My issue with it is even more fundamental than this: it is structurally impossible for Proof of Work to become more efficient. If you find a more energy-efficient way to mine, the correct decision is not to reduce your energy usage, but rather to use the same amount of energy to compute more hashes.

The energy-per-transaction ratio is only relevant if the utility of Bitcoin is to compete with Visa to process transactions.

However, if the ultimate utility of Bitcoin is to store trillions of dollars of value, transmit it into the future, across borders, quickly, pseudonymously, and without capital controls from coercive and corrupt governments, then energy-per-transaction is an inappropriate metric.

Bitcoin is the only asset with these properties, and it doesn't need many transactions to do what it does best.

If you find you're angry about Bitcoin's energy usage but you're not even more angry about the energy wasted by industries and technologies that have much less social utility, then maybe it's not the wasted energy that you're angry about after all.

> without capital controls from coercive and corrupt governments

And conversely too! letting corrupt governments and corporations transfer money with ease!

That's always the problem with freedom, it works for good guys and bad guys alike.

Let's not forget druglords, illegal arms dealers, hitmen...

Paging argument from 2010...

You would use Monero or physical cash for those activities.

Bitcoin is much easier to trace once you have a lead.

It's just a tradeoff between opsec and ease of use. Monero is harder to turn back into cash, precisely because it carries a worse reputation.

Note that my argument was pro btc - I can see underground demand sustaining its price in the long run.

Because you definitely can't do that with fiat...

This is such an old, tired take.

No, you really can't do that easily and sometimes at all with "fiat". Which is why "money laundering" is a thing. Or drug cartels having the problem to stash or move hundreds of thousands of small dollar bills over North and South American borders.

A somewhat trustworthy state having control over or transparency into money transfers generally is a very good thing. If your government isn't trustworthy enough, you have bigger problems than the currency.

When it comes to truth, old != bad. The oldest profession is still around, for example, because demand will always be there.

There are industries with less social utility than Bitcoin?

> However, if the ultimate utility of Bitcoin is to store trillions of dollars of value, [...] then energy-per-transaction is an inappropriate metric.

The amount of energy used in equilibrium in Bitcoin mining is proportional to the price of Bitcoin (between halvings and holding the fraction of total mining costs spent on energy constant). If the price of Bitcoin rose enough that it had a 2TN market cap tomorrow, miners would have incentives to burn 3-4x more energy.

I agree that transaction-based accounting is flawed, but it doesn’t make the problem go away.

By computing more hashes, you raise the difficulty for everyone else mining so you do ultimately push out the less efficient miners.

That's efficient if you define efficiency as hashes computed per watt. That's the metric by which less-efficient miners are driven out.

However, this does nothing for the overall energy efficiency of Bitcoin, which could be defined as transactions processed per gigawatt. If energy gets cheaper, that number only gets smaller.

The problem is that there is, by definition, economic incentive to spend an amount of electric energy equal to the economic value of transaction confirmation to all protocol participants.

If someone has money and power already, they can take a better chance to get more bitcoin. How about the other guys and the other guys ... 10/20 years later?

There’s also proof of stake which requires a lot less energy. I think a lot of nuance is missed when discussing Bitcoin and its consensus mechanism as the only way of running a blockchain

So far no one has shown in practice how to run a decentralized proof of stake network.

How do you bootstrap a node in the face of conflicting history? How do you handle network partitions?

There are also open question surrounding security, prevent actors from colluding and so on. Most other blockchains do not even bother. At least the Ethereum people have the ambition to have something practical. Let's study the design when it you can make an eth transfer with it.

These discussions are like when in a discussion about uranuium waste management someone inevitably cries about how we should just scrap it and go with fusion energy instead.

Great idea, but let's have something working first.

It is quite noteworthy that Ethereum trusts proof of stake so much that they want to move to it though. Ethereum has some really good devs.

Last time I checked for altcoins that are trying to solve Bitcoin‘s various scalability problems I ended up with a proof of stake coin too, Nano (using the rather new idea block lattice, small but so far quite low-toned development team that seemed to focus on problem solving rather than scammy marketing).

I’m surprised that Ethereum is convinced that they can „just move“ from proof of work to proof of stake. Should be possible for Bitcoin too then, shouldn’t it? (Maybe I’m missing something about this transition.) Nano required a whole new development.

Ethereum's social contract has evolved to become much more "fluid" and open to change than Bitcoin's. Mainline Ethereum today is is very different in many ways to what it was two years ago. There is even a proposal with a lot of support to have transaction fees mostly burned rather than mostly given to miners. Aside from minority forks, something like that would never happen in Bitcoin in a million years.

So, while Bitcoin techinically could switch to proof of stake, it probably won't.

Tezos is up and running and working just fine.

Plus Cosmos, Polkadot, Solana, Cardano, Avalanche, NXT, and probably others I'm forgetting. There are plenty of PoS systems in production these days!

