Bruce's take can be summarized like this: with a blockchain, you're choosing to trust code and cryptography instead of a person. And when immutable code does the wrong thing, you can't trust them to fix it like you could a person.
This is a good take, and underappreciated - but there's a larger point being missed.
In high-social-trust situations, there are fewer benefits to blockchains. In low-social-trust situations, people aren't a good option. Bockchains provide an entirely new and welcome option for that scenario.
Relatedly, there are different kinds of distrust: mistakes and attacks are one, but platform risk is another that blockchains can alleviate. Putting it another way: I trust Twitter/YouTube/PayPal to generally work without issues, but I don't trust them not to change their offering in an exploitative way or become rent-seekers.
> In high-social-trust situations, there are fewer benefits to blockchains. In low-social-trust situations, people aren't a good option. Bockchains provide an entirely new and welcome option for that scenario.
How does this actually work in practice? I've never seen an actual non-convoluted / non-far-fetched case where this would apply.
For one: irreversible payment. Dealing with lots of scammers as a seller? They can PayPal dispute. They can chargeback. But if they send you cryptocurrency, the money is guaranteed.
This doesn't prevent scams in the other direction, but it's something. This is a common problem many independent sellers face when doing legal and legit business.
I believe he’s saying the opposite - that people like to think of blockchain as trust in technology, but in reality you are still trusting in people.
Any evaluation of the security of the system has to take the whole socio-technical system into account. Too many blockchain enthusiasts focus on the technology and ignore the rest.
> In low-social-trust situations, people aren't a good option. Bockchains provide an entirely new and welcome option for that scenario.
We've been handling these set-ups very well for millennia. You find a trusted intermediary. These days, for better or worse, globalization has created more and more of them.
Iran and Venezuela would disagree, or any country the US government has banned from the international banking network. The fact that one country can unilaterally block bank transfers between third parties solidifies the need for non-politically centralized banking.
Those intermediaries can be moderately to very expensive. Look at how much value credit-card interchange skims.
Blockchains offer a technological solution to a social problem- and threaten many "trusted brokers" with replacement with near-zero marginal cost software.
> Bruce's take can be summarized like this: with a blockchain, you're choosing to trust code and cryptography instead of a person. And when immutable code does the wrong thing, you can't trust them to fix it like you could a person.
Maybe Bruce hasn't heard of the Ethernum DAO fork [0], or the two subsequent forks. The immutable block chain seals a bad decision, then suddenly it wasn't so immutable any more. Just like someone paying you via direct bank transfer is irreversible, right up the money they are transferring was stolen from the bank - then it turns out it's real easy to reverse.
Crypto currencies are the tools of people, not the other way round. If we, or in the particular case the miners decide they want the block chain to be mutable, then that is what happens.
Bruce's take can be summarized like this: with a blockchain, you're choosing to trust code and cryptography instead of a person
It can't be summarised like that. He explicitly says: "Any evaluation of the security of the system has to take the whole socio-technical system into account".
His issues revolve around the fact that blockchain ecosystems have become centralised: "Just about everyone using bitcoin has to trust one of the few available wallets and use one of the few available exchanges. People have to trust the software and the operating systems and the computers everything is running on. And we've seen attacks against wallets and exchanges. We’ve seen Trojans and phishing and password guessing. Criminals have even used flaws in the system that people use to repair their cell phones to steal bitcoin".
While I agree with Bruce assessment that a lot of Blockchain evangelists push the whole "decentralized trust" or "trustless" or "in code we trust" ideas, I actually think you don't need to buy these ideas to see some hope in how things may work "better" if the promises of Blockchain materialize:
I agree entirely that we would need still be trusting humans even with Blockchain (shifting trust around). But I do see us gaining something with Blockchain: standardization. Rather than having multi-national and multi-lingual team of lawyers and diplomats to establish contracts and rules and how to interpret it and enforce them we can (try to) standardize all in code.
We would all still need to trust our lawyer/engineers, but now they don't need to trust some intermediary, translator/etc to get things done.
People think lawyers are just about making sure there are no loop holes, but its really not that. Lots of it is about making sure that two parties really agree on something including in the fine details.
