The other day, I received a dollar bill with a stamp inviting me to visit WheresGeorge.com.  That site allows you to track currency as it circulates naturally through the economy.  One “Georger”, as users are called, enters the note’s serial number, (optionally) marks it, and then spends it.  The idea is that someone else coming into possession of it will see the stamp, and update its location before releasing it back into circulation.  Interestingly, about one piece of circulating currency in a thousand is registered at this site.  The dollar bill I was holding was spent in Bismarck, North Dakota, a few months ago.  Whoever recorded it in Bismarck was presumably informed that I received it, and will be able to keep tracking it as it continues to circulate.

My son saw me on that site, asked what it was, and in response to my explanation, he proclaimed that it was “cool”.  I tend to agree.  It appears to be an interesting hobby, not unlike Amateur Radio, (or sending out a message in a bottle or helium balloon) in that something is sent out into the environment to fend for itself, and it’s interesting to see just where it winds up.  One can order the rubber stamp for a mere five dollars (or an “Abe”, as Georgers would call it) postpaid, and I suspect Santa will be bringing us one.

In short, it appears to be a fun hobby.  But it also demonstrates something about privacy in an era in which the NSA can spy on its citizens with seeming impunity.  I have no doubt that the NSA, if it wants to, can easily look up every phone call I’ve made, every e-mail I’ve sent, and every letter I’ve sent and received.  I suspect that they can easily retrieve the contents of the phone calls and e-mails.

I suspect the U.S. Postal Service, bless their hearts, still stubbornly defends the integrity of the envelope, and there’s probably little chance that the NSA is privy to contents of the mail.  There’s something to be said for bureaucratic momentum, and I suspect there’s enough bureaucratic momentum at the post office for them to cling to the archaic notion that the mail can’t be opened without the right procedures being followed.  Since Ben Franklin was the one wrote those procedures, they’re probably still more or less in line with the Fourth Amendment.

I also have little doubt that the NSA can easily access all of my banking transactions.  If they care to find out, they can get a pretty good idea of everything I’ve purchased.  (And as you can see, Mr. or Ms. NSA Contractor, it’s all pretty mundane.)

I think that most people realize, or at least suspect, that all of this is true.  So to preserve their privacy, some people resort to spending only cash.  The idea is that the cash is anonymous.  Perhaps I’m wrong about the sanctity of an envelope at the post office.  But they are certainly wrong about the anonymity of their cash.   This is proven by the existence of WheresGeorge.com.  I know that one particular piece of currency was in Bismarck a few months ago.  I know (and now, thanks to me, the whole Internet knows) its location this week.  And it would surely be a trivial matter for the NSA to check my identity and that of the person who spent the bill in Bismarck.

It is quite true that this doesn’t really reveal anything particularly interesting.  The note in question passed through hundreds or thousands of hands before it got to me.  Even if the Georger in Bismarck was a notorious terrorist, the link to me is very tenuous.  But we do know a few things.  We know to an absolute certainty that someone crossed the Red River (or at least went around it) carrying this particular piece of currency.  We know this fact because of the note’s presence in two different places at two different times.  This is not a very profound piece of information.  But the reason why it’s not very profound is because only one note in a thousand is registered, and there are only so many Georgers doing the data entry.

But as more notes are tracked, the amount of information we can derive does not increase linearly.  Instead, it increases exponentially.  In other words, if twice as many bills were being tracked (1 in 500, instead of 1 in 1000), or if twice as many people were tracking them, then we would be able to figure out about four times as much information.  We would be four times as good at tracking down when the note was at a certain point, and where it was at a certain time.

Let’s look at the example of having twice as many people tracking the bills.  Currently, we know that a bill traveled 500 miles in 6 months.  But with twice as many people tracking, then we would probably know that it moved 250 miles in 3 months.  Increasing the number of bills being tracked would probably also increase the total amount of information available in a similar exponential fashion.

And if we increase both the number of people tracking and the number of notes, then the total available knowledge increases by a power of 4.  In other words, we now know 16 times as much information.  In other words, we would be able to narrow down the location of the bill in terms of movements of about 30 miles every 11 days.  We’re still not at the point where we can get much information about individuals, but we are getting much closer.  If we double everything again, we can now narrow down the location to about 2 miles, and less than one day.  Double it again (in other words, track 8 bills out of a 1000 instead of 1, and have 8 times as many “Georgers”) then we can narrow down the location to an eighth of a mile, and less than an hour.  We are now at the point where we can start to monitor individual behavior.  If we know that one individual was in that 1/8 mile radius, and he was in another 1/8 mile radius an hour later, it’s very likely that this is the person with the dollar bill.  And of course, if we have additional information available, then the likelihood becomes even greater.

In order to do this, you probably need to be smarter than me, and you probably need more computing power than I have.  I’ll leave it to the reader to consider who might have a lot of smart people with a lot of computing power.

Now, I’m sure the NSA has better things to do than hack into the WheresGeorge.com database.  All of that data is already readily available to them.  But I suspect they want cleaner data and more data.  And it seems to me that such data will become more and more available.

The cashier at the Kwik-E-Mart probably counts the till by hand, and doesn’t keep any records of what notes are in the till.  But he deposits some of them, and he hands the rest of them out as change.  The deposits are counted at the bank, and at some point, they are counted automatically.  And presumably, most large retailers count the money themselves, and probably do it automatically.  And many of the people who got change at the Kwik-E-Mart will spend it at a large retailer.

In fact, if the Kwik-E-Mart takes in enough cash, it might be worthwhile for them to buy their own counter. This money counting machine, for example, is very reasonably priced.  It can automatically distinguish between denominations, and it’s even set up for numerous currencies.  There is no indication that it can scan serial numbers.  But adding that capacity would be a trivial matter.  After all, if it can distinguish between a five dollar bill and a fifty dollar bill, or between a U.S. dollar and a Canadian dollar, or between a Euro and a Peso, then adding the additional capacity to scan serial numbers could easily be handled by software.

Nor is there any indication that this machine can be connected to a computer.  But this could be done with very little additional cost, and would add both convenience and security for the owner of the machine.  The manager of the Kwik-E-Mart could balance his books at the end of the day quite easily–just put the money in the counter, and let the computer print the deposit ticket.  This could also add security, since an employee counting the cash wouldn’t have the opportunity to “accidentally” enter the wrong amount after the bills were counted.

I don’t know if more expensive counters already have these features.  But they certainly will sooner or later.  The cost to add them is very low, and at least a few customers will want them.

Of course, many money counting machines are already connected to computers.  These include the automated check out lanes at the supermarket.  There’s no reason why they couldn’t record serial numbers.

In other words, if it hasn’t happened yet, it will very soon.  And as a result, there will soon be a large database of serial numbers, times, and locations.  This data will be stored on individual computers owned by retailers and banks.  Of course, there will not be any need to store this information in a centralized location, and it will probably be stored securely.  In other words, it will be quite similar to cell phone metadata.  And probably more useful.