Bitcoin, Energy, And The Future Of Money
medium.com
medium.com
The desire for currency to have "intrinsic value" appears to be driven by the fear of a sudden collapse of currencies that exist only as a system of accounting (like the US dollar). Or the desire to have a store of value that never suffers deflation or inflation. We all want that, but most of us acknowledge that it is no more possible than a utopia.
You can't freeze the economic system. Whether you back your money with energy, cryptography, flowers, or seashells, you can never control all the other factors, and thus your currency will fluctate.
> The systems developed by cryptography researchers invoke a central issuer of currency in their security definitions, which is useless for Bitcoin.
The first part of that sentence is not completely true, because Bitcoin itself was (presumably) developed by cryptography researchers.
(Obviously, it would be meaningless if people could just move money from wallet to the other.)
First of all, that is not a good definition, since it rules out the possibility of a unit of currency "coming back" to the party that spent it (imagine if I gave you a dollar, and then you were not allowed to give it back to me). It also fails to rule out the possibility of dividing one 1BTC unit into two 0.75BTC units, which is another form of double spending.
There are deeper problems too. You say that if you "can" do it, it is a security problem, but there is no restriction on the attacker's computing power. What if the attacker forges ECDSA signatures? What if the attacker finds SHA256 collisions? What if the attacker computes a longer block chain, which is a polynomial time problem, reversing his transaction and getting the chance to spend the money again?
The reason I said the security notion is vague is that there is no precise definition that Bitcoin actually achieves. If someone can develop one, and if it actually captures a meaningful notion of security, great! So far, no such definition exists.
"The first part of that sentence is not completely true, because Bitcoin itself was (presumably) developed by cryptography researchers."
I seriously doubt that, given that the original Bitcoin paper made absolutely no mention of the decades of previous work on digital cash that preceded it, nor any of the work on secure multiparty computation, nor did the paper give a security definition, nor did the author seem to be bothered by the fact that there is a polynomial-time attack on the system. It reads like a paper written by cryptography amateurs or enthusiasts, the sort of people who make most of the posts in sci.crypt and comp.compression. In a pedantic sense, Bitcoin is the product of research, but if the term "researcher" means anything beyond "human being," I think it is fair to draw the line at being familiar with and building on previous work.
[Edit: Of course, that definition of "researcher" is not quite what we want, since it does not account for people who open entirely new fields. That is a rare edge case, however, and the issue was whether or not Satoshi was a researcher in an already established and well-developed field.]
"Perhaps this is so obvious that people consider it not worth explicitly stating"
If the security definition is not explicitly stated, it is very hard to analyze whether or not the system is secure. How do we even talk about security without defining our terms? There is a polynomial-time algorithm for "reversing" a Bitcoin transaction and spending your money again that is detailed in the original Bitcoin paper -- does that mean that Bitcoin is insecure (I would say yes, but it seems that most Bitcoin proponents say no)? Without a definition of security, it is hard to say.
There is also the matter of determining whether or not Bitcoin's security meets my needs. As a simple example, consider the ElGamal encryption system. It is secure -- but only against chosen plaintext attacks. If you need a system that is secure against adaptive chosen ciphertext attacks, ElGamal is probably the wrong choice (you need Cramer-Shoup). With Bitcoin, however, there is no clear security definition, so how does one even decide if Bitcoin satisfies their security needs?
I think that is a precise definition of "double spend" in bitcoin. Do you agree with that definition?
As for all the other stuff you're talking about re: polynomial algorithms, collusion, etc.... I am just puzzled that you are bringing all that stuff up. It is widely known that bitcoin is resilient to any attempt to double-spend unless there is a coordinated attack by greater than 50% of the total hashing power. And double-spend is the _only_ attack at the distributed algorithm level.
Can you comment on the above paragraph? I mean, were you not aware of the 50% issue, or were you aware of it but have a problem with it? And do you agree that double-spend is the only kind of attack at the algorithmic level?
Of course, there is always DDoSing the peers at the protocol layer, and stuff like that, but I think we can agree that that is not what we are interested in discussing.
And then there is the crypto level, which you bring up by mentioning ElGamal. Bitcoin uses ECDSA, but only for signing; nothing is encrypted. So one can talk about whether ECDSA is breakable, but I'm not losing any sleep over it, personally.