Blockchain isn't going to magically solve the problem of having to figure out what the two sides actually want.
This is why the Stellar blockchain looks promising. On the Stellar blockchain, you put your trust in anchors and issuers (human entities) and each blockchain-subset (roughly the trust-union of specific-token issuers and specific-token anchors you choose) can create a highly distributed / decentralized blockchain from each users perspective, depending on how many anchors and issuers you are willing to trust.
It's a wonderful hybrid of "high-social-trust" networks run by humans (banks, governments) and distributed trust via code / your choice.
No one ever seems to address the fact that Blockchain still doesn't magically provide trust at all. It still needs people to participate for that and provide their resources (miners).
And for people to participate they need some incentive or reward. So how do you ever motivate enough people to participate in your distributed ledger?
You can give them blockchain tokens of course, but now you've created a new currency again with all its associated problems. Hardly a solution for your original problem of low trust.
> with a blockchain, you're choosing to trust code and cryptography instead of a person. And when immutable code does the wrong thing, you can't trust them to fix it like you could a person.
There are soft and hard forks that do change the behavior of blockchains so I'm not sure I entirely agree with the premise of this argument. Help me out, what am I missing?
I found the value of blockchain in a low trust setting easier to understand after reading up on the Byzantine Generals Problem fault tolerant distributed systems.
So far I haven’t seen a usecase that absolutely needs a blockchain either. I mean one that can only be addressed by a blockchain and will fall apart without it. But many people have used it as a general enough hammer to hit every nail they can find with it. As last it as people make money on it, even hypothetical money, it’s going to stay popular.
Buying drugs/murders on the darknet and getting untraceable ransoms from your cryptolocker malware are at least two solid use cases. One could argue evading currency controls is a 3rd.
The problem is that there's been no new uses found for blockchain after the initial uses discovered.
Blockchain isn't completely anonymous though. Bitcoin, for instance, can be and has been deanonymized to catch criminals. The entire concept of a blockchain is create an extended ledger, which isn't exactly an idea focused on privacy.
I think those are all basically the same use case - they're all about routing money of nontrivial value around governments. Much like Tor, it is not 100% unblockable, but it is quite difficult to block if you keep the internet flowing at all.
And like Tor, it can be used for both moral good and moral bad. (In fairness, you can decide that the good outweighs the bad for Tor but not for cryptocurrency; you can also believe that the good doesn't outweigh other harms like the computational/energy cost of running a secure blockchain, which necessarily needs a nontrivial fraction of the world's purchaseable computing power.)
Number 4: Having a highly volatile, non regulated investment option with a low barrier of entry that is probably controlled by a handful of bitcoin heavyweights.
A blockchain decentralizes trust, you can always solve the same problem a blockchain tackles by introducing a central point of trust. The question is if this choice is acceptable, more resilient, cheaper...
I actually (unfortunately) don't think this is true in a financial sense.
The only people with an incentive to use a semi-anonymous "currency" with extremely high transactions fees are criminals. You can argue that not all crimes are harmful or immoral, and you'd be right, but that's outside the scope of my comment.
The more crypto is used as a payment for crime, the more value it has. It becomes an investment vehicle for crime. If you think global crime will increase (and increasingly use crypto), then you can invest in it using Monero or whatever people think offers the best combination of liquidity and privacy these days.
What does "generating value" mean in this context. Aren't most transactions just shifting resources around?
If i have a bunch of cookies and my friend has a bunch of cup-cakes, and i trade some cookies for some cupcakes, what abstract value was created? The total number of cupcakes and cookies stayed the same, they just shifted around (like musical chairs). Neither me nor my friend are any richer, although we may be happier. Was this transaction useless?
No, when the flow of money is constrained it interupts actual labor. The great depression is the classic instance where if money had been made work would have been done and everyone would have been better off.
Blockchain can be used to issue something like Apple stocks in a decentralized manner which achieves million times more scale and lower cost than what's possible now.
Obviously a centralized entity like New York Stock Exchange could also copy the technology of using Merkle hashes and possibly achieve much bigger scale, at the end of the day it would still be the coordination issue for a single organization.
