The most powerful supercomputer in the world reaches 143 petaflops, bitcoin network in comparison reaches 80704290.84 Petaflops. Granted that a majority of this computing power is in the form of custom ASIC's, but the figure is really staggering.
And all of the 80704290.84 Petaflops, consume 73.12 TWh to repeatedly calculate SHA256 ! What could be the world's most powerful network does absolutely nothing but crunch hashes billions of times to discard almost all of the results anyway. Sheer waste of computing power.
It's incredible that people on HN, a technical readership, still don't understand why the energetic or computational "waste" of proof-of-work is needed.
There is simply no other known way of implementing a decentralized, censorship-resistant, robust digital currency. Without PoW you lose one of these properties.
Proof-of-stake doesn't work. It isn't robust. Eg. PoS cryptocurrencies can't resolved which chain is correct after a network split. There are many other unsolved problems. That's why Ethereum is years behind schedule in designing and deploying PoS.
Replacing PoW with a useful algorithm (eg. protein folding) loses decentralization (a central trusted authority must verify/sample who performs the work correctly).
Trust-based systems (Stellar, Libra) lose censorship-resistance.
A "dumb" PoW is literally the only practical solution.
If the social benefits of a cryptocurrency are worth the computational power, and if PoW is the only technical solution to implement it, then by definition PoW isn't wasteful.
If the only way to have a decentralized, trustless, robust digital currency is to waste an inordinate amount of energy and increase our global warming and pollution problems, then we'll have to do without that kind of digital currency. It's a matter of priorities.
From a position of power it is easy to do without, because you derive no security or finances from this solution. After all, the status quo is paying you just fine. Why is what you say relevant for most people, who to varying degrees definitely don't have all the things that you apparently take for granted?
Is having food on the table and a roof on their head relevant to "most people"? Because, when we're talking about climate change, that's what is at risk for hundreds of millions of people.
Said the man living one of the most privileged lives possible on this planet. Bitcoin has a very clear use to the oppressed people of the world, and for that exact reason it’s popular among criminals.
It's comparable to what a single hydro dam like the Three Gorges Dam can produce. Don't fall prey to the scary comparisons BECI employs to mislead its readers.
Global warming is critical, but there are other energy wasters much, much bigger than Bitcoin miners.
Besides, as it's been said many times, miners tend to use renewables since they have become cheaper than fossil fuel power plants. A recent paper by Stoll et al. estimated CO2 emissions as being comparable to what a single city like Las Vegas emits: https://mobile.twitter.com/zorinaq/status/113943906857019392...
Right now the share of the world economy running on Bitcoin is negligible, yet its energy consumption is relevant. If that share became consistent, the price of Bitcoin would increase by many orders of magnitude, and mining bitcoin would become more and more competitive in relation to other uses of energy, and would increase in tune - that's why I say inordinate. PoW is just unsustainable by design.
Bitcoin's energy consumption is also by design self balancing - increase the price of electricity, less electricity will be used. With the same benefits for the network.
It's not a "waste" if the result is a decentralized, trustless, robust digital currency.
And the global warming argument is a many times debunked hoax - if anyone has better use of that electricity, they are welcome to use it, which will make mining cryptocurrency unprofitable. Except in the other cases, the benefits go to the single entity that owns the business, whereas with cryptocurrency every participant benefits from the stronger network. Global warming is caused by using fossil fuels to make electricity, mining crypto doesn't require fossil fuels. There are other ways to make electricity, let's focus on them.
> if anyone has better use of that electricity, they are welcome to use it, which will make mining cryptocurrency unprofitable
I'm sorry, but you got this backwards. If bitcoin becomes a relevant part of the world economy, then it HAS to use a relevant part of the world energy, because consuming energy for PoW is the only limiting factor against a 51% attack. So, either Bitcoin is irrelevant - and thus its power usage is a literal waste - or it's relevant, and then its power usage has to become relevant in terms of global warming.
From the point of view of the miner, they have a supply of electricity at a certain price and can decide what to do with it. They might run their TV, computer, heating, industrial machinery, datacenter, or a bitcoin miner. They have to compute what profits can each of the options give them. If running bitcoin miners at that electricity cost and at that profit they can get from selling the bitcoins is more profitable than the others, then it's logical to do that. If on the other hand the price of bitcoin goes down, or the price of electricity goes up, or people start paying more for server hosting, then the decision changes. The fact that miners are now mining, means it's currently most rational to do that, given what society (users, participants) are willing to pay for the services.
