. The profit and loss is what will tell if we are "wasting" resources or not, if it is profitable, then that means that people value bitcoin more than the electricity it uses, that is not a problem.

Not only is there no way around it, but it's actually impossible to start new networks from scratch based on proof-of-stake due to the "nothing at stake" problem. If attackers don't care if their stake gets burned (because the value is low) then there's no longer an incentive for attackers not to validate on multiple chains at once (which is what staking is supposed to prevent). All new networks must go through a proof-of-work stage of evolution.

Further, the existence of proof-of-stake coins at all means that decentralized exchanges exist, which makes it impossible to ever kill proof-of-work in a legislative sense. Even if you killed fiat exchange for $SHITCOIN entirely, once there is a coin that has sufficient value, people will be happy to have bitcoin instead of USD, or trade to Bitcoin and then over to USD. You can either ban them all entirely, or you have to live with both PoW and PoS together.

Unfortunately, crypto is a memeplex, it's like a corporation, it's a self-reinforcing living entity that supports itself (maintains homeostasis) through a series of rules and incentives, and now that it's been created it can't be easily killed. Every individual is financially incentivized to keep playing their part, so everyone would have to collectively agree to not make money by taking part, and as people fall out the incentives become steeper and steeper to participate. Would you mine on your card for $100 a day? $1000 a day?

We will just have to live with destroying our planet, because now that it's been created it's virtually impossible to kill. That was the whole point. With apologies to Alien: it’s biologically engineered to be a survivor... the perfect predator (of certain other memeplexes - meaning, states and state-backed currencies, which are also self-sustaining memeplexes with their own rules and incentives). That was the stated goal of the design, and it works! There's no "off button", otherwise it would be trivial for "statists" to push it. There is no "oops we forgot to add a limit to the paperclip maximizer" or "don't kill the people who try to stop paperclip maximization" button either. People whose livelihoods depend on crypto are highly incentivized to come up with ways (like PoS) to keep it going even if the outcomes of the overall system becomes undesirable. The day it became legitimized by the SEC, it certainly became impossible to stop it. Now you will have Wall Street money fighting you too.

You wanted an unstoppable decentralized cryptopunk future with assassination markets, that's exactly what it does. It wouldn't be unstoppable if individual or collective action (like that of a state or other group of individuals) could stop it, and it was designed knowing that people would try to stop it. All according to keikaku.

(cue the gif from dr strangelove of general turgidson's horrified realization of what he's just said)

> it's actually impossible to start new networks from scratch based on proof-of-stake due to the "nothing at stake" problem. If attackers don't care if their stake gets burned (because the value is low)

most new PoS chains do just fine because validators are owned by VCs that want to grow the network. whether they can successfully decentralize is an open question but the bootstrap problem is solved

> the bootstrap problem is solved

it's solved if your only holders consist of trusted parties, which is a trivial (but unrealistic) solution to a lot of crypto problems ;)

it seems entirely certain that cryptos that are not VC backed will launch in the future. It happens almost daily actually.

PoS investors burn their reputation if they double spend on any network they invest in.

this is a stronger guarantor of safety than PoW for every chain aside from ones that dominate their computing platform (BTC/ETH). anon miners will eat a naive network and move on.

yes, requiring external trust to exist between parties is a very effective strategy, I agree, but it's not a solution to the general problem where trustlessness is required.

And if you are requiring trust between parties... you might as well simply have an auditable database instead of requiring staking or mining at all.

(requiring oracular knowledge (external to the system) that other VCs have a financial stake is still a trust-based system... and they could still sell that to someone who doesn't, and fuck you over, and your knowledge that they won't do that for social or contractual reasons is still oracular knowledge. You haven't solved anything, you've just created an oracle that everyone agrees to trust, which is a trivial solution to almost all challenges within crypto.)

databases don't have the self-organizing infrastructure to eventually decentralize. vc tokens might.

"it's actually impossible to start new networks from scratch based on proof-of-stake" is simply incorrect. almost all crypto networks will start this way.

Agreed. This is obvious both empirically and analytically.

As BTC price goes up, the network starts to yield larger mining profits in total (in fiat terms) and this creates an incentive for people to launch new miners to absorb the surplus profits... The increased competition then causes an increase in mining difficulty which reduces the profit margins back to what they were before the increase but now there are more miners in the network with more total hash power which are consuming more electricity in total. The equilibrium point is the point where most miners are teetering on the edge of unprofitability.