At the end of the day blockchain solves the coordination problem in a low trust environment in a definitive manner. As it turns out that the cost required to solve this coordination problem is really high when you compare it to the high trust, but low scale environment. but blockchain manages to achieve planetary scale throughput (obviously not right now because this is the foundation layer, things need to be built up).
It's like cash versus Visa. Cash is possibly billion times more scalable than Visa because Visa needs to reconcile a single database but cash transactions do not, as long as the recipients locally verify that they're receiving a real bill.
So far I haven’t seen a usecase that absolutely needs an Array either. I mean one that can only be addressed by an Array and will fall apart without it. But many people have used it as a general enough hammer to hit every nail they can find with it. As last it as people make money on it, even hypothetical money, it’s going to stay popular.
Store your private key on a paper wallet all you want, but what are you going to do during a tiger kidnapping? People look at me weird when I bring this up, but know nothing about real security. I could advice some ISO certified procedures to keep your money safe but your family is going to be the weak link. Go read In Cold Blood by Truman Capote to get an idea on what could go wrong.
A large chunk of the arguments tend to assume Proof of Work as the consensus mechanism which isn't a flaw inherent in blockchain (eg. could have Proof of Stake, etc.)
Yeah but if I can trust them in 10 seconds with very little effort on my part, I can be much more nimble with my strategy. If i can quickly verify the validity and source of primary data, I can make decisions on that data without hesitation. The value is in automating the trust, having a tiny little notary saying yes this is true, go ahead and make evidenced based decisions. Now of course we could do this with other tech, but I find the standardization of this trust mechanism to be the substructure of new business building tools and services around this quick trust, much faster than they can build them around trust accumulated by institutions, governments, notable individuals.
I've not seen any satisfactory explanation for how a working PoS system would operate. Granted I'm not an expert, but anytime I delve into whitepapers, they are written confusingly and without crucial details, which makes me wonder if some of these alternate chains are just vaporware or at the very least a lot less than they promise.
This is a misunderstanding of the concept of "trusted third parties".
Incredibly, yes, you have to rely on the software that you use to perform tasks to perform them. But this is not the same as placing another person in a mediating position where they can choose to betray you.
Confusing these two senses of "trust" results in an intuitively appealing but entirely mistaken family of arguments that aren't the gotchas people think they are.
The Very Short Introduction book on Trust covers different senses of the concept of "trust", including the ones I mention here, in a readable and compact way that I highly recommend.
> Yes, bitcoin eliminates certain trusted intermediaries that are inherent in other payment systems like credit cards. But you still have to trust bitcoin—and everything about it.
.... which is very tolerable to me.
Right now, it is more expensive to use blockchains do to both the novelty and specialization of it. As that becomes more ubiquitous it doesn't.
Whether we are talking about FX spreads to and from fiat currencies, or onchain transaction fees due to congestion, these all get better due to pressure and competition. A reason to improve. A permissionless ability to develop the improvements.
Even in things people pretend not to care about (after they lose money) like token sales, it costs token sale issuers more money to list on a random Asian exchange, than it costs to list on the Nasdaq. When the permissionless nature of the technology should mean the opposite is true. More competition and it fulfills that.
Enterprises are just moving away from their private blockchain to public blockchains, pretty much just this year, following Ernst & Young's Baseline framework.
Outside of the private sector, the regulatory frameworks specifically for blockchain technology have only begun to been ratified over the last 18 months, with the key market under the US Federal Government just getting providing clarity to financial institutions last week, at the Office of the Comptroller.
The metric you are referring to is a misattributed S-Curve of adoption that has condensed for a few select technologies over the last century.
Bitcoin is not the primary area of development for industry use of blockchains. There are various virtual machine technologies proliferating that bitcoin does not feature.
I'm not technical enough to disagree with Bruce. I'm also not an international law lawyer, fixer, intermediary, etc, to agree or disagree with how we have been trusting intermediaries for thousand of years and why this is impossible to replace.
The way I see it is that blockchains and smartcontracts have the potential to replace treaties, lawyers, corporate intermediaries, etc in an increasingly multipolar world where treaties would serve as a legal protocol to regulate transactions.
Maybe it is naive (I'm still trying to workout how this can ever happen if nation-states actively oppose these desintermediations) but I do think that privately orgs and people could benefit from technology that helps them transact with less middlemen.
In a world where China/Russia/etc is trying to avoid exposure to US money/laws maybe this could serve as alternative?
Readit News
This is a good take, and underappreciated - but there's a larger point being missed.
In high-social-trust situations, there are fewer benefits to blockchains. In low-social-trust situations, people aren't a good option. Bockchains provide an entirely new and welcome option for that scenario.
Relatedly, there are different kinds of distrust: mistakes and attacks are one, but platform risk is another that blockchains can alleviate. Putting it another way: I trust Twitter/YouTube/PayPal to generally work without issues, but I don't trust them not to change their offering in an exploitative way or become rent-seekers.
How does this actually work in practice? I've never seen an actual non-convoluted / non-far-fetched case where this would apply.
This doesn't prevent scams in the other direction, but it's something. This is a common problem many independent sellers face when doing legal and legit business.
Bitcoin isn't convoluted or far fetched even if you really want to disagree.
It's nuclear chemistry proof (unlike gold), and outside the reach of government.
Any evaluation of the security of the system has to take the whole socio-technical system into account. Too many blockchain enthusiasts focus on the technology and ignore the rest.
We've been handling these set-ups very well for millennia. You find a trusted intermediary. These days, for better or worse, globalization has created more and more of them.
Blockchains offer a technological solution to a social problem- and threaten many "trusted brokers" with replacement with near-zero marginal cost software.
Maybe Bruce hasn't heard of the Ethernum DAO fork [0], or the two subsequent forks. The immutable block chain seals a bad decision, then suddenly it wasn't so immutable any more. Just like someone paying you via direct bank transfer is irreversible, right up the money they are transferring was stolen from the bank - then it turns out it's real easy to reverse.
Crypto currencies are the tools of people, not the other way round. If we, or in the particular case the miners decide they want the block chain to be mutable, then that is what happens.
[0] https://en.wikipedia.org/wiki/Ethereum#The_DAO_event
It can't be summarised like that. He explicitly says: "Any evaluation of the security of the system has to take the whole socio-technical system into account".
His issues revolve around the fact that blockchain ecosystems have become centralised: "Just about everyone using bitcoin has to trust one of the few available wallets and use one of the few available exchanges. People have to trust the software and the operating systems and the computers everything is running on. And we've seen attacks against wallets and exchanges. We’ve seen Trojans and phishing and password guessing. Criminals have even used flaws in the system that people use to repair their cell phones to steal bitcoin".
I agree entirely that we would need still be trusting humans even with Blockchain (shifting trust around). But I do see us gaining something with Blockchain: standardization. Rather than having multi-national and multi-lingual team of lawyers and diplomats to establish contracts and rules and how to interpret it and enforce them we can (try to) standardize all in code.
We would all still need to trust our lawyer/engineers, but now they don't need to trust some intermediary, translator/etc to get things done.
Blockchain isn't going to magically solve the problem of having to figure out what the two sides actually want.
It's a wonderful hybrid of "high-social-trust" networks run by humans (banks, governments) and distributed trust via code / your choice.
And for people to participate they need some incentive or reward. So how do you ever motivate enough people to participate in your distributed ledger?
You can give them blockchain tokens of course, but now you've created a new currency again with all its associated problems. Hardly a solution for your original problem of low trust.
There are soft and hard forks that do change the behavior of blockchains so I'm not sure I entirely agree with the premise of this argument. Help me out, what am I missing?
The problem is that there's been no new uses found for blockchain after the initial uses discovered.
Regardless of the morality/legality of a use, the fact that those are real use cases for the currency do give actual value to bitcoin as a currency.
Now it's worth noting that "digital gold (speculation)" seems to far surpass ransom+online drugs as the reason for bitcoin to have value.
http://www.americankingpin.com/
About the rise and fall of The Silk Road
And like Tor, it can be used for both moral good and moral bad. (In fairness, you can decide that the good outweighs the bad for Tor but not for cryptocurrency; you can also believe that the good doesn't outweigh other harms like the computational/energy cost of running a secure blockchain, which necessarily needs a nontrivial fraction of the world's purchaseable computing power.)
