Do not deny the electrical problem of Bitcoin

Do not deny the electrical problem of Bitcoin

Andreas Antonopoulos, a leading figure in the cryptocurrency community, respected because of the excellent book he wrote about Bitcoin, recently spoke about the electrical spending of the Bitcoin network . The text of his intervention is interesting for several reasons because it takes the classic arguments of the Bitcoin defenders, because it introduces what seems new arguments, and also because it forgets to answer certain objections.

A less partisan point of view concerning the reality of the nature of Bitcoin mining deserves to be expressed. I will try to give it by proposing a point-by-point answer to the adroit and vehement speech of Andreas Antonopoulos.

– – – – –

“Mining is one of those few industrial activities that do not depend on a territory at all. You need cheap electricity, the place where you find it does not matter and you can choose the place of installation only on this criterion. This means that mining can be a form of arbitrage in the electricity market for the best performing sources. ”

– – – – –

What does it mean ? That if somewhere the electricity is cheaper because overproduced bitcoins miners will buy it and that it will facilitate the establishment or maintenance of plants where it would be difficult without mining to consider them. For miners where the mining takes place is not very important and therefore, if an industry or firm (eg Google data farms, or an aluminum plant that wants to pay electricity as cheaply as possible and is ready to move) is interested in this cheap energy, so it will have to do without it or pay more, since it will compete with bitcoin farms. This “arbitration” of which Antonopoulos speaks does not seem positive to me. Use the word arbitration is clever, but a less partisan way of speaking would be to say, “bitcoin farms are taking the cheapest sources of electricity, picking them up for themselves and driving up the price of that electricity that might not have been Dear. ”

– – – – –

“In many cases the reason electric is cheap is because the production is not in line with local demand. In regions where the electricity distribution network is underdeveloped, the solution is often to invest in a power station not just to cover current demand, but in anticipation of the demand of the next fifteen years […]. When you invest in a 50-megawatt [hydroelectric, geothermal, …] plant in a place where today’s demand is only 15, you can not turn it off or slow it down. The energy produced will be wasted unless you can convert it into an alternative value pool allowing you to pay back the cost of your installation more quickly. Bitcoin can be considered as a global environmental subsidy since it makes it possible to absorb investments in alternative energies more quickly […]. By balancing the balance of supply and demand in private areas of distribution network, the decentralization of Bitcoin promotes the decentralization of energy production. ”

– – – – –

If you build a 50-megawatt plant when you only need 15, it means you made a mistake. This should not happen very often and certainly not to the point of meeting the big needs of bitcoin miners. The way in which the problem is presented forgets that other users are ready to move and that they are penalized because the competition of minors bitcoins raises the price of electricity that they are therefore obliged to pay the above the price they would have had access to without the bitcoin farms. To say that we avoid waste is false for two reasons: (a) the overproduced electricity could have interested another industry;

The usefulness of mining is the consolidation of the blockchain – the increase of its resistance to attacks 51% – but this resistance is already very good, and therefore any expense to increase it is a form of waste. If, moreover, it encourages the construction of plants that otherwise would not have been, it is an economic and ecological absurdity indefensible.

Saying that ” Bitcoin can be considered a global environmental subsidy because it allows to absorb more quickly investments in alternative energies Is making fun of the world. From the moment the blockchain is robust enough to withstand attacks 51%, burning electricity by computing SHA256 is equivalent to throwing electricity (so coal resources, or others). Since this growing demand for miners’ electricity is due to a pure mechanism of competition and has no real utility – even for Bitcoin – it can only be considered as a global environmental perversion. The demand of the miners encourages the creation of over-sized power plants, or even the construction of new plants solely for the purpose of calculating SHA256, calculations that are immediately deleted. This is an environmental madness and an insult to all those who really need electricity and do not have any to heat themselves,

Andrea Antonopoulos’s reasoning is analogous to the following reasoning: “If a baker produces ten times more bread than he is able to sell to his usual customers, and sells his surplus to serve as a bad fertilizer. or fuel for heating, it will more quickly amortize ovens too large, and so it is a subsidy to the production of cheaper breads. No ! Loafed breads will not help to dampen the ovens, the waste of having built ovens too large will remain a mistake and will not help in the production of cheaper breads, and the flour used for the bread produced will not even not refunded!

The mining of bitcoins does not promote the decentralization of energy production, it favors the production of energy where it is not needed. Ultimately, it increases the cost of electricity in the areas concerned, for which it is therefore an economic nuisance in addition to being an ecological nuisance for the entire planet.

– – – – –

“Bitcoin is easy to criticize for its power consumption because it is obvious and visible [unlike the hidden, traditional banking system]. ”

– – – – –

Let us first note that Andreas Antonopoulos does not dispute the high power consumption; he knows that it is really important and is careful not to give the slightest figure that even minimizing as much as possible the result would remain frightening, as shown by the various possible minimalist calculations.