First of all, it is hard to even analyze "total hashing power." There are far too many slippery details for any formal analysis -- you need to know how the hash function is actually implemented, what sort of hardware, etc.
That being said, how do you know there are no other double spending attacks? Where is the proof of that?
"double-spend is the _only_ attack at the distributed algorithm level."
Prove it.
"then there is the crypto level"
There is no such separation. Bitcoin is not a signature system, nor is it a hash function. The fact that ECDSA and SHA256 are used in Bitcoin does not imply that Bitcoin is secure. All that you can really say is that Bitcoin does not meet its security goals (whatever those are) if ECDSA or SHA256 are not secure.
I don't understand why you're bringing that up. You don't need any of that. Anyway, the total hashing power is currently 79,854.53 gigahashes per second (where one "hash" is actually a double SHA-256).
> That being said, how do you know there are no other double spending attacks? Where is the proof of that?
Any double spend in a block will be immediately rejected by all nodes. So the only question is, "Which blocks are actually valid blocks, and which blocks are orphans?" Blocks that form part of the longest extant blockchain are considered valid; all others are orphans. So the only question at that point is, "Which is actually the longest extant blockchain?" If you have massive hashing power, you can create an alternate blockchain and release it, allowing double-spends of transactions in all the blocks back to the point that your alternate blockchain forks from what was previously considered to be the longest blockchain. Mathematically, you need > 50% of the total hashing power to do this. The proof of that is in the paper. (Though it's worth acknowledging that there are a lot of details that are _not_ in the paper, and you have to talk to bitcoin devs or read the wiki or the source or the forums to figure them out.)
> Prove it.
Preventing double spends (and spending of money that doesn't exist - semantically also a double spend) is the sole purpose of bitcoin's distributed algorithm. All the rest is just verifying that transactions are signed properly.
> The fact that ECDSA and SHA256 are used in Bitcoin does not imply that Bitcoin is secure.
True, but there are other things that imply that Bitcoin is secure, and my only point here is that the use of ECDSA and SHA256 are nothing to worry about.
There certainly can be and have been security flaws in the code, but they're not related to the actual fundamentals. That's why they've always been easy to fix and why Bitcoin has never been "hacked."
Edit: changed one to six.
In other words, under your definition, what does it mean to be secure against double spending?
I responded to another comment from you elsewhere, and I hate to confuse things by having one conversation in two places, but this is a really good question, and there is a straightforward answer.
To make a longer blockchain, you have to have enough computing power to "catch up" to all the other people mining bitcoins, and then mine faster than they are. So, you have to have > 50% of the total hashing power.
If you look into the mining "business," I think you will become convinced that only a nation-state or maybe Google could perform this kind of attack right now, and once ASIC mining technology is widely diffused, even that would be difficult.
The "6 confirmations" thing the grandparent said is kind of arbitrary, but that is the number of blocks that bitcoin devs widely agree is enough to wait before you can be "effectively certain" that your transaction is in the longest block chain, and always will be. So that's how long you wait before handing over the keys when somebody purchases your mega-yacht with bitcoin.
By "kind of arbitrary" I mean, in the context grandparent used it. If you do the math (as is given somewhere near the end of the bitcoin white paper), that is where you start approaching a 0% chance of the transaction not ending up in the longest chain.
The "prevention of double-spending" is a means, not an end. You have to ask yourself why you don't want "double-spending".
This seems sensible on it's face, but you're ignoring the effect that speculators have on the market. If consensus is that a commodity will increase 10% year over year then investors will pile on, driving up the price until the expected return is 0%.
When the expected return is 0% (like it is for buying a bushel of corn on a random Tuesday and selling it a month later) then your currency is no longer deflating.
In other words, expected economic growth will be priced in to Bitcoin.
Of course, when the speculators turn out to be wrong about economic growth, you will see the price move. But it won't be deflationary just because the economy is growing.
Yeah, that's kind of the point. When economic growth is "priced in to" a currency that is tied to a fixed commodity, it results in value fluctuations of the currency.
A major issue is that we don't have anything that's usefully dense in energy content to use as money (except nuclear fuels, I suppose, but I have a hard time seeing them in circulation with the current fear of nuclear power...).
> A major issue is that we don't have anything that's usefully dense in energy content to use as money.