In the end you do trust the code, a computer running the code or some currency exchange.
I actually (unfortunately) don't think this is true in a financial sense.
The only people with an incentive to use a semi-anonymous "currency" with extremely high transactions fees are criminals. You can argue that not all crimes are harmful or immoral, and you'd be right, but that's outside the scope of my comment.
The more crypto is used as a payment for crime, the more value it has. It becomes an investment vehicle for crime. If you think global crime will increase (and increasingly use crypto), then you can invest in it using Monero or whatever people think offers the best combination of liquidity and privacy these days.
If i have a bunch of cookies and my friend has a bunch of cup-cakes, and i trade some cookies for some cupcakes, what abstract value was created? The total number of cupcakes and cookies stayed the same, they just shifted around (like musical chairs). Neither me nor my friend are any richer, although we may be happier. Was this transaction useless?
Blockchain can be used to issue something like Apple stocks in a decentralized manner which achieves million times more scale and lower cost than what's possible now.
Obviously a centralized entity like New York Stock Exchange could also copy the technology of using Merkle hashes and possibly achieve much bigger scale, at the end of the day it would still be the coordination issue for a single organization.
At the end of the day blockchain solves the coordination problem in a low trust environment in a definitive manner. As it turns out that the cost required to solve this coordination problem is really high when you compare it to the high trust, but low scale environment. but blockchain manages to achieve planetary scale throughput (obviously not right now because this is the foundation layer, things need to be built up).
It's like cash versus Visa. Cash is possibly billion times more scalable than Visa because Visa needs to reconcile a single database but cash transactions do not, as long as the recipients locally verify that they're receiving a real bill.
Deleted Comment
Bruce Schneier makes the case against both cryptocurrencies and Blockchain technology, as I understand it.
I think it should get some more attention, because it makes a very good case, centered around the topic of 'trust'.
https://news.ycombinator.com/item?id=19097613
Dead Comment
There are always inherent risks that you just can't reduce with tech.
Incredibly, yes, you have to rely on the software that you use to perform tasks to perform them. But this is not the same as placing another person in a mediating position where they can choose to betray you.
Confusing these two senses of "trust" results in an intuitively appealing but entirely mistaken family of arguments that aren't the gotchas people think they are.
The Very Short Introduction book on Trust covers different senses of the concept of "trust", including the ones I mention here, in a readable and compact way that I highly recommend.
.... which is very tolerable to me.
Right now, it is more expensive to use blockchains do to both the novelty and specialization of it. As that becomes more ubiquitous it doesn't.
Whether we are talking about FX spreads to and from fiat currencies, or onchain transaction fees due to congestion, these all get better due to pressure and competition. A reason to improve. A permissionless ability to develop the improvements.
Even in things people pretend not to care about (after they lose money) like token sales, it costs token sale issuers more money to list on a random Asian exchange, than it costs to list on the Nasdaq. When the permissionless nature of the technology should mean the opposite is true. More competition and it fulfills that.
Outside of the private sector, the regulatory frameworks specifically for blockchain technology have only begun to been ratified over the last 18 months, with the key market under the US Federal Government just getting providing clarity to financial institutions last week, at the Office of the Comptroller.
The metric you are referring to is a misattributed S-Curve of adoption that has condensed for a few select technologies over the last century.
Bitcoin is not the primary area of development for industry use of blockchains. There are various virtual machine technologies proliferating that bitcoin does not feature.
The way I see it is that blockchains and smartcontracts have the potential to replace treaties, lawyers, corporate intermediaries, etc in an increasingly multipolar world where treaties would serve as a legal protocol to regulate transactions.
Maybe it is naive (I'm still trying to workout how this can ever happen if nation-states actively oppose these desintermediations) but I do think that privately orgs and people could benefit from technology that helps them transact with less middlemen.
In a world where China/Russia/etc is trying to avoid exposure to US money/laws maybe this could serve as alternative?