Calculations based on Marc Bevand’s known expenditure on ASIC circuits indicated a consumption of around 4 TWh per year when the network hashrate was 3.25 10 18 hash / s ( here ). The hashrate Bitcoin network is now (late December 2017) 4.5 times greater. The electrical expenditure is therefore 4.5 times greater, except that there has been a progress in the energy efficiency of the ASICs used to undermine that can be estimated at 1.3 over the short period concerned. The electrical expenditure of the network is therefore in order of magnitude of 16 TWh according to the calculations De Marc Bevand ( a defender of the idea that Bitcoin does not spend much ) updated by the factor 4.5 / 1.3 = 3.5.

Another minimalist calculation is possible: we make the wildly optimistic assumption that all the miners use the most efficient equipment available today (the Antminer S9 here ), we use the data on the power in Hash per Watt of this best device , and we arrive at 12 TWh of annual expenditure for the Bitcoin network.

These minimum figures correspond to about two nuclear power plants. Most worrying is that the price of Bitcoin has increased significantly recently, it is certain that competition between minors will lead in less than a year to at least three times the energy expenditure of the network. We will then be with these optimistic assessments at about 36 or 40 TWh or more than 4 nuclear power plants.

By comparison, all data centers of Google (making a real service to hundreds of millions of people every day) are globally only 700 MW, so 6 TWh per year ( here ). We understand why Antonopoulos prefers not to mention any numbers, and not even those that are minimal yet no doubt!

The argument of the traditional banking system is easy, especially since those who use it do not detail reasoning or calculation. In reality, Bitcoin also has significant off-line costs, that is, in addition to mining. There are, for example, dozens of exchange platforms, numerous information sites, a whole world of startups dedicated to cryptographic currencies that are put forward to prove that this technological innovation is something of economic importance that it do not hinder. This world spends energy on its offices, its machines, this world pays wages, occupies buildings that are heated, organizes meetings and therefore travel, etc. This cost around cryptographic currencies is not counted in mining (which only concerns the calculation of SHA256 doubles per billion billions). It is probably weaker today than the banking system, but if the cryptographic currencies were to replace the fiduciary currencies, would that cost be less important than that of the traditional banking system? It is not certain. It is rather likely that it would be equivalent. The problem is that at this cost out-of-the-chain, it is necessary to add in the case of Bitcoin that of securing the blockchain – the cost of the mining – which day by day becomes more important. It is equivalent in the case of fiduciary currencies only the cost of issues and replacement of used notes which is in the case of the dollar less than one billion dollars per year, here ).

– – – – –

“However, you must keep in mind that Bitcoin’s current level of security could enable it to cope with a global attack by a coalition of states. This level of security is required because it guarantees a fundamental property: resistance to censorship. ”

– – – – –

It seems naive to believe that the political powers will oppose cryptographic currencies by implementing 51% attacks that have now become very expensive (several billion dollars). The political power, if it is embarrassed, or if it considers that the effect on the flow of electricity is troublesome (which may occur) will ban bitcoin mining plants and close them, which will is already the case in the Sichuan province of China (see here ). It may also completely prohibit the use of bitcoins and other cryptographic currencies as is the case in Morocco. Political power will have weapons other than attacks in good standing network. A global ban on cryptographic currencies will not completely prevent them from existing, but by making their use a clandestine activity and outlaw, by sending in prison its recalcitrant actors who will be caught, it would make escape most of all the activity that is developing today around them. It is to understand nothing of the truly decisive balance of power on earth that to think that 51% resistance to attack is the heart of the problem and therefore that mining is so important. It is to understand nothing of the way in which our societies operate that to think that clandestine cryptocurrencies would continue to flourish.

Antonopoulos’ way of expressing another danger. By appearing to believe that the “level of security” produced by the mining protects Bitcoin from any attack including by a coalition of states, it propagates the misconception that Bitcoin is definitely resistant to any attack because of the existence of mining farms and their 10 ^ 19 calculations of double SHA256 per second. It is true that the threat on the Bitcoin network by the attacks 51% is actually made expensive by the amount of calculation put in the blockchain by the minors, and that the risk of this attack of the network is today minimal … without however to be inconceivable. But what is totally misleading in the way Antonopoulos is expressed is that he forgets that other attacks are possible against which the mining does not produce any protection. If someone succeeds in inverting the SHA256 function, then he will be able to win all new bitcoins issued making all mining farms unnecessary and giving himself the power to stop all transactions (by systematically adding empty pages to each new page) . Inverting the SHA256 is a mathematical problem that no one has proven is impossible to solve. What happened recently for the SHA1 shows to those who would doubt that in mathematical cryptography this kind of thing happens: see Inverting the SHA256 is a mathematical problem that no one has proven is impossible to solve. What happened recently for the SHA1 shows to those who would doubt that in mathematical cryptography this kind of thing happens: see Inverting the SHA256 is a mathematical problem that no one has proven is impossible to solve. What happened recently for the SHA1 shows to those who would doubt that in mathematical cryptography this kind of thing happens: see here or here .