I don't really understand this comment. If you're going to design a commodity backed currency, the energy density of the commodity is irrelevant. You don't have to swap uranium bars at the store in order to use an energy commodity as currency. You issue contracts representing your ownership of a specific amount of the commodity in a warehouse somewhere.
Sound familiar? Commodity backed currency is not new. Swapping out the commodity just ties you to another system of supply & demand.
In fact, for a developed country like the US the anticipated trend in energy use is in the opposite direction to the anticipated trend in real GDP. http://www.eia.gov/forecasts/aeo/MT_energydemand.cfm
An energy-based theory of value is more wrong than the labour theory was in the time of Smith and Marx, and similarly it's increasingly wrong. This forum is full of people who create economic value with computer technology that decreases the demands on physical infrastructure.
Hm. I certainly wasn't suggesting that anything be pegged to the kJ, or whatever.
I do think that someone with twice the available energy is, to a first approximation, twice as well off, and that reducing the amount of energy it takes to do whatever you want to do also makes you better off, since both of these make you able to do more of the things you want to do (or have more of the things you want to have).
It's true that there is an enormous amount of things we can do now that fall below a cost floor such that we can have however much we want, but in general I'm far more concerned with lowering costs and increasing options than with figuring out what the "value" of something is. Is a Kindle download as "valuable" as a physical book? I'm not sure that that has any objective answer. Is the cost of production lower (either in pounds or dollars or energy)? Enormously so. And that makes us all richer in comparison.
So if we're not pegging money to a kJ (or whatever), then we're no longer talking about energy backed money.
> I do think that someone with twice the available energy is, to a first approximation, twice as well off, and that reducing the amount of energy it takes to do whatever you want to do also makes you better off, since both of these make you able to do more of the things you want to do (or have more of the things you want to have).
Yes, but we're talking about currency here. I get what you're saying: that the increases in efficiency reduce the "value" of your energy holdings, but it also increases the amount you can accomplish with the same energy. That is not a zero sum in a broader economic sense though, because the cost of different goods vary disproportionately with the cost of energy.
Once again, you're stuck with a currency with a supply that fluctuates out of sync with other major economic factors. You'd be better off backing your currency with an index of commodities from multiple sectors, but then you're making currencies in to something they don't need to be. As I said in my original post, if you want to buy assets of a specific kind, go buy them!
Currency is a form of short term accounting, and that's a good thing. Long term, you want currency at work in the form of exchange buying and selling things. You really only want people holding as much currency as they need to operate a means of production. Stagnant currency is stagnant means. It's the very definition of an economic slow down.
> It's true that there is an enormous amount of things we can do now that fall below a cost floor such that we can have however much we want, but in general I'm far more concerned with lowering costs and increasing options than with figuring out what the "value" of something is. Is a Kindle download as "valuable" as a physical book? I'm not sure that that has any objective answer. Is the cost of production lower (either in pounds or dollars or energy)? Enormously so. And that makes us all richer in comparison.
That is efficiency. It's a significant driver of economies. Everyone wants that. What's unclear is how a move from the USD to something commodity backed would benefit efficiency.
> So if we're not pegging money to a kJ (or whatever), then we're no longer talking about energy backed money.
You appear to be talking about pegging the USD to a given amount of energy. I think that's a bad idea, for the same reason as pegging the USD to a certain amount of gold was when the US didn't have enough gold to actually redeem it. Using energy as money is not about a fixed exchange rate from some non-energy system. I guess the problem here is that "based" and "backed" are being used interchangeably, and they aren't the same thing.
Instead, I think it would be an interesting experiment to price everything else in energy.
> I get what you're saying: that the increases in efficiency reduce the "value" of your energy holdings
Well, increase the "value". Efficiency increases are personally deflationary if we're using energy as a currency, but without the macro effect of reducing the monetary supply or velocity, I would think.
The fed has only half the toolsets nowadays - it can create and pump liquidity, but cannot extract excess amount and cannot direct where the liquidity will go.
For example if bitcoin is so volatile that every time people earn one they immediately want to buy X with it to make sure they hold their value, you might as well skip the bitcoin step and directly use X as a currency, especially since a bitcoin transaction isn't free. X could be stocks, bonds, gold, dollars, etc.