As someone who hasn't been watching closely for a while, any thoughts on which of those seem most interesting?

To me Mina and Zcash are interesting since they (plan to) use cryptographic arguments to show that chain states are valid, which leads to some nice features, such as better light clients. I'm probably biased since our own startup was influenced by them. Besides that, I'd say NEAR since it seems like the most promising sharded chain that's in or close to production.

(They weren't on the list since Mina hasn't launched quite yet, Zcash isn't PoS though it might switch, and NEAR isn't PoS in its current form, but will be soon.)

All of those are run by a singular entity. This makes it trivial for them to implement PoS.

That's not the use case something like Bitcoin is built for.

That's flatly untrue.

How can that be controversial? They all had an ICO to begin with. Some of them have even registered a company or organization to back it up.

The controversy is in the language "started" vs. "run."

Tezos, as an example, is governed by voting and on-chain governance. It isn't "run by a single entity."

Ripple or Tether are, though, sure.

Governance is more than a voting protocol. Tezos is not governed by equal actors.

There is a "Tezos Foundation". They control 30% of all staked XTZ. They checkpoint the blockchain regularly to provide correct history.

It seems to me that admin rights to the bitcoin core github repo count for a lot more than a 30% vote.

You can make near-zero-cost ETH and ERC-20 transfers right now using L2 solutions like state channels and rollups. There are also a few POCs of applications like Uniswap running on L2.

> I think a lot of nuance is missed when discussing Bitcoin and its consensus mechanism as the only way of running a blockchain

Bitcoin holders are financially incentivized to steer everyone else toward Bitcoin.

As long as early adopters are financially invested in alternative currencies failing, we’re going to continue seeing a focus on Bitcoin.

Proof-of-stake distributed ledger technology is extremely efficient. Hedera Hashgraph uses a gossip protocol and has extremely high transaction-per-second support with finality. Already being used by major corporations.

https://hedera.com/

Proof of space, too. See e.g. https://www.chia.net

Ugh, that seems like it just exchanges one form of waste for another. If it takes off the price of storage is likely to rocket in the same way GPUs have.

Let's not shoot ourselves in the foot, again, for another digital gold rush?

IIRC some Proof of Storage algorithms allow you to store actually useful data, and there shouldn't be any incentive to get fast storage, so it should have less of an effect on the average person.

True, and I understand this is how projects like Sia work. But IIRC Bram Cohen's proposal was effectively PoW but with storage.

When is that rolling out for Ethereum

It's a really complicated transition, but it's already started. The estimate is that there's probably about $1.5B worth of transactions left to mine, but most transactions will be switching over to PoS in the coming year. No one knows for sure when the full transition will occur, it's more of a gradual phase-out.

I like Ethereum not just because of the more responsible PoS but also because mining is currently asic resistant. A lot of the waste going into bitcoin right now is building specialized miners that will go into a landfill after a generation or two. Eth can still be mined with hardware you already own.

It's coming with Ethereum 2.0. The transition process to Eth2 is on the way, it is estimated that they will finish by 2022.

> "Proof of work" is just a synonym for "having wasted tons of energy." This is my main beef with these systems. Gaining fundamental value from the act of wasting energy [...] is not a sensible basis for a 21st century technology.

This is my main beef with the system we live in already.

See ; Bullshit Jobs

https://en.wikipedia.org/wiki/Bullshit_jobs

"Bullshit Jobs: A Theory is a 2018 book by anthropologist David Graeber that argues for the existence and societal harm of meaningless jobs. He contends that over half of societal work is pointless, which becomes psychologically destructive when paired with a work ethic that associates work with self-worth."

I am not contesting the Bitcoin situation is ... interesting.

(edit = comment layout )

Hypothetically this could be amortized better by increasing the transactions per block. Just because Bitcoin was first doesn't mean it's the best wrt efficiency. PoW can be competitive with legacy infrastructure with good clearing design.

I think Donatella Meadows made it quite clear a while back that fiddling with constants and coefficients doesn't really have an Effect with capital E.

At the very least the proof of work ought to compute something useful.

It can't be made useful because the output of a useful calculation can't be known in advance - otherwise it wouldn't be useful.

I think Gridcoin offers a happy medium. It's proof of stake, but it also offers rewards for participating in BOINC-based research projects.

https://gridcoin.us/

Not exactly "proof of work" in the same sense its being used, but I wish that the project would get more attention.

I've been trying to get people to pay attention to this project for years. I think a lot of the "useful computation" chains are much more likely to survive the next cryptocurrency crash because of their inherent utility. For example, the various storage coins, stuff like Nym or Oxen, Namecoin, etc.

You can gain use computing to gain specificity from complexity though, and recognize that new specificity - i.e. brute force protein folding calculations.