One interesting thing to note, however, is that while increases in the price create an incentive for more miners to join the network, the Bitcoin network itself does not require so many miners to exist to support itself at those higher price levels. If you assume a hypothetical scenario were all governments around the world required a license to mine Bitcoin and limited the number of machines allowed in total, then price could keep going up without an increase in the total number of miners. This would be very convenient for existing miners as they would be able to earn more profits without having to buy more machines as the price increases.

The fact that this obviously negative-sum game has been profitable for miners in the long run is proof that our current monetary system is broken. If our monetary system was efficient and becoming more so, it would be unprofitable to run so many miners. Bitcoin profits from the inefficiency of monetary system whilst revealing to us the fiat system's negative-sum nature. As dumb and unethical as this all seems, it is the smartest investment to make in the context of our current system.

I suspect that Bitcoin saves more electricity than it wastes by turning the world economy into a giant squidgame... It attracts capital towards it instead of towards other economic activities which would end up consuming more electricity... Those other activities would certainly have made better use of the electricity (since, unlike BTC mining, they would have provided tangible value to people in return) but Bitcoin doesn't care about people's well being. It profits from the destruction of opportunities and the concentration of wealth. Poor people consume less electricity. Bitcoin saves electricity by ensuring that poor people stay poor by soaking up all the excess capital so that it doesn't get invested in real economic activities which would bring people out of poverty.

> world governments would rightly shut it all down before the majority of our electric output went to mining BTC

They can regulate it otherwise, e.g. demand that miners use their own source of green electricity, and make other type of mining illegal, or demand that less than 1% of the total electricity production of the country is used in mining, outlaw the rest and fine illegal operations, etc. It doesn't have to be completely banned everywhere.

> BTC is guaranteed to fail at some point

You'd have to define what "fail" means here, especially given that many proponents of BTC argue in its favor for different (and often contradictory) purposes. What's the failing scenario you're talking about, once mining takes all the electricity it can (be it 100% of production or whatever regulations still allow it to take)?

> by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin.

You have to factor in the amortized cost of specialized hardware over it's lifetime, and a few other factors. But it's a good approximation.

BTC tried to solve this with diminishing rewards for mining. The problem is BTC's market cap doubled far faster than the rewards halved.

But I think you are wrong at the high end. If BTC was worth 100x tomorrow, the current miners cannot just push a button and draw 100x the power. I think governments would be far more likely to ban new rigs, but may let grandfathered rigs continue to operate. If that's the case, I can see miners becoming quite profitable.

> I think governments would be far more likely to ban new rigs, but may let grandfathered rigs continue to operate.

This seems irrational given the ever increasing efficiency of hashes per energy.

It's only irrational if you think the government cares about maximizing hashing efficiency. It's perfectly rational if you recognize that in most (western) legal systems there is a difference in the ease and acceptability of stopping someone from taking a new action and the allowing them to continue taking that same action. This is due to various legal underpinnings, but also makes perfect sense in that allowing someone to invest millions and then pulling the rug out from under them is very different from telling them they were never allowed to invest those millions in the first place.

Building codes are an excellent example of this. It's perfectly legal to maintain, sell, lease, etc. a building that would be illegal to build today.

I agree 100%, though I am interested in chains which are doing "Proof of useful work". I haven't looked into it too much, but I believe Banano participants run folding@home (https://en.wikipedia.org/wiki/Folding%40home) to secure their network.

This seems like it could be an alternative to classical proof of work that at the very least, isn't actively harmful to society. But I'm unsure if it offers the same security guarantees.

> Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin.

I'd amend that to read: Proof-of-work requires, by design, that the money wasted on electricity and electro waste is proportional to the total market cap of Bitcoin times the mining reward (currently 6.25 BTC per block).

This is beautifully put, I'm saving it in my notes to remind myself of it later. The point about market cap was especially striking.

What I don't understand though is once all bitcoins have been mined, doesn't that mean that transaction fees are then also required to be proportional to the bitcoin market cap and electricity usage value?

