The defense of the massive, extralegