4 ms·

You wouldn't have to make a payment from "JOHN DOES ACCOUNT" to "PERVERTED PORN ACCOUNT". You could easily set up several wallets in the middle, obfuscation the

You wouldn't have to make a payment from "JOHN DOES ACCOUNT" to "PERVERTED PORN ACCOUNT". You could easily set up several wallets in the middle, obfuscation the transaction and gaining deniability. However, beware! If your wife is an paranoid data scientist, she could still find circumstantial evidence for your porn subscription.

saurik 13y ago
So, if you aren't going through a remixer (which is the point at which my paranoid data scientist friends point out "that doesn't even really help, at least in the long run, if not even in the short run"; and it isn't like any normal user is going to be doing this) the obfuscated wallet approach isn't really helping you much, if nothing else due to timing correlations. Even then, most normal people (the kind of people who are likely going to run into this problem most often) aren't going to understand how to do anything other than use the default features encoded into their wallet software (which are remarkably deterministic in function, almost never obfuscates the change making the middle steps trivially transparent, and generally don't provide any control over which transactions will accidentally cluster). Bitcoin is simply not trying to solve this problem: there are things bitcoin claims to solve, and things bitcoin succeeds in solving (whether on purpose or as a side effect), but "I want my transactions private" was never one of them. While it would be awesome if it were true, not all things related to payments are made better by bitcoin :/. It also must be pointed out that you aren't just dealing with your spouse at this point: you are dealing with everyone who might possibly care; maybe your pastor is a "paranoid data scientist", or maybe the noisy person across the street has taken up statistical correlation and inference as a new hobby. Maybe someone makes a new service that helps companies screen employees, and one of the things it does is scan the blockchain using some really epic algorithm coming up with some kind of bogus (and questionably-legal) "morality score" (I bring this up as there already have been companies constructed to do this for data on Facebook). The transaction record of bitcoin is out there for everyone to datamine, and if you make a mistake the record will still be there and still be public years from now. There might even be new techniques (whether bitcoin-specific, or generally in the fields of statistics or machine learning) that will make this data more transparent in the future. The defense model for "I don't want people I deal with to know about my porn addiction" of bitcoin vs. even a normal credit card does not come out very strong.
- pgsandstrom 13y ago
  Those are good objections. I guess the only safe way of dealing with bitcoins is to get money into a wallet without in any way associating it with your real life persona, which could be very tricky. And after that never ever use your wallet to pay a service that is assigned to your real name. Are there any good resources to read up on how the remixers work, and why they cant be relied upon?
  - saurik 13y ago
    I do not have any resources related to that; the summary (as I vaguely remember the arguments) came down to evidence being accumulated due to reasonable caps on fees and delays (if giving your money to the remixer occasionally ate 99% of the money and occasionally made it hold onto the money for a year, you are no longer in a position to effectively utilize it) combined with correlations on repeated transactions. To be clear, though, my argument was simply to move the "data scientist" line from "wallet obfuscation" (for which I then provided more concrete issues) to "remixers", and then to point out that normal people (the kind of person who didn't solve the underlying problem long ago using prepaid credit cards ;P) aren't using remixers and are unlikely to begin using remixers (as they probably aren't even using the more useful forms of wallet obfuscation: they are probably relying on what their wallet software is doing by default).

saurik 13y ago

So, if you aren't going through a remixer (which is the point at which my paranoid data scientist friends point out "that doesn't even really help, at least in the long run, if not even in the short run"; and it isn't like any normal user is going to be doing this) the obfuscated wallet approach isn't really helping you much, if nothing else due to timing correlations. Even then, most normal people (the kind of people who are likely going to run into this problem most often) aren't going to understand how to do anything other than use the default features encoded into their wallet software (which are remarkably deterministic in function, almost never obfuscates the change making the middle steps trivially transparent, and generally don't provide any control over which transactions will accidentally cluster). Bitcoin is simply not trying to solve this problem: there are things bitcoin claims to solve, and things bitcoin succeeds in solving (whether on purpose or as a side effect), but "I want my transactions private" was never one of them. While it would be awesome if it were true, not all things related to payments are made better by bitcoin :/. It also must be pointed out that you aren't just dealing with your spouse at this point: you are dealing with everyone who might possibly care; maybe your pastor is a "paranoid data scientist", or maybe the noisy person across the street has taken up statistical correlation and inference as a new hobby. Maybe someone makes a new service that helps companies screen employees, and one of the things it does is scan the blockchain using some really epic algorithm coming up with some kind of bogus (and questionably-legal) "morality score" (I bring this up as there already have been companies constructed to do this for data on Facebook). The transaction record of bitcoin is out there for everyone to datamine, and if you make a mistake the record will still be there and still be public years from now. There might even be new techniques (whether bitcoin-specific, or generally in the fields of statistics or machine learning) that will make this data more transparent in the future. The defense model for "I don't want people I deal with to know about my porn addiction" of bitcoin vs. even a normal credit card does not come out very strong.

pgsandstrom 13y ago
Those are good objections. I guess the only safe way of dealing with bitcoins is to get money into a wallet without in any way associating it with your real life persona, which could be very tricky. And after that never ever use your wallet to pay a service that is assigned to your real name. Are there any good resources to read up on how the remixers work, and why they cant be relied upon?
- saurik 13y ago
  I do not have any resources related to that; the summary (as I vaguely remember the arguments) came down to evidence being accumulated due to reasonable caps on fees and delays (if giving your money to the remixer occasionally ate 99% of the money and occasionally made it hold onto the money for a year, you are no longer in a position to effectively utilize it) combined with correlations on repeated transactions. To be clear, though, my argument was simply to move the "data scientist" line from "wallet obfuscation" (for which I then provided more concrete issues) to "remixers", and then to point out that normal people (the kind of person who didn't solve the underlying problem long ago using prepaid credit cards ;P) aren't using remixers and are unlikely to begin using remixers (as they probably aren't even using the more useful forms of wallet obfuscation: they are probably relying on what their wallet software is doing by default).