There is a separate argument surrounding currency debt. That is, current policy doesn't strictly regulate the money supply in a way that tracks economic growth. They use debt as a means to "create" growth. I won't pretend that this is some sort of foregone conclusion, but the whole debacle involving Reinhart and Rogoff and their Excel error dealt a serious blow to those who object to the theory that debt drives growth. Again, I'm not saying the case is closed, but right now, the debt-driven-growth folks are winning the debate.
When you fix your currency to a commodity or other non-controllable factor, you give up control of your money supply, which causes volatility of its own.
As an example, in the 17th century, the Netherlands experienced a large bubble in the value of... tulips.
Indeed! And before long, people will want to lend and borrow BTC once usage expands beyond the current early-adopter "hard money" crowd. That's just human nature and is OK on a small scale. But then BTC IOUs may start to circulate. Then some new-fangled "BTC Goldman-Madoff" figures out ingenious ways to "make your BTC work for you, let us do the IOU magic of fractional reserves for you. Just don't all attempt to cash in your chips at the same time, which you won't given the incredible growth in IOU 'returns' (not hard BTC but hey, by now one is as good as the other in the minds of most) we can deliver for a few decades".
Bam, here we go again?
While this is true to a limited extend, do we really /want/ it to be true? Using energy as the measurement of an economy's size would naturally lead to economies across the globe investing in raw energy output, rather than other equally (or more) important areas of technological progress like making future infrastructure more energy-efficient. This wouldn't be so bad if it weren't for the fact that the sources of energy we most commonly turn to today, fossil fuels, are inherently limited and have huge negative externalities. We need to come up with a standard that convinces markets to grow the standard of living in a way that doesn't require increasing the energy supply. If we don't, we risk exacerbating the global environmental and resource crises.
The first part is probably true, but surely if your currency is based on energy, you will probably spend a hell of a lot of effort on squeezing the most out of each dollar/joule?
Currency/money is a data-structure for a kind of cooperation algorithm. That is the way to understand it in an informational world. Note that money is not really about being a 'medium of exchange', a 'unit of account', 'store of value' -- those are tautological ways to look at it, they are its particular forms of activity, the real purpose is cooperation (more) generally.
The radical thought is that we do not really want currencies. We want something more information-rich -- something made for our current information tech, rather than originated from cowry shells or whatever. Hardly anyone is thinking in this way though, disappointingly. Yet it is myopia: if you look around non-currency systems are beginning to evolve. What is reputation, or 'karma', are they currency? Not really. What are things like Flattr and Kickstarter, are they markets? Not really.
Start deliberately thinking beyond currency.
You wouldn't purchase my heating bills from me...
Exactly, technology improvements can and do change the energy used.
It makes sense to me, since in principle the amount of currency in circulation could naturally adjust to the size of the economy. To make that work maybe you'd need demurrage, to prevent utilities from printing up kilowatt vouchers for production in the distant future.
Since a watt is a watt, you could have as many issuers as energy producers, without worrying about exchange rates, though limitations of the electric grid might force them to be somewhat local.
Other factors that affect the pricing of a "Joule": convenience of application (I can't burn coal in my car, I pay a premium for gasoline; gasoline won't get me to the moon, I pay a premium for rocket fuel); ease of refining the mined resource into finished product form; ease of transporting fuels from source to "refinery" to distribution to end user; we'll pay a premium for fuels with a reliable steady supply; outside risks, e.g., political unrest in the Middle East increasing production costs and risk.
The "joule" measure for the value of a fuel is flawed. We already have a great measure for the value of a fuel, it is the US$.
Also, as many point out, the energy required to create a bitcoin has little bearing on its value. Energy costs help determine the mining effort and technology put toward mining bitcoin, but the value of a BTC will fluctuate depending on other factors. The total energy required to manufacture, transport components of, and assemble a house, along with the "inherent" value of those materials, is less than the US$-denominated utility of that house. If I shock a turd with the same energy used to build the house, it doesn't have any value approaching that of the house unless I'm Andy Warhol.
"The energy supply defines how much productive work can be done, while the efficiency with which we use energy defines how much utility can be created from that energy supply (widgets per kilowatt-hour)."
the $ is driven by more factors such as FX, inflation, interest rates and therefore not a clean measure of energy value.
possibly the purest current value of energy today would be in energies futures markets?
