The future of crypto-assets? (from the comments)


I think one problem with this view is that the bitcoin storage cost is likely a function of the capitalization of bitcoin.

Bitcoin security is a function of the amount of miners. The amount of miners is a function of mining revenue streams. If there are not enough miners, stealing bitcoins becomes possible. So the bitcoin network needs to generate enough mining revenues to keep enough miners interested. This amount has to be a function of the capitalization as if it ever diverges stealing bitcoin would become a valid strategy. So bitcoin must have a negative yield, either because you need to pay for the miners energy or because your bitcoins are going to be stolen. This is currently hidden by the wave of investment in the sector (and the funding from the bitcoin seniorage), but that’s got to stop at some point. I don’t think gold for example has similar features. Securing gold is not a function of of the dollar value.

If miners are a competitive industry, miners revenue has to roughly equal miners costs, so it’s a real cost. Compared with other fiat currency cost of storage it seems incredibly wasteful. Storing government bonds is basically costless. You might get negative yields, but those are transfers, not consumption. Gold was valuable because it doesn’t rust. It doesn’t require maintenance. Bitcoins require a lot of maintenece… I think that’s a weakness…

Here is the link.



"Bitcoins require a lot of maintenece… I think that’s a weakness…"

HUGE fundamental weakneness

That's like saying electric lights and hospitals are fundamentally weak because the electric grid has to be maintained. I don't think you appreciate what the data center business is like. It operates with way more efficiency and reliability than say.. oil production. The price of compute power is very steady and going lower. There is a fundamental cost to do the compute that Proof of Work requires, but that's part of the bargain. If you want a tractor, you need a large engine. So a lot of people have already invested substantial sums into mining equipment for Bitcoin. It's a feature.

It does not matter if computing prices are going lower when it comes to Bitcoin. Mining prices stay the same. This is because mining is a zero-sum game, you compete with the other miners for slices of the same pie.

Mining prices are coupled to the price of bitcoin, because if you are getting paid in bitcoin you have more incentive to spend on hardware and compete with everyone else if it goes up. If people ever stop investing large sums in mining, a single miner who does do that can monopolize the network and start forging transactions, destroying the currency.

The original idea of bitcoin was that mining would be cheap enough to do in average hardware, and thus mining would remain decentralized, but that went face first into forces that are well described in Wealth of Nations. It's not surprising that crypto libertarian utopians just forgot to read the very basic on economics.
This specific problem is why other cryptocurrency plays are trying to use Proof of Stake vs Proof of work. At the same time, that approach hasn't lead to the strong growth forces that Bitcoin has. This is not surprising when it comes to tech: The tools that are best designed in for the long haul often lose to those that optimize for immediate marketshare.

I suspect that at some point, someone will take the current bitcoin chain as a base and change a bunch of key algorithms to make more sense. They could do the same right now, but taking in the bitcoin chain as a baseline gives people good excuses to buy in: 'If this fork succeeds at all, and you already have some bitcoin, you start with free money!'

As of now, there's no real solution to the glaring deficiencies in PoS. PoW still has much better security and incentive guarantees. The only way PoS can be made to work with the algorithms we have today is either by relying on some centralized authority (Ripple, Iota), hybridizing with PoW (Dash, Decred), or just rolling with it and saying "we'll just fork if the network gets Sybil'd" (Neo, Lisk, Pivx).

If we want society to stop "wasting" the resources used by mining, a more fruitful avenue is probably figuring out how to shoehorn some socially useful problem into a viable PoW algorithm. That's easier said than done. Bitcoin uses SHA2 because it has really good properties as a cryptographic OWF. There's no viable improvements to just pure brute forcing. Otherwise a single miner could 51% by improving the underlying algorithm.

The question is what kind of useful real world problem works as a OWF. Maybe protein folding? Animation rendering? Hypersonic aerospace design?

Wow, I must be getting old. I am a trained financial economist and I didn't understand a word of that.

@Larry Siegel- it's not hard to figure out, and I'm sure you could do it LS if you put your mind to it. I'm not in the field but I figured out what Doug is saying (I have in the past invested in bitcoins so I knew a thing or two about them). The only thing I did not know was that Proof of Work vs Proof of Stake, I assumed they were largely the same (that once some miner 'mines' a new bitcoin, it's trivial to verify it's real, but I guess it's not what I assume). The part about forking the network only works if the 'good guys' have > 50% of network computing power, and if the "bad guys" have not yet taken over the network.

The OP was very good, and another reason why BTC is a pyramid scheme. One reason I got out (when BTC was less than $1k a coin) was I realized due to the huge blockchain a non-centeralized version of BTC would never work to replace Mastercard/Visa/Amex, it's just too slow. If you centralize the bitcoin, as Doug suggests as a workaround, you solve the problem of speed but lose the "pseudo-anonymity" of BTC.

People said similar things about the internet's inability to scale throughout its life. Although it doesn't look like Bitcoin could handle Visa's scale now, let's see what future changes bring us.

SHA256 dehashing of SETI signals

What about neural networks in AI? They are also GPU hungry, like bitcoin mining by the sound of it. But they are basically a giant linear algebra problem, does linearity make them unsuitable for PoW?

Funny, I don't know if it's viable as an OWF, but I bought a GT960 card when they were only kinda-new and was surprised how quickly I rocketed into the top .1% of folders in Folding@home... granted I had lucked into the GTI build that actually bothered to cool the converters, so I didn't have to worry about overheating, but still... I wonder what the ratio is between PF and bitcoin right now, haha.

Gold has negative yield too. Doesn't stop all those Indians from buying it by the ton.

Because they are demon-worshippers.

Could someone elaborate more on the nature of bitcoin storage costs? Why is it high? And how can it be reduced?

Let's say you fly to a faraway third world foreign country - Wallonia. You need to take out some cash and decide to use an ATM. One ATM is in the lobby of a huge skyscraper with a giant sign outside saying "Bank of Wallonia". The building is teeming with bankers wearing impeccable suits. Bank of Wallonia has very public advertisements all over the country. It sponsors the premiere soccer team and the country's prince sits on the board of directors.

The other ATM is in the back of some sketchy alley. It is labeled with some cheap logo of a supposed bank that you haven't seen anywhere. The ATM looks like it's a leftover from before the Soviet era. There's no office in sight.

Why do you feel more comfortable using the first ATM? The answer is signaling. The latter could easily be some scam, designed to rip off tourists a few hundred dollars at a time. Bank of Wallonia in contrast has clearly invested millions and millions in impressive buildings, ubiquitous advertisements, well-dressed bankers in suits. Well more than the plausible revenue from ripping tourists off for a few hundred bucks. Hence it's inherently more trustworthy, even though no one you trust is vouching for it.

Same story with bitcoin miners. The miners produce a single canonical transaction log. If someone tries to sell you some counterfeit bitcoin, they'd have to produce a blockchain with more accumulated computing power than the honest blockchain. That would be a very very expensive endeavor. Just like putting a giant skyscraper to scam a few tourists, the revenue from counterfeiting bitcoin falls far short of the costs.

However just like Bank of Wallonia, the resources going into that signaling don't come for free. Someone needs to pay for that computing power, just like someone needs to pay for skyscrapers, advertisements and bankers' fine tailored suits. Ultimately that comes from the customers in some way or another.

Thanks. I get this.

And I understand why the cost of mining gets progressively high. But what is meant by cost of storage here? Isn't it different from cost of mining?

I also get this, and you do know Wallonia is a real country, don't you?

Haha, no just learned that from you. Always thought it was the Alice and Bob of trade economics for some reason.

Calling Wallonia a third world country increases the probability that Doug is from Flanders ...

And if Wallonia is a real country, does that promote Belgium to the status of a real country as well?

I think the OP is talking about the integrity of the public ledger. If the ledger is vulnerable to fraudulent transactions, then bitcoins will not represent a secure store of value and will be worth less or probably zero (in terms of real goods).

In order to keep the value of my bitcoins secure, I depend on the miners to keep on mining. And the very fact that mining is costly is what makes things work; if mining was costless, then anyone could infiltrate the ledger. So in equilibrium, the system only works if miners are willing to spend resources to keep the ledger secure, which requires that the miners get paid. The other wrinkle OP adds is that as the total market cap of bitcoin goes up, then so must the equilibrium mining cost (otherwise buying enough processing power and submitting fraudulent transactions would become a profitable strategy).

So who is paying the miners? It must be the people who own the bitcoins.

Got it.

So the term "storage" is a bit of a misnomer here. What's meant by the cost of storage is basically the cost of solving the math problems in order to be rewarded bitcoins.

Yes the integrity of the public ledger depends upon having enough miners. But that has been proven by Bitcoin. The problem is not that mining is expensive, it's that to a certain extent, liquidity dilutes price. Until the very last Bitcoin is mined, the self-regulated consensus of the miners on the chain will change the difficulty of mining such that new blocks are created on pace. When somebody lucks out and solves a block too soon, the network bucks up the difficulty for the next block. So no matter how much compute you throw at the problem, the creation of blocks is restricted.That congestion artificially inflates price by constraining liquidity in a market with ridiculously high demand. Shorts get wrecked. So the bigger problem is that Bitcoin is operating now at scale. It can't get much bigger, especially in light of the trepidation of well capitalized market makers, and intransigent regulators. It's actually rather hilarious to watch the Bloomberg talking heads' confusion of why Bitcoin markets don't respond to the signals they are used to. Bitcoin is tiny. It's distributed enough for all the geeks in the world, but nowhere near enough for the world itself. Some fork of it may down the line after institutional money inflates it even more with giant, slow buys. In fact, that very well may be the way to kill BTC as we know it. Bitcoin's main problem, as I see it, is that it's hogging the spotlight and its ecosystem is not large enough to fulfill its promise. People are buying into the ecosystem that exists, which is right now a money machine, not a currency.

