Here’s what I understand is up with the NSA surveillance program. I’m going to be brief and cursory here, but I can fill in some of the cracks later from the notes and cites I’ve been accumulating.

It’s not ‘wiretapping,’ per se, and it isn’t a searchable database of call data. The ‘big database’ that people keep mentioning apparently refers to two things — or, rather, one thing with two separate components. The first and less interesting is a very large store of archived communications traffic, comprising phone calls, email, web pages, and probably IM and text messages, etc.

Now, the problem with giant piles of data like this is sorting them so that they’re in any way useful. People have been speculating about keyword searches and other such kinds of data mining, but that would be a stupid way to set things up. Besides just the practicalities, somebody who’s a terrorist would probably be less likely than other people to use (for example) the words ‘bomb’ or ‘terrorism’ in his or her communications. The way you want to organize a database like this is by personal connections. In fact, this is a standard Internet-age solution for such things, and that’s what they did.

The NSA program is based on a network of personal connections very much like Friendster. Except what it is, is ‘enemyster.’ For instance, if Osama bin Ladin’s cousin’s valet sent an email to your sister-in-law’s friend, who also emailed you, you’re in the database. People are scored according to their degree of separation from other people who are considered high-risk or interesting. Presumably, there’s a threshold or series of thresholds at which levels of surveillance are automatically kicked in, such that a person with a high enough score would have their communications monitored.


Now, everyone knows the thing about six degrees of separation, from the Stanley Milgram study. It’s easy to trace a single connection from one individual to another using a small number of steps. A grandmother in Dubuque would tend to have not a more distant connection to interesting or high-risk individuals, necessarily, but fewer close connections. One truly interesting thing about the NSA program is that people in fields such as journalism and politics are going to be scored extremely high because of their proximity to newsworthy figures including genuine terrorists. Christiane Amanpour, for instance, who is Iranian and is chief international correspondent for CNN, would have multiple, perhaps hundreds of high-interest connections lighting up the system like a Christmas tree. We’ve already heard an intriguing niblet about Amanpour being surveilled. Arianna Huffington, as well, is going to be all up in that shack. And that’s where things get very, very interesting.

Here, you’re the NSA. You can do this right now. Amanpour is married to James Rubin. Click on a name in James Rubin’s profile. Then go to that page and click another name, and so on.

That’s who’s in the system at levels of high interest, unless a given individual is intentionally and manually exempted from the database. You can see why this would be a problem, and why an investigation of this program is strenuously not desired by the current administration.

Are they abusing it for political purposes? Well, that’s a question, isn’t it?

In passing — and this is the thing I know the least about — it seems that the ‘net neutrality’ issue has a lot to do with Congress offering an incentive to the telcos for allowing their networks to be used (illegally). That’s not a point I’d be confident arguing right now, but it has the telltale corona of something that may later turn out to have been obvious.


Comments: 22


Ha ha ha… Wonder how Dana Rohrabacher feels about this?


It’s true the way you get to the data you want is via some sort of network/relationship analysis.

And USA Today was that broke it. I bet the editorial by what’s his name was the signal on the direction they were going to go with it. I love how USA Today was the absolute mockery of the old school papers for so long and this story is a huge f*** you.


Gavin, how do you know that it’s not a searchable database? If you want to pin down the source of a leak, for example, why not query for the reporter’s received calls, and then do a little research?


Gavin, lefties are so paranoid. Imagine thinking that the NSA would ever listen to our phone calls! President Bush himself said today that he would never allow spying on “innocent Americans”, so that should settle that!

By the way, regarding that 900 number you keep calling. When the lady who answers keeps calling you “Saucy Boy”, that’s not some kind of code name, is it?


My assumption has been that they just sort of collect everything they can, emails, phone records, whatever, and don’t actually bother to look at it unless something happens. Then they have a practical usage: they know what general time period to look for information, they have a good idea of who was involved that needs looking at, they have a better idea of what *kind* of conversations they should be looking for.

I see it as a law-enforcement tool more than anything else. It won’t *prevent* a damned thing, it’s all too complicated and massive. But it will help them put on a good show trial if anyone’s still around.

Of course, they also then have a little pile of pre-acquired info if you later draw yourself to their attention in a way that has nothing to do with terrorism, say, by asking the president if he ever feels ashamed of himself or attending animal rights meetings or whatever.