That would be great, but unfortunately the structure of problems that are well suited to proof of work (hard to compute but easy to verify in a decentralized way) doesn't seem to have too many practical applications.

Aren't there lots of NP problems that have this property?

Yes, but how many of them are distributable in the way required by the constraints of mining? (I really hope I'm wrong and would love to see a counterexample of "constructive" mining!)

All of them, by definition.

It's really hard to see how BTC does not get outlawed by most governments. Since financial transactions are zero sum and provide no economic value, it's the equivalent of a giant methane leak.

We detached this subthread from https://news.ycombinator.com/item?id=26090317.

The correct comparison is with the transaction and storage costs, in electricity, of storing gold in fort knox and moving it once a century in a boat across the ocean. Tesla's 1.5 billion USD in bitcoin probably won't move in your lifetime.

Mining for gold is even more inefficient. You've got dig earth several kilometers down, build massive infrastructure, use harmful chemicals to purify it, lose lives etc etc

That's falsely equating fixed and variable costs. A mined bitcoin is useless if the network isn't burning electricity, whereas mined gold is continually useful after mining.

At least gold is useful for something though

Gold also doesn't require constant electricity expenditure to keep it's value and "utility"z

A kilo of gold once mined is perpetual. A bitcoin needs the whole network to keep running (and be large enough so that miners can't collude) to keep it's value.

proof of work is cleverly misnamed, because what's actually going on is proof of wasted work, or simply proof of waste. you are proving you threw a precise measure of something valuable away (money/resources/energy... whether your own or someone else's).

the behavioral econ assumption is that you will seek to recover that loss someday, thus requiring you to hold and protect the receipt token representing it until that time.

Ethereum has already started transitioning to proof-of-stake which will significantly lower transaction costs in Ethereum.

Indeed, the comment shows how valuable Bitcoin actually really is - it's pretty impressive.

Interesting take. Does the amount of money spent on Herbalife show how valuable it actually really is? What about essential oils or healing crystals?

In a tautological way, sure, everything's worth what somebody's willing to pay for it. But the above examples must call into question whether there is a distinction between what something is worth in dollars, and what its merit is.

Good points. To give it a totally different spin, imagine Bitcoin's value is pretty constant actually and what changes is the dollar (its being devalued at a rapid pace at the moment).

There is only a limited supply of Bitcoin, but there will be more and more dollars printed over the next decades (making the dollar even less valuable compare to the constant Bitcoin). So Bitcoin will be interesting to watch - quite exciting for future and to see what's gonna happen.

Most newer cryptos are built on proof of stake. So your criticism is mainly aimed at Btc.

Naturally. While POS cryptos share many other bad sides with Bitcoin at least they don't ruin the planet so they're far far better.

POW only ruins the planet if you're using non-renewables / non-nuclear. It's not intrinsic.

quibble on "you're". It's a shared planet, so it gets ruined if enough mining is done on non-renewables. Most people complaining aren't mining at all.

Was this a useful comment? "You're" is just a stand-in for "people are". The second person is a fairly common way to structure such a sentence.

Sure, you can argue that nuclear (or whatever) is a better way to produce energy than most/all others currently known, but that's a very low bar to clear.

There are more and less sucky ways to produce energy, but let's be clear: all of them suck to some degree. There is no type of energy production where the mere act of producing that energy is a net positive.

Then the argument turns to whether or not the gains from proof of wasted energy-type systems outweigh the drawbacks of producing the required energy -- even if that energy is produced in the least sucky way. I'm firmly in the no camp.

Which is where most energy is derived from, so it's ruining the planet.

You may be interested in Ada coin. It uses a "proof of stake" algorithm instead which does not rely on processing power. Virtually infinitely more eco-friendly.

Cryptocurrency are inefficient by the design only at creation. But once mined, in principle with the right setup transactions can be arbitrary cheap. It is similarly as with gold. It is expensive to dig it once, but once a coin is minted, it can be used for millions of transaction making transaction cost rather low.

This is completely incorrect. Transactions are processed as part of mining. You cannot simply stop mining and continue doing transactions for free; the whole point of the mining process in a PoW cryptocurrency is that mined coins are given out as a reward for helping to secure transactions.

What happens when all the bitcoins are mined? There still has to be some settlement process or else the blockchain ends.

Mining continues but the reward comes solely from transaction fees.

The algorithm makes it such that the last coin will be mined sometime around 2140. My assumption is that people just plan on moving to another fork before that point.

Miners are incentivised with transaction fees alone after all the bitcoins are mined.

> What happens when all the bitcoins are mined?

Then miners will get only the transaction fees. But it's still the same process. (Anyway, that won't happen, the number of available BTC to mining will asymptoptically approach 0, but will never get there.)