It refers to this, and storage cost as cost of securing keys and accounts:

> If there are not enough miners, stealing bitcoins becomes possible.

This is a common belief, but it's plain wrong. Miners do not secure against mis-spends (e.g. taking bitcoin that doesn't belong to you). That is accomplished by public key cryptography. The only way for Alice to send Bob bitcoin, is for Alice to sign a transaction moving her coins to Bob's address. Unless Alice's private keys are compromised, that's impossible. Security against mis-spends/theft is based on ECDSA cryptography. (And if ECDSA gets broken, we have a lot bigger problems than bitcoin theft).

What mining/proof-of-work protects against is *double-spends*. Think writing bad checks. Alice can prove to Bob that she's signing over her bitcoin to him. But Bob has no way of knowing that she hasn't also done the same thing to Charlie. If they both get duped, ultimately once the network reaches consensus one of them will be out of the bitcoin they're owed. That's where the block chain comes in. Miners create a canonical source of truth, by provably spending more resources than any other collection of party.

Alice signs her transaction and publishes it to the block chain. Once Bob sees it accepted and published, he can be sure that she didn't use those coins to pay anyone else before him.

So to be very clear, mining is **not** a cost to cryptocoin storage. Mining is a cost to cryptocoin transactions. Someone who just wants to buy some bitcoin and sock it away for 30 years, doesn't have to worry about the security of the mining pool. Even if it's compromised for a substantial amount of time during that period, his coins are still secure. The people who have to worry are the ones accepting bitcoin for payment.

Thanks, useful information.

However, it seems like the mining issue is still just as much a risk to storing value in Bitcoin, even if "storage costs" isn't the right description. If the mining was compromised such that it wasn't possible to verify transactions, surely the price of Bitcoin would collapse, wouldn't it?

Sure. But the sharp distinction to other assets looks a little more blunted. Surely if caravans of gold keep getting robbed that would diminish the market price of gold, and your portfolio of gold would suffer. Even if your own gold was never robbed.

I guess I would go beyond my original point to directly disagree with the original post. The thesis is that if mining revenue significantly falls behind market cap, then miners would break bad. I disagree because the rewards to a 51% attack on the mining pool are the largest transaction that you can get accepted by some counterparties in the space of a few blocks. This isn't a function of market capitalization, it's a function of transaction volume.

Market cap could continue to rise precipitously, and transaction volume as a percent of total coins could slow precipitously. Miners' incentives would still stay aligned with honest transactors. That's basically what's happening now as the fixed block size leads to congestion in the network. Miner resources would grow sub-linearly with capitalization, and cost per total circulating coin would fall.

This process can even be accelerated if we develop better off-chain technological solutions. E.g. lightning network or side-chains. The fixed rewards per block will keep halving, so miner revenue will be increasingly tied to the total volume of on-chain transactions.

If you have posted (or you ever end up posting) a comprehensive overview of crypto-assets, I would be very eager to read it. Your comments on this post have clarified a lot of my confusion surrounding bitcoin Thank you.

Excellent points Doug.

As best I can tell the major real danger to bitcoin is some kind of exogenous shock to its value (e.g. most gov'ts outlaw it, the Butlerian Jihad, quantum computing eats ECDSA, etc.) that makes bitcoin mining suddenly so unprofitable that a canonical blockchain cannot be maintained. Then Han isn't shooting first and Mary Sue defeats Superman, cats and dogs are living together... but at that point we have bigger problems anyway.

Thanks Doug for a good post. Beyond the mining costs incurred in verifying the integrity of the block chain. I think the costs of securing the private keys should also be considered. There are a lot of bad actors ( say North Korea) and the wallets’ security costs must be borne by the coin exchanges that pass them on to their customers. If your key is stolen it’s not like you can cancel your credit card or put a watch on your account. Any transaction with your stolen key will be considered valid and irreversible by the block chain. So you’re out of luck.

Doug is right. Thank you Doug for knowing what you're talking about.

Great clarifying comment, thanks Doug.