If people don't actually own money as joules, but money as "widgets" or "joules of oil"+"joules of hydrogen".., then you can just abstract away the energy stuff.
I'm not impressed by the energy as currency concept. Money should reflect all resources, including energy, human labour etc. Binding currency to something (like gold) doesn't seem to be useful.
1. Hopefully the words "Monetary Policy" and "Fiscal Policy" ring a bell. The ability to control market fluctuations, let alone wield some level of political power as a government, lies heavily within a government's ability to utilize its money supply. Not all regulation/rule is evil - what would a programming language be without regulation/rules? Additionally, the great depression was more than proof at the need for some level of market regulation (John Keynes).
2. Basing money on a fixed supply, IE FUCKING ENERGY (since it can't be created or destroyed), would imply that there is a fixed amount of wealth. That's called Mercantilism and it was disbanded by Adam Smith... and 1000s of other economists/governments/times/etc
3. Bitcoin and friends decrease the transaction costs of money exchange (government regulation, forex, taxes). Money creates liquidity for the value you've created, and Bitcoin and friends makes that value more liquid -in certain ways- but it does not replace money.
4. Stability. Knowing that there is an entity that will back their currency, filled with fail safes, etc. encourages proliferation of that currency. It's why most investments (forex wies) are done in US dollars. Thing such as bitcoin, do not have that type of security. (FDIC anyone?)
5. Decentralization of currency often leads to destabilization of currency. If anyone can make it (and yes, it is far more likely someone will create a bitcoin instead of forging a US dollar), then the value in and of itself will drop.
6. "But WallStreet and all the finance people are up in a yipyap about it!" That's because it's an investment - many people play the forex (foreign exchange market) because there is much profit to be made from arbitrating currencies.
7. Inflation is needed - Milton Friedman has this at about 4% or so a year to keep up (check on that). Why? Try playing monopoly with 4 players with the set of cash they give you. Then try it with 10. There's not enough money to represent the players, therefore you need more. Inflation isn't always bad.
Though that's kind of crazy, it's far from a possibility. And, given the increasingly-global nature of business, anything like online sales could easily track this to the minute by watching exchange rates. It's definitely not there yet, but it almost seems inevitable given time.
The fact that energy is fundamentally conserved has little to do with the amount of energy available to the economy in usable form.
There's nothing about bitcoin or energy-backed currency that would prevent someone from insuring it.
Hayek published a book in 1977 arguing that a decentralized system of competitive currencies would lead to natural price stability without centralized intervention.
Fees paid could be distributed to citizens, making the whole thing work just like Hansen's fee-and-dividend, but adding an incentive to absorb carbon when it's more efficient to do that than to avoid emissions.
The US has insanely massive oil reserves given the new fracking technology.
All money has only the intrinsic value of your labor. That's it. Gold, Silver, U.S. Dollars, wooden coffee tokens, they all represent a claim on human labor. There is effectively an infinite supply of gold on earth(unfortunately most is believed to be in the molten core and no one is able to process it of seawater yet), the only reason it's reasonably valuable by itself is the labor input required to harvest it.
At some point, maybe robots will fix themselves and human labor will no longer be required, in which case the whole basis for a money system is going to disappear. However, that hasn't happened, and doesn't look to happen in the near future.
People should keep this in mind when thinking about money. Fluctuations in value in a currency leads to the arbitrage of labor. Currency itself has no value, it only represents value. Backing it with energy or some other resource, still simply links it to how much labor that energy source requires.
So far, there has been no perfect currency, and I doubt there ever will be. Bitcoin may solve some problems, but it doesn't solve all of them, and it's not clear it solves more problems than many other solutions. While some people may get rich from the arbitrage, not everyone can sit around trading 1's and 0's. Someone has to do the labor.
I'll reference Adam Smith again, who's theory was that Late Stage Capitalism would end in the destruction of the Capitalist State when the profits of finance were so high in proportion to the profits of labor, labor ceased. Right now the trading of bitcoins is much more profitable than selling stuff for bitcoins, so bitcoin has pretty much skipped the labor input part and gone straight to finance, so I still don't think it's addressed the main problem with currency, which is that people still think of currency as the value, not the labor it represents.
The other point was that it is not important for a currency to be a long term store of value. It's only needed to facilitate transactions.