6 ms·

I'm not an expert, but I read the paper. So you don't need an intermediary to confirm that a particular digital transaction took place between two people. By

I'm not an expert, but I read the paper. So you don't need an intermediary to confirm that a particular digital transaction took place between two people. By analogy to the original problem as stated, that's like saying you don't need an intermediary to confirm that a message with particular content was delivered between two generals. You still need an intermediary to confirm that the recipient is going to act on it as intended. And I believe the history of Silk Road and its would be successors is evidence that an intermediary is still required post-Bitcoin. So you don't have to worry about the money not being there. You still need an escrow agent. http://www.nytimes.com/2014/01/19/business/eagle-scout-idealist-drug-trafficker.html http://www.nytimes.com/2014/01/19/business/eagle-scout-ideal...

eurleif 13y ago
>By analogy to the original problem as stated, that's like saying you don't need an intermediary to confirm that a message with particular content was delivered between two generals. No, that's not analogous. Bitcoin prevents double-spending. It doesn't just deliver messages. You're talking about a whole other level of cooperation. It's true that Bitcoin doesn't guarantee that, but no one said it did, and that doesn't imply it doesn't really solve BGP. Bitcoin does have a built-in feature, multisig transactions, that makes a more-trustworthy form of escrow possible.
- riemannzeta 13y ago
  Well I found a partial answer to the question I asked right in Andreessen's essay, but it's not the one you give here. I guess I think of the authentication of sender and content for recipients as being equivalent to the elimination of double-spending, and I don't see either as being responsible for the failure modes of fraud in digital payment systems. As a practical matter, recipients don't have has to question whether a particular person actually has the credit to complete a given transaction right now, and I'm not aware of double-spending credit as being a big problem for credit card payment systems. Who cares about solving that problem when we have a perfectly workable solution right now? But sure, I get the point about other forms of fraud might be harder -- namely, stealing account information necessary to initiate transactions. The question there is whether that problem too goes away in a world in which credit numbers are kept as private keys on a physical device (like a smartphone) and exchanged with merchants via something like SSH. Is the overhead of Bitcoin still worth it?
jjh42 13y ago
There is some academic interest in ability to use bitcoins scripting language to build "smart contracts." In practice, escrow isn't necessarily needed. In a typical consumer transaction the merchant is more trustworthy than the consumer (think me ordering from amazon). Amazon could just steal my money but that'd quickly destroy their brand. OTOH amazon accounts are free and so amazon is much more worried about me defrauding them. If amazon accepted bitcoin I'd pay them without worrying about escrow.

eurleif 13y ago

>By analogy to the original problem as stated, that's like saying you don't need an intermediary to confirm that a message with particular content was delivered between two generals. No, that's not analogous. Bitcoin prevents double-spending. It doesn't just deliver messages. You're talking about a whole other level of cooperation. It's true that Bitcoin doesn't guarantee that, but no one said it did, and that doesn't imply it doesn't really solve BGP. Bitcoin does have a built-in feature, multisig transactions, that makes a more-trustworthy form of escrow possible.

riemannzeta 13y ago
Well I found a partial answer to the question I asked right in Andreessen's essay, but it's not the one you give here. I guess I think of the authentication of sender and content for recipients as being equivalent to the elimination of double-spending, and I don't see either as being responsible for the failure modes of fraud in digital payment systems. As a practical matter, recipients don't have has to question whether a particular person actually has the credit to complete a given transaction right now, and I'm not aware of double-spending credit as being a big problem for credit card payment systems. Who cares about solving that problem when we have a perfectly workable solution right now? But sure, I get the point about other forms of fraud might be harder -- namely, stealing account information necessary to initiate transactions. The question there is whether that problem too goes away in a world in which credit numbers are kept as private keys on a physical device (like a smartphone) and exchanged with merchants via something like SSH. Is the overhead of Bitcoin still worth it?

jjh42 13y ago

There is some academic interest in ability to use bitcoins scripting language to build "smart contracts." In practice, escrow isn't necessarily needed. In a typical consumer transaction the merchant is more trustworthy than the consumer (think me ordering from amazon). Amazon could just steal my money but that'd quickly destroy their brand. OTOH amazon accounts are free and so amazon is much more worried about me defrauding them. If amazon accepted bitcoin I'd pay them without worrying about escrow.