LOL!
Call me Mr. Gullible, I never once thought of that.

I don't have any questions for the man, I'm trying to take Orlov's advice and learn to ignore politicians.

Of course Nate, Robert, Jeffrey and the other named posters are protected by their exposure to the public eye. On the other hand, Stuart no longer posts here...

I have no way to know for sure, but NSA has some (secret) capability to seed learning classifier algorithms to datamine the "meaning" of Internet data streams. Classifying Obama's email response traffic would be a trivial subset of all email traffic.

Would NSA actually do that? Would Obama direct NSA to do it? My first impression is that NSA would not. My second impression is that the problem at hand is so important and has so much to do with public opinion that they might do it afterall. Add to that, signals intelligence has become a deeply entrenched, highly reliable feature of our national security apparatus.

See The Spy Factory at PBS.com.

Transcripts

http://www.pbs.org/wgbh/nova/transcripts/3602_spyfactory.html

Here are some parts of the transcript.

BRIAN REID: The most curious piece of equipment in that room is a completely flexible monitoring system that can be told on a moment's notice, "Please monitor all conversations that contain the word hummingbird. Please monitor all conversation that goes to Mobile, Alabama. Please monitor all conversations that contain both the word hummingbird and go to Mobile, Alabama."

NARRATOR: NSA has turned its giant ear to listen in on America.

BRIAN REID: Based on everything I know, I believe that there are between 15 and 30 of these secret rooms around the U.S.

NARRATOR: The post-9/11 rules authorized NSA to listen in to Americans both inside and outside the U.S., without any special court approval.

NARRATOR: Calls and data from the Middle East and North Africa are collected and relayed to a listening post, tucked in the hills, outside Augusta, Georgia.

NARRATOR: As NSA began tapping in to fiber optic cables as well as satellites, information began to flood in like never before.

According to a Congressional study in 2008, some intelligence data sources grow at a rate of four petabytes—that's four quadrillion bytes—per month, the equivalent of 12 filing cabinets of new information for every American citizen, every year. But what does it all mean?

ERIC HASELTINE: Computers, today, tell people what things are: "Here's some data that you asked for." They don't tell you what it means. So there is some work going on to try to marry the power of computers to the power of humans.

NARRATOR: Specialized software can help extract important information based on context and meaning. Dr. Robert L. Popp does advanced research on these kinds of programs, known as classifiers.

DR. ROBERT L. POPP: Say you wanted to build a classifier for Al Qaeda, the term, the concept, Al Qaeda. The way it would work is you, as an analyst, would go find all these documents—whether they're emails or things on the web or whatever—but all these documents that in your judgment are narratives associated with the concept of Al Qaeda.

NARRATOR: In the future, by refining the software and harnessing enough computing power, these classifiers could potentially reduce the mountain of information human analysts have to examine.

The concept of clustering categorical data by means of a learning classifier algorithm is effectively visualised in Chapter 5 "Intelligence Overload".
http://www.pbs.org/wgbh/nova/spyfactory/program.html