If bitcoin is the dominant currency, another option exists:
51% attack, double-spend to get free electricity.
This is useful even when it makes an apparent loss by damaging trust in the currency: If you are, say, the USA president and you’re at war with Iraq, and Iraq uses bitcoin, you can outspend on energy until they surrender.
Unless the whole world uses bitcoin, but then the first few nations individually face the same problem, regardless of who else actually uses bitcoin — gotta keep the USA and China happy at the same time! (The EU isn’t integrated enough to do that sort of thing yet, but is a similar sized group).
Right now, 51% needs the cost of one very large power station from when you start until when you win — large nations, the sort with global ambitions, can spare a lot more. In the previous example, that’s close to all the power Iraq produces, but 5.4% of the USA’s output.
Then there’s the fact that most countries like being in charge of their own currencies as the ability to create or destroy units is a useful economic lever.
> The fact that miners are now mining, means it's currently most rational to do that, given what society (users, participants) are willing to pay for the services.
Or it’s speculation, like so many other things before and probably yet to come.
There is no such thing as speculation by mining - if they wanted to speculate but didn't profit from the mining itself, it would be better for them to speculate by buying - instead of giving dollars to the power company, they'd just give them to other people on the exchange. Basically the only reasons for miners to "speculate" are long term contracts like lease contracts for warehouses. And those may or may not be long term.
And 51% attack has the designed side effect that all other people stop using the currency - they may keep following the previous fork without the double-spend - as shown by ETH and ETC. You'd also need some way of forcing all participant to stay with the now corrupted currency. Otherwise the attacker who performs 51% attack will gain the ownership of a network that immediately becomes completely worthless.
> And 51% attack has the designed side effect that all other people stop using the currency
That’s my point. That is literally the point. That is why it is a bad thing and why no sane nation would ever allow it to become their main currency. It is an attack surface. It is a vulnerability to your economy.
I see no benefit in a trustless currency: If I have no trust in a currency, why should I trust the goods/services?
I see no benefit in a decentralised currency: gold is one (anyone can mine it) and there’s a reason we moved away from it.
I don’t buy that bitcoin is either decentralised (they who control the algorithm steering committee control the currency); trustless (why should I trust irreversible transactions? Why should I trust those who wrote my wallet? Why should I trust those that wrote my mining app?); nor robust (the domain of money is law, not logic, so always subject to government interference; the price is currently highly volatile; and apps always have bugs yet to be discovered).
Trustless currency means exactly the opposite of what you think. It means you don't need to trust anyone, because everything can be proven and verified mathematically without any authoritative figure that you'd have to trust. That's why you can trust the system itself.
I think you have fundamentally misunderstood my point.
Let’s say I buy a widget. How can I trust that the widget will arrive? That it will do what widgets do? That it will not break? None of these are payment issues, but they might call for a refund. How can I get a refund? How can the refund system be resistant to abuse?
The answers we currently have are “the law”. If the law functions, I don’t need a trustless currency.
Once we have enough green energy I'd totally agree with you. But currently, I don't think we can afford adding that much CO² for a currency that isn't widely used anywhere in the world.
Which is negligible compared to the number of credit card transactions in the US (> 300 times as many per month) or bank transfers in Germany alone (50 times as many per month). And that's not even counting other ways or countries.
Disregarding the energy cost, couldn't a "dumb" PoW be more computationally complex than Sha256, such that a relatively larger number of computers could compete in the consensus algorithm, thereby increasing decentralization? To me, Bitcoin doesn't seem that decentralized, considering the ASIC chip manufacturers control the majority mining stake. Intel has even gotten into the game.
More complex PoW, such as a memory hard one requiring several GB of memory to solve efficiently, can lower the efficiency gap between GPUs and ASICs to about one order of magnitude, but not much less. That means that ultimately, only ASICs can mine profitably (and only if they have access to very cheap power).
I am not that old and still I remember Ring networks (before ethernet), they were highly impractical and rudimentary, but they filled a purpose at their time.
Then Ethernet came. Anyone remembers the dumb hubs? when a computer wanted to send a message, the hub actually broadcasted it to all computers in the network... so inefficient, and yet, it served its purpose.
Later, came the network protocols. There was this protocol used for discussion groups/news, remember NNTP? wasteful, because nodes had to download / replicate the full history of posts. But still... served its purpose for its time.
Then came a messaging protocol, SMTP, very useful for sending "electronic mail", very convenient. But people started sending binary data files on it... by converting it to text (UUEncoding anyone?) so wasteful and inefficient. But it is still being used.