The number of miners is related to the mining reward, but the reward is not necessarily related to the capitalization. Most of the rewards come from fees paid on transactions, not on new coins generated on each block. They just need to maintain a high enough demand for transactions (relative to the fixed supply) to maintain security.

Not at all...Reward for mining a block is currently 12.5 BTC, so about $200k. Each block can contain about 2000 transactions, average of about $30 per transaction. Fees provide about 23% of mining reward only, an all time high. It used to be as low as 10%.

At the moment that’s true but as time goes on the mining reward decreases and the transaction demand will increase if adoption increases.

Also, fees are currently at 45% of the reward.

Three words:

Proof. Of. Stake.

Let me say first that I have no problem with other people inventing things, and profiting from them, even to the billions.

But your three words trigger four:

"Federally insured financial institution."

For millions of Americans that will remain the safest route for savings and payment.

This will be true right up until it isn't.

So, forever.

Yes, you're right: USD will be the global reserve currency for the rest of human history. Idiot.

It will be well past you and I are long dead.

That is certainly one possibility.

In the great recession I had a few of my banks fail. They did not go through dissolution, but I benefited from Fed managed and financed mergers to save those accounts.

So you know, it worked for me.

My mom actually had a bank fail, and got the FDIC payout.

The font in the body-part of this post is now about twice as large as the font in the comments. Wasn't that way yesterday. I kind of like it; hierarchical.

Bitcoin isn't cash, is another way of saying this.
The blockchain stops the double spending when miners agree. Hence we have to wait for the transaction to complete. If you had a small number of miners, we can get a conspiracy, a plot to fake the clearance and allow double spending via a cooperating agent.

Eventually they get caught, however.

Wow, there is a lot of confusion in this thread. Let me try to clear some up:

* Miner reward (in BTC terms) is predetermined by a fixed schedule (i.e., there will be 21mm BTCs 100 year from now, and there is currently 16mm BTC; so, there will be about 5mm BTC "printed out" in the next 100 years)
* If you hold BTC currently, you are being diluted roughly at the rate of (21mmBTC-16mmBTC)/100 per year
* The dilution above is the "price you pay" for keeping the network going
* As BTC rises in price, obviously the dollar cost of storage (i.e., (21mmBTC-16mmBTC)/100 per year ) rises. HOWEVER, in BTC terms the rate at which you are being diluted is constant

That’s about a 4% growth of the BTC supply per year, currently

Your key premise:
> If there are not enough miners, stealing bitcoins becomes possible.

Is only sort of true. You'd have to assume that the people who are remaining are all bad actors, which is hard to believe since people would see this coming long before and bitcoin would lose value along the way. The bad actors would be hurting themselves along the way, and their end game would be owning a LOT of something that's worthless. Not to mention the massive collusion that would be necessary to make this happen.

Check out this shitcoin:

Though to be fair this was only a problem for people who stored their money in their mattress, for people who put their money in interest-paying banks(remember those?) it was a non-issue.

On the cost of storing blockchain. Parts of the blockchain from the oast get r-written as transactions 'complete' their oath and end up agglomerate at WalMart. So some blocks with a lot of recent Walmart activity may get re-written, say after the winter sale, when all the money is collected. This is the part about reducing empty space in existing blocks by repacking valid blocks, it comes under the moniker orphan block in the literature. It is not much, and repacking reduces look up times later.

However, if that guy at the top of the chain goes on a spending spree with his gazillion, then there will be high probability of having to recompute a significant chunk of the total block chain, it will jam

So, let me get this straight. If Ethereum, Ripple, and Iota use Proof of Stake rather than Proof of Work, is this why they are cheaper to run but somehow also less secure from fraud and theft? If so, why have we not heard about fraud or theft happening with any of those cryptocurrencies?

How liquid are crypto-currencies? I notice Tyler calls them crypto-assets.

I saw somebody mention that he'd bought a tractor for Bitcoins but I have no idea how common that is. Maybe it is.

I also find it odd that Bitcoin is popularly represented as a gold coin. There are probably lots of unsophisticated people out there who don't know it's a string of code. I don't foresee too many people trading used cars for it. It's still priced in the government's money, so I think Tyler is using the correct term. This begs the question, to me anyway, what's the utility of the asset? What makes Bitcoin a better investment than real estate?

This betrays an almost complete misapprehension of how Bitcoin works. Miners provide NO protection against stealing coins. The protection is supplied entirely by cryptography and is free. Miners perform the role of preventing "double-spend" - telling two different parties you are paying them with the SAME bitcoin.

Mining is protection against fraud. Not theft.

It is not that different from gold. Mining gold is not required for using it as a value store - but in practice it is inevitable. For value store gold mining is a pure waste.

Comments for this post are closed