No, new bitcoins run out around 2140. From then on there will be no new bitcoin

Actually your analogy is flawed. Because the way bitcoin works you have to pay some energy cost per block and you have a limited time to use that block because the next block will be coming in in 10 mins. It would be more like we had to make new coins every 10 mins.

Or am I missing something here?

Bitcoin probably is not fixable, but other designs are much better at supporting cheap transactions. There is nothing inherently expensive in maintaining the block chain.

this thread is about bitcoins. So I thought its implied we are talking about bitcoin. Maintaining a POW block chain is inherently wasteful, but yes POS blockchains can maybe save this.

Even with bitcoins just consider what happens once all bitcoins are mined. Does the whole thing collapse? Or do bitcoins continue to be used similar to POS?

Then the transaction fees associated with every transaction fund the miners.

The block mining algorithm doesn't change, right? So I don't see how it will move to POS. That will require some sort of buy in from the users. Every time something new is introduced in Bitcoin the legacy remains and the new ideas fork the chain into their own thing. Usually these alternatives either die or ignored.

Could you elaborate on that? How can future transactions be accommodated if everyone stops mining new blocks?

Miners get compensated by transaction fees too.

In the grand scheme, the "tons" (or rather, gigajoules) of energy wasted by the Proof-of-Work are nothing. The Sun radiation that hits Earth every second dwarfs this consumption.

Also, one of the definitions of life is "a local decrease in entropy, at the cost of a global, larger, increase in entropy". By this criteria, every living being is a bad thing for the universe, because it accelerates the global heat death ever so slightly.

I guess my point is, some uses of energy are acceptable, even desirable. And every use of energy accelerates the heat death of the universe, but we humans are insignificant on this scale; there is really nothing that we can do to even accelerate our Sun's death. We're not even Kardashev-1 :)

I was going to point that the universe is not relevant here, only humans, the energy we produce and how we produce it matters. But then I realized that either your comment must be tongue in cheek or you are trying to hard to rationalize this than there is no point on try to convince you.

Okay, I admit my comment was a bit rushed, and I apologize for that. It's not tongue-in-cheek, and I don't feel that I'm rationalizing things (but then, I guess I wouldn't feel it even if I were).

I was trying to view things in a bigger picture. Ofcourse we are using too much fossil fuels at this time. Ofcourse our functioning is pretty unsustainable.

What I am trying to say, is that sometimes high energy usage is acceptable. Should we stop launching things in orbit? Should we stop the LHC?

I think that for _some_ people, having a thing such as bitcoin is worth the energy expenditure. It is perfectly okay that for you, it _isn't_ acceptable. I am not trying to convince you of anything. Just affirming that there exists a subset of humans that find value in it, and therefore are willing to allocate resources -- be that electrical energy, fiat money, or mindshare.

> Also, one of the definitions of life is "a local decrease in entropy, at the cost of a global, larger, increase in entropy". By this criteria, every living being is a bad thing for the universe, because it accelerates the global heat death ever so slightly.

The discussion about energy use of Bitcoin is within the context of the effectively closed system of Earth.

Until such a time that solar generated electricity is too plentiful to meter and ubiquitous, Bitcoin mining will cause huge demand for fossil fuel based energy, with the associated consequences for climate change.

> effectively closed system of Earth.

All the plants, algae, and everything with chlorophyll would beg to differ. I really think we cannot consider Earth a closed system, because we're not a rogue planet in the insterstellar void.

Our solar _system_ would be a better approximation of a "closed system".

Solar radiation is the only significant exogenous input to the otherwise closed system of earth, and I directly addressed it in the context of solar electricity production needing to grow by a huge amount to offset the carbon footprint of BTC mining.

> Our solar _system_ would be a better approximation of a "closed system".

That's both pedantic and irrelevant. The rest of the solar system (other than the Sun) have practically zero bearing in any way on the energy consumption of BTC mining, or the damage to the climate it causes when powered by fossil fuels.

>In the grand scheme, the "tons" (or rather, gigajoules) of energy wasted by the Proof-of-Work are nothing.

>but we humans are insignificant on this scale; there is really nothing that we can do to even accelerate our Sun's death. We're not even Kardashev-1 :)

I am shocked that you are brave enough to write these things in the same comment. The problem with Bitcoin and proof of work is that it's literally capable of consuming our sun and all energy in the universe.

Is that you, Gary Johnson?

No idea who's Gary Johnson. I'm just a guy from Eastern Europe, watching too much Isaac Arthur.

Just a playful joke. He's a US politician who ran for president in 2016 as a third-party (i.e. no chance of victory) candidate.

At one point he was asked about his long term view on global warming, and he said "In billions of years, the sun is going to actually grow and encompass the Earth, right? So global warming is in our future."

Ah, I see. Thanks for explaining :)

Gary's response was obviously tongue-in-cheek, even somewhat flippant. My comment wasn't trying to be ironic, though I now see that it could be interpreted so.