So, bitcoin for me is just that early system, that early test that shows a way to do decentralized, censorship-resistant, digital currency. Humanity will find more efficient ways to do it (and maybe they won't be used, as with Email) or at some point in the distant future, the power side of the equation won't be relevant (renewable energy? nuclear? who knows). But for me, that does not change the fact that bitcoin proposition is a stepping stone for society that sooner or later will change the way people transfer value.
I invoke Zawinski's Law of Software Envelopment, and predict that Bitcoin will be replaced by artificially intelligent spam filtering software that can read mail.
> If the social benefits of a cryptocurrency are worth the computational power, and if PoW is the only technical solution to implement it, then by definition PoW isn't wasteful.
But they aren't, and the whole exercise is comparable to trying to repeatedly solve ever-larger NP-complete problems by brute force. Something no technical person in their sane mind would consider a correct course of action.
And just like we use good enough polynomial approximations to get near-best solutions to NP-complete problems, we can do the same for running economy. Trust is not a liability. Trust is what makes economy efficient.
The bitcoin network asics reach exactly 0 petaflops. Why? Because they cannot do any floating point operations. They can literally do just SHA256 hashes, nothing else.
The "worlds most powerful network" simply cannot do anything else than this. It's effectively just an expensive set of electrical heaters.
That's like saying a water bottle factory has only an output of N water bottles per day while the Three Gorges Dam can output N million liters of water a day with the difference that the Three Gorges Dam actually produces something useful, electricity, while doing so.
Just because it's a higher number, it doesn't mean it's equally useful.
It's more like instead of using the Three Gorges Dam to generate power you use the giant flow of water to fill 7 water bottles per second and divert the rest to dry up in a nearby salt flat, citing the renewability of the solution.
So supercomputers produce nothing useful but the bitcoin network does? In the long term I'm not sure we gained anything so far by all the energy spent on bitcoins. Some people got richer, some got poorer but I've yet to see any benefits for society from bitcoin.
So many times this argument comes up and every time I am amazed at how ignorant it seems to me. But it is very hard to discuss...
Can't we all just agree that what is a valuable use of someones watts or petahashes per second is a subjective choice? That people are free to value their resources as they see fit? You personally might not value securing Bitcoin network, but many other people find it valueable and it is perfectly fine in a free country to allow them do what they please if you are similarly allowed to do what you please (no one forces you to buy Bitcoin miners)?
The problem is the impact these things have on the rest of us when we ourselves have no interest in bitcoin.
If bitcoin existed purely inside a virtual world — say, if coin miners were a virtual good that WoW players could buy which made WoW money appear in their inventory according to similar rules but without the actual hard work of computing anything beyond a lightweight O(num_players) random number generator on the server, then it stops being anyone else’s problem.
Are my christmas lights that I put on during winter also everyone else's problem because they waste energy (that I bought with my own money)? Are you going to get into my business there as well and tell me what I can and cannot spend my resources on, because you claim that they have an "impact on you and you have no interest in them"?
I am pretty sure that the best consensus the western world has come up with is that if there are any externalities, they can simply be taxed in like for example a carbon tax. But telling other people what they should and should not find valueable as if you have some sort of ultimate authority on knowing what is valueable, well, I hope you can already see what is wrong with that. Value is subjective. As in, other people have a right to decide for themselves.
> Are my christmas lights that I put on during winter also everyone else's problem because they waste energy (that I bought with my own money)?
Yes but too little to bother about. Transportation and inefficient heat management are the only common ones that matter.
> they can simply be taxed in like for example a carbon tax
On that we agree. I believe car fuel is taxed appropriately in the UK, but in general these externalities are not properly accounted for. If they were, it would be a different matter, but they’re not.
Yes but the bitcoin network computers don't _really_ coorperate (well they do, with a bandwidth of the order of 2MB/15mins), unlike supercomputers, so hard to compare. But yes, lots of waste.
Maybe there are more values in investing into SHA256 hardware other than Bitcoin. That's 80704290.84 petaflops dedicated to break the SHA256 algorithm (finding a duplicate). Finding such duplicate could advance knowledge, mathematics, cyber security, society, etc
And all of the 80704290.84 Petaflops, consume 73.12 TWh to repeatedly calculate SHA256 ! What could be the world's most powerful network does absolutely nothing but crunch hashes billions of times to discard almost all of the results anyway. Sheer waste of computing power.