Image result for bitcoin electricity

Another possible attack is the double-key signature system. He, too, rests only on a supposed difficulty of a mathematical problem. Adi Shamir (the S of RSA) thinks moreover that this protocol of signature is today in danger: “Elliptic curves will fall out of favor; There’s a very strange current situation with the NSA moving away from it with no explanation. See here .

Other attacks have also been designed which the mining does not protect. There is for example the ” Eclipse Attack “. They are certainly difficult to implement, and today no one has been able to exploit them, but they are among the threats to the Bitcoin, and it is misleading to suggest that the massive electricity expenditures of mining farms protect it.

– – – – –

“However, if the system was ten times larger, with ten times more users, it would not need as much as ten times more computing power […]. It is a mistake to believe that if Bitcoin becomes widespread, its long-term cost will continue to grow. It’s even the opposite that will happen. As time passes, the reward awarded to the miners decreases. The result is that the cost will eventually reach a plateau. ”

– – – – –

Yes, it’s true, the cost of mining is not related to the capacity of the bitcoin network (today 4 transactions per second) but during Bitcoin, taking into account the corrections made by the halving which bring down the reward awarded to minors. However today two things are certain.

(a) Bitcoin network power in number of transactions is ridiculous compared to what other transaction systems allow (several thousand transactions per second for Paypal or the bank card network) while the cost of mining is already very important . In other words, huge quantities of electricity are being mined for a powerless system that is naively believed to make it resistant to any technical attack or intervention of political power.

(b) The halving will arrive slowly (the next in June 2020, in 2 years and a half, the following in 2024 in 6 and a half years) while the electrical expenditure of the mining increases rapidly. This increase (which can be evaluated by following the network hashrate and the progress of ASICs) is in order of magnitude of a factor greater than 4 for 2017 and 3 per year on average for the 4 previous years. It would therefore be necessary to compensate the only increase in Bitcoin’s electricity expenditure for 2017, wait six years of halving (you have to wait for the halving of 2024). Given the increase in prices, it is foreseeable that we will have in 2018 an increase in the electrical expenditure of bitcoin mining equivalent or greater than that of 2017 (in fact most probably more), which will at the very least a multiplication by 16 in two years demanding to wait 14 years of halving to be compensated. When the plateau arrives, the expense of mining will become huge.

It is not certain that this plateau happens when we see what happens with commissions: on December 22, 2017, some pages received more than 9 bitcoins of commissions (see ic i); we are not far from canceling the halving 18 months ago that has increased the pay from 25 to 12.5 bitcoins.

It is not possible to wait for the hypothetical plateau of Antonopoulos. I bet that something economic will happen – a crash – or a political one – an intervention that prevents the continued growth of the electrical expense of mining – before the expected plateau is in sight.

– – – – –

“I do not think, moreover, that it is relevant to propose several proof-of-work systems , I think only one is enough. Maybe everything else could work in proof-of-stake with an anchor in a unique proof of work system. We need a proof-of-work system around the world to provide us with a solid energy-based immutability, but we only need one, and maybe that’s Bitcoin killer app. ”

– – – – –

The idea that there is a single currency based on proof of work others relying on it is interesting but it encounters some obstacles. (a) If even one is too troublesome, what I believe, that there are no others does not solve anything; (b) It seems very unlikely that the cryptocurrency community will be able to agree that there is only one. I do not believe in this killer app . I believe instead that without a serious and quick treatment of the problem of electricity Bitcoin because of electricity will be a victim of a killer crash or a killer action of a political nature that will end the anarchist dream of money. digital liquid escaping states.

– – – – –

Conclusion
To deny that there is a problem with the electrical expense of Bitcoin mining can only be done by disguising the reality by hiding many unquestionable points about this cost. It is wrong to defend the idea of ​​crypto-currencies that refuse to see reality. It would seem more reasonable to continue to look for ways to stop using catastrophic evidence of work . Above all, we must find ways to convince or force the main actors of this infernal machine leading to madly burning electricity (minors) to get on with each other to move on.

About the Author
Jean-Paul Delahaye is a mathematician – he has a doctorate in mathematics on the theory of transformations suites – and computer scientist – he is a professor at the University of Lille 1 and researcher at the Research Center in Computer Science, Signal and Automatic Lille CNRS and member of the Algorithmic Nature Group . His current work focuses on computational games, the algorithmic theory of information, the definition of chance and its perception. Since 1992, he has been writing Logic and Calculus (an article of 6 pages each month) in the journal “Pour la science”.

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: