5 ms·

From what I understand, people other than you and your friend will only trust one of your accounts (the real one), and will only accept money from it. So even t

From what I understand, people other than you and your friend will only trust one of your accounts (the real one), and will only accept money from it. So even though you might have 100 FakeYous, you will only be able to pay with RealYous.

cdancette 9y ago
But what if I pretend to be two different people? It seems we'll need some kind of centralized organization to verify identities here (the government? )
- hectorlorenzo 9y ago
  Yes, I thought exactly the same. And immediately after: "then what's the point...?"
- dllthomas 9y ago
  It's not quite "identity" that we need to verify, but rather uniqueness. Centralized approaches are theoretically easy; I'm not sure there isn't a possible decentralized solution. The one proposed does seem to be a partial solution, but probably not enough of one.
  - cdancette 9y ago
    I agree that the proposed solution seems not enough, but I'm impatient to see what will happen when they try it out. Maybe the trust chain will work out
  - baddox 9y ago
    I don’t think you need an explicit decentralized solution. I think the individual incentives are such that this type of scam is unlikely, because it’s probably prohibitively difficult to maintain two significant non-overlapping social graphs. I explained this more in this comment: https://news.ycombinator.com/item?id=15898504 https://news.ycombinator.com/item?id=15898504
    - dllthomas 9y ago
      I followed up over there :)
- baddox 9y ago
  I assume the point is that an individual wouldn’t just form a trust link with another individual without some confidence of their “true” identity. For me to trust someone’s currency, I would either need to be very sure of their identity through some trusted third-party system (e.g. a centralized system like government ID), or be involved in an extensive social (and probably in-person) relationship that would be difficult to fake (like friendship). For the latter, what I mean is that it’s probably prohibitively difficult to maintain two significant in-person social graphs that have no overlapping members. Sure, people can get away with it for things like romantic affairs, but the incentive there is very different than money exchange, not to mention that romantic affairs often get discovered.
  - dllthomas 9y ago
    I don't see that overlap in the social graph implies a high enough chance of discovery. Wouldn't discovery require that members from each group are actually involved in the same financial transaction (in a way such that people are paying attention to who's involved)? It may well be possible to keep that rare enough for long enough that people try it, and are sufficiently successful to be a problem. (This is certainly not confidence that it wont work, just lack of confidence that it will, pending further reading/experimentation/analysis...)
- SolarNet 9y ago
  Read the linked website. They include validators, a decentralized set of organizations, that can check government IDs to provide that sort of trust. (This is like the only response I seem to be contributing, no one seems to have read the full thing, so people keep asking questions like this that are answered by the source, so there is no interesting conversation).
dpark 9y ago
But of course when I pay you, you’ll have no idea which account is my primary. I’ll tell you it’d FakeMe23 and you’ll accept the money because why wouldn’t you?
- Jtsummers 9y ago
  Per the authors, they want people to use a real peer-to-peer trust network. I have never met you, dpark, and so the only way for you to pay me (or me to pay you) is for there to be a web of trust between you and me. And then we will only be paid with currencies of people who we trust. This also means they have a stake in it, they could end up with 50 dpark-fake coins that no one else will accept so they need to be careful of trusting people they don't know well. You won't pay me with dpark-fake because I don't trust it, instead you'll pay me with pg because I've trusted him and he's trusted you. However, this gets to the fundamental problem of the web of trust: It's really hard to do. Physical key signing parties are impractical to scale. You need systems like keybase which correlate your keys/accounts with each other if we aren't able to meet and communicate in person (which also requires trust, I wouldn't trust you after a single meeting, but would readily trust my friends who I know well but are physically remote from me). Quoting from the paper: This example demonstrates that Bob can only ever receive money that he trusts, and Alice can only ever spend money that other users trust in turn. Even if Alice makes 100 fake accounts and has them all trust each other, she will never be able to spend more than the amount of AliceCoins she has, since that’s the only account that other users will trust. This is why it is crucial that users take direct peer-to-peer trust relationships seriously.
  - dsr_ 9y ago
    Or you could sidestep all those issues by having a single currency which everyone who participated would accept, the same way that dollars and euros work.
    - Jtsummers 9y ago
      Sure, which is what we do now and generally works well. I'm not advocating for this system, merely addressing the question by pointing to their documentation where they answer it.

cdancette 9y ago

But what if I pretend to be two different people? It seems we'll need some kind of centralized organization to verify identities here (the government? )

hectorlorenzo 9y ago
Yes, I thought exactly the same. And immediately after: "then what's the point...?"
dllthomas 9y ago
It's not quite "identity" that we need to verify, but rather uniqueness. Centralized approaches are theoretically easy; I'm not sure there isn't a possible decentralized solution. The one proposed does seem to be a partial solution, but probably not enough of one.
- cdancette 9y ago
  I agree that the proposed solution seems not enough, but I'm impatient to see what will happen when they try it out. Maybe the trust chain will work out
- baddox 9y ago
  I don’t think you need an explicit decentralized solution. I think the individual incentives are such that this type of scam is unlikely, because it’s probably prohibitively difficult to maintain two significant non-overlapping social graphs. I explained this more in this comment: https://news.ycombinator.com/item?id=15898504 https://news.ycombinator.com/item?id=15898504
  - dllthomas 9y ago
    I followed up over there :)
baddox 9y ago
I assume the point is that an individual wouldn’t just form a trust link with another individual without some confidence of their “true” identity. For me to trust someone’s currency, I would either need to be very sure of their identity through some trusted third-party system (e.g. a centralized system like government ID), or be involved in an extensive social (and probably in-person) relationship that would be difficult to fake (like friendship). For the latter, what I mean is that it’s probably prohibitively difficult to maintain two significant in-person social graphs that have no overlapping members. Sure, people can get away with it for things like romantic affairs, but the incentive there is very different than money exchange, not to mention that romantic affairs often get discovered.
- dllthomas 9y ago
  I don't see that overlap in the social graph implies a high enough chance of discovery. Wouldn't discovery require that members from each group are actually involved in the same financial transaction (in a way such that people are paying attention to who's involved)? It may well be possible to keep that rare enough for long enough that people try it, and are sufficiently successful to be a problem. (This is certainly not confidence that it wont work, just lack of confidence that it will, pending further reading/experimentation/analysis...)
SolarNet 9y ago
Read the linked website. They include validators, a decentralized set of organizations, that can check government IDs to provide that sort of trust. (This is like the only response I seem to be contributing, no one seems to have read the full thing, so people keep asking questions like this that are answered by the source, so there is no interesting conversation).

dpark 9y ago

But of course when I pay you, you’ll have no idea which account is my primary. I’ll tell you it’d FakeMe23 and you’ll accept the money because why wouldn’t you?

Jtsummers 9y ago
Per the authors, they want people to use a real peer-to-peer trust network. I have never met you, dpark, and so the only way for you to pay me (or me to pay you) is for there to be a web of trust between you and me. And then we will only be paid with currencies of people who we trust. This also means they have a stake in it, they could end up with 50 dpark-fake coins that no one else will accept so they need to be careful of trusting people they don't know well. You won't pay me with dpark-fake because I don't trust it, instead you'll pay me with pg because I've trusted him and he's trusted you. However, this gets to the fundamental problem of the web of trust: It's really hard to do. Physical key signing parties are impractical to scale. You need systems like keybase which correlate your keys/accounts with each other if we aren't able to meet and communicate in person (which also requires trust, I wouldn't trust you after a single meeting, but would readily trust my friends who I know well but are physically remote from me). Quoting from the paper: This example demonstrates that Bob can only ever receive money that he trusts, and Alice can only ever spend money that other users trust in turn. Even if Alice makes 100 fake accounts and has them all trust each other, she will never be able to spend more than the amount of AliceCoins she has, since that’s the only account that other users will trust. This is why it is crucial that users take direct peer-to-peer trust relationships seriously.
- dsr_ 9y ago
  Or you could sidestep all those issues by having a single currency which everyone who participated would accept, the same way that dollars and euros work.
  - Jtsummers 9y ago
    Sure, which is what we do now and generally works well. I'm not advocating for this system, merely addressing the question by pointing to their documentation where they answer it.