The core argument is that digital payment systems can fail very easily during times of "crises or war". This is because typical digital payment systems by design only work if the devices are connected to the internet. Moreover, credit mostly only works if you get permission from one of two foreign companies.
The question I have is, can you design a digital currency that works without the internet?
The way bitcoin and friends work is that you prevent the double spend problem by announcing the transaction on the network and waiting enough time that it becomes computationally infeasible to reverse the transaction.
However, can you design a hardware device, with some sort of security enclave, using which when the private key corresponding to some coins are transferred from the sender to receiver, the receiver can confirm that the keys have indeed been deleted by the sender.
Obviously, a look-alike device with a fake enclave can be built to scam someone. But if the practical difficulty of creating such a device is more than the difficulty of printing fake cash, then you can use such a digital currency.
It is very hard to clone a single modern SIM card or security chips in modern phones, the cost is significantly higher than 10K USD. So an off-line bank card or phone app holding, for example, up to 1K USD equivalent should be entirely possible.
EDIT: I missed the problem with retailers. A shop needs to be able to accept off-line electronic payments and be able to hold it for days before syncing with a bank for such payments to be an alternative to cache in case of emergencies and then we are talking about sums that are worth for criminals to be able to access. So such system will need highly non-trivial security including physical security and may end up more expenses than simply using cache.
Let me explain a bit more what I am thinking. I imagine that the central bank issues many many 1 cent digital notes. Each note is unique (much like cash) with a unique private+public key. In this system, whoever holds the private key for a note owns the note (similar to cash).
Now, the central bank also issues to each citizen a device that holds all the notes they own. If I want to give you a cent, I transfer the private key to you (by holding our devices close - no internet needed). But in order for you to trust that I indeed deleted the private key off my device, I need some sort of confirmation. This confirmation comes from a security enclave (consisting of a processor and the memory that stores the private keys) within my device. It should be able to emit a cryptographic certificate signed by some bank issued key that confirms the transaction.
The idea is that security enclave can't be hacked to extract either the bank issued key or the algorithms that the enclave performs.
Once a shop excepts money, they should be able to use it. Even having to wait for sync for a few days can put people in the red.
In my country, for example, it takes car-ride drivers to get their money from Uber/etc in about 1 week. So they almost entirely operate on cash. They need liquidity to pay for fuel and daily expenses.
Current mainstream card payment systems allow for offline payments (at some risk) as long as the merchant terminal has intermittent access.
EMV payment chips probably have enough compute and storage to do some sort of stored value application. With the right design, you could store transactions on the card and the terminal, and when using an online terminal, build a tunnel from the card to the issuing bank to a) report the card stored transactions and b) refresh stored value for offline transactions.
You don't need a block chain, PKI should be enough. Especially in countries without 5000 banks and 5000 credit unions which are pretty much banks.
What happens during the war, where the enemy systematically takes out your ISPs centers, so you no longer have internet in parts of the country for months? I think that is the kind of threat model that OP is discussing. Which is why cash is recommended.
The way I see it, when I carry cash and use it for transactions, there's at most a a few dozen people physically close enough to me to even try to exploit the transaction, and the most they could capture is the cash I have on hand. Chances are, if I dropped a wad of cash, they'd pick it up and hand it to me. Sure, it's easy to rob someone, but the chances of me being around someone that would even want to try is near zero, and the repercussions are significant.
When I make an electronic transaction, there's tens to hundreds of millions, if not billions of people that could attempt to exploit the transaction. Despite it being significantly more difficult, the chances of one of them trying and succeeding are still much greater, and they can take far more.
The core argument is that digital payment systems can fail very easily during times of "crises or war". This is because typical digital payment systems by design only work if the devices are connected to the internet. Moreover, credit mostly only works if you get permission from one of two foreign companies.
The question I have is, can you design a digital currency that works without the internet?
The way bitcoin and friends work is that you prevent the double spend problem by announcing the transaction on the network and waiting enough time that it becomes computationally infeasible to reverse the transaction.
However, can you design a hardware device, with some sort of security enclave, using which when the private key corresponding to some coins are transferred from the sender to receiver, the receiver can confirm that the keys have indeed been deleted by the sender.
Obviously, a look-alike device with a fake enclave can be built to scam someone. But if the practical difficulty of creating such a device is more than the difficulty of printing fake cash, then you can use such a digital currency.
It is very hard to clone a single modern SIM card or security chips in modern phones, the cost is significantly higher than 10K USD. So an off-line bank card or phone app holding, for example, up to 1K USD equivalent should be entirely possible.
EDIT: I missed the problem with retailers. A shop needs to be able to accept off-line electronic payments and be able to hold it for days before syncing with a bank for such payments to be an alternative to cache in case of emergencies and then we are talking about sums that are worth for criminals to be able to access. So such system will need highly non-trivial security including physical security and may end up more expenses than simply using cache.
Let me explain a bit more what I am thinking. I imagine that the central bank issues many many 1 cent digital notes. Each note is unique (much like cash) with a unique private+public key. In this system, whoever holds the private key for a note owns the note (similar to cash).
Now, the central bank also issues to each citizen a device that holds all the notes they own. If I want to give you a cent, I transfer the private key to you (by holding our devices close - no internet needed). But in order for you to trust that I indeed deleted the private key off my device, I need some sort of confirmation. This confirmation comes from a security enclave (consisting of a processor and the memory that stores the private keys) within my device. It should be able to emit a cryptographic certificate signed by some bank issued key that confirms the transaction.
The idea is that security enclave can't be hacked to extract either the bank issued key or the algorithms that the enclave performs.
Re your edit: it's not just a question of crime.
Once a shop excepts money, they should be able to use it. Even having to wait for sync for a few days can put people in the red.
In my country, for example, it takes car-ride drivers to get their money from Uber/etc in about 1 week. So they almost entirely operate on cash. They need liquidity to pay for fuel and daily expenses.
Current mainstream card payment systems allow for offline payments (at some risk) as long as the merchant terminal has intermittent access.
EMV payment chips probably have enough compute and storage to do some sort of stored value application. With the right design, you could store transactions on the card and the terminal, and when using an online terminal, build a tunnel from the card to the issuing bank to a) report the card stored transactions and b) refresh stored value for offline transactions.
You don't need a block chain, PKI should be enough. Especially in countries without 5000 banks and 5000 credit unions which are pretty much banks.
Something like the US Military Eagle Cash Card: https://en.m.wikipedia.org/wiki/Eagle_Cash although that is apparently being ended https://www.stripes.com/branches/army/2025-02-11/army-bids-f...
What happens during the war, where the enemy systematically takes out your ISPs centers, so you no longer have internet in parts of the country for months? I think that is the kind of threat model that OP is discussing. Which is why cash is recommended.
The way I see it, when I carry cash and use it for transactions, there's at most a a few dozen people physically close enough to me to even try to exploit the transaction, and the most they could capture is the cash I have on hand. Chances are, if I dropped a wad of cash, they'd pick it up and hand it to me. Sure, it's easy to rob someone, but the chances of me being around someone that would even want to try is near zero, and the repercussions are significant.
When I make an electronic transaction, there's tens to hundreds of millions, if not billions of people that could attempt to exploit the transaction. Despite it being significantly more difficult, the chances of one of them trying and succeeding are still much greater, and they can take far more.
https://news.ycombinator.com/item?id=43377517