No, fee costs are based on supply and demand of block space. It's simple, miners just scoop the highest paying transactions into the next block. If there's congestion, people pay higher fees to expedite their transactions.

Yes but once there's no longer the financial insentive of bitcoin rewarded for mining, fees need to have the same insentive relative to bitcoin market cap? Otherwise double spending becomes the most profitable for miners surely.

Yes, total fees need to be high enough to secure the network.

Electricity generation could be different and better in the future.

To emphasize, it does not matter. Let's say we have some great new technology that enables us to get electricity for a tenth of what it currently costs. It just means miners will be willing to use 10 times more electricity to mine for a given amount of reward (and, for the umpteenth time, block rewards, transaction fee, whatever, doesn't matter - the BTC they get in return for being the first to mine a block).

Indeed, this is the whole idea behind specialized mining rigs. These rigs essentially lower the cost of electricity for a given hash rate.

But assuming other miners have access to the same equipment, in the competitive system that is proof-of-work they will just bump up their usage of electricity until the point that, again, it no longer makes sense for them to do so because they'd be spending more than the rewards that they get in return.

The use of mining for hashes in general seems such a completely arbitrary thing to select for Proof of Work. It may have suited the academic cryptography/CS mindset back when the whole thing was an intellectual exercise on whitepaper, but now, as a victim of its own success, it is a threatening technical debit burden.

My working theory is that a system could be devised that re-defines Proof of Work to align with some demonstratable real world physical benefit like planting a tree or recycling waste or placing a renewable kwh on the grid. Yes there are tremendous process invention challenges on a Manhattan Project level scale but the stakes seem able to justify a Moonshot approach.

Mining under existing Proof of Work conditions is a weird arbitrage at scale of physical compute cycle cost/performance and raw electrical overheads measured against the perceived speculative value of the minting benefits returned.

Even reducing mining to a barebones fintech set of transactions would make sense at this point. A single Cash (as proof-of-work) account where each miner adjusts their relative position in real time in response to market conditions. An opensource code as law snapshot Lottery system that mimics the existing minting benefit and proportional voting rights distributions on chain. Yes it interrupts the purely distributed zero trust consensus mechanism, but, there are other compatible mitigation and trust designs, and, as an overall step change, the ability to grow without consuming nation state level power resources in this day and age is well worth the tradeoff -- if not absolutely mandatory.

Also on HN the other day [1]

frankenst1 reply to mondoveneziano :

START QUOTE > For Bitcoin, the most popular one, on the order of 100.000.000.000.000.000.000 of hashes get calculated to mine a single block, multiple trillion per second. Within ten minutes, only a single one of those 100.000.000.000.000.000.000 hashes is actually used, depending entirely on luck. The rest are thrown away entirely. They do not form part of the final hash or anything else, the energy spent on them is lost.

You need to dig through 1 million gram of ore to find 1 gram of gold, depending entirely on luck. The rest is thrown away entirely. The energy is of course not lost but was necessary to find something of value (which also creates a lower price boundary). In contrast to gold, Bitcoin mining ensures provably fair probability of winning, has a tremendously lower barrier of entry, provides incentives to be run on renewable energy and produces far less waste. END QUOTE

[1] https://news.ycombinator.com/item?id=31376647

The value of bitcoin needs to increase more than x2 indefinitely for the power consumption to increase.

> Proof-of-work requires, by design, that the electricity value expended on mining is proportional to the total market cap of Bitcoin.

This premise isn't true. Funny you have this big confident argument when you're wrong about the basics.

The electricity expended is proportional to the value of mined bitcoins plus transaction fees. Note that the mining rate goes down over time, approaching zero.

You’re saying that as the price of Bitcoin has risen, the price of electricity has risen? I think something’s missing from your argument.

I think he meant that if the price of 1 bitcoin rise, the energy cost of mining 1 bitcoin should also rise in proportion

After the each “halving” you should expect the energy expenditure would also halve by that same logic. The halving would could continue until the energy requirement is minor!

Haha, get it, minor?

This is probably part of the appeal of BTC as a value store. Gold typically sits in a vault, it’s hard and energy intensive to mine.

BTC marginally improves upon gold as a value store, it’s trade able as a hard asset across national borders in real time.

Time will tell if BTC is as durable as gold.

Reply on news.ycombinator.com