Oh, you should certainly be able to do that. The basic structure is a way of prioritizing people; it doesn’t (or shouldn’t) prevent searching the database in other ways.


Just what do we know about this Colbert fellow?

Run the logs. I want a report on my desk in the morning.


There are all sorts of mining possibilities. A lot of it comes out in how to display the data, how to make things visually apparent that are too complex to discern in the raw data. For that much data any viewing program would need some massive power. Also, extra dimensions would help.


That is my thoughts also. The log data can be mined easily and with more effect than the actual content of the data.

For instance, a company is more interesting in their own web log data than the actual content of the web pages they serve you.

The data is OLTP data and is not just collected, it is put in the warehouse (OLAP) and most likely mined. This same sort of thing was proposed as a trial balloon in Total Information Awareness, which is why both the right and the left cut its funding. These guys just put it into practice without testing the efficacy using the one agency with enough resources to accomplish it, but least expected to implement it—the NSA’s Golden Rule is to not spy on American citizens.

Businesses do this sort of thing all the time. As Gavin points out, it’s sort of like Friendster (I prefer to use MySpace and FaceBook as they’re more profitable), but from the perspective of the actual companies, not the users.

These companies make a lot of money because their database (mined from the connection data) has a lot of value to advertisers and advertisers want to sell into that. This is why FaceBook is profitable and MySpace was bought for half a billion dollars.

Ignoring the legality, the difference is the government database would be larger, more connectively rich, and more consistent than any done by a single private entity. As that entity is a large utility (like the phone company), laws are placed on them to prevent them for this reason to ensure the privacy of its citizens.

Many problems in this world are homomorphic to such a database. One example is tracing likely terrorists. Another is figuring out which companies are which in order to give a competitive advantage to one party. Another is finding political ammunition against your enemies (say, deducing who is a closet homosexual and who you could talk to to “out� him). Another is using it to build more accurate election and voting models. Another is disseminating a wedge issue among the parts of your base who will be most supportive of it in order to manipulate their vote.


they just sort of collect everything they can, emails, phone records, whatever

Can you imagine the size of the heat-sink that would be needed on a hard-drive doing a task like that?

Hrm… you know I heard a stat recently that 1 in 3 homeless men are veterans, and I also know that the streets can get mighty cold during winter. Now if you put two and two together, not only could we solve this coldness problem, but I can think of no better tribute to the veterans and the freedoms they fought for than to have them warm themselves with the heat generated by a domestic surveillance program.

Should I email Bush directly, or just say “terrorist osama peace activist”, and trust he gets the message?


Actually the NSA needs all that data to find the ELS codes that will lead them to the terrorists.


Hell, I’d be curious to hear what convicted felon John Poindexter thinks about Son of TIA.


Well, it’s not set up the same way as TIA, but it’s pretty close in practice.

It ought to have its own Internet-type acronym, like ‘TIA.’ Maybe the OMDFG! program?


I really think it’s time this administration goes the way of the dinosaurs and gets buried by God to test our faith.


TIA was never “defunded.” Some few small portions were(like the bit to protect privacy), the rest was simply transferred to the classified part of the budget to remove it from public scrutiny – the “defunding” was a shell game. This has been reported multiple times, and I personally heard it from Poindexter’s own lips.


I have nothing intelligent to add, so I will just comment that the enemyster graphic is oh so cute. Tee hee! (I’m giggling like an airhead, if that wasn’t clear.)


That system is actually a lot more clever then I gave the government credit for, when the data mining rumors were floating around months ago, I imagined it as a keyword trigger system, which would be effectively useless.



If you were really high up on the list of people to observe, what would it do to the list if you started spamming people?


I think Net Neutrality is going to make it easier for the government to spy on us. If you want the government closely monitoring the actions of Internet users, of course you would put a Net Neutrality bill into place so they can have more oversight over Internet activity, no? This is what’s worrying me now.


http://lipanabsketball37.blogspot.com verizon free ringtone verizon free ringtone verizon free ringtone http://lipanabsketball37.blogspot.com verizon free ringtone verizon free ringtone


what i want to know is the beta code of it…how could we enter the code?…i really want to know the code^^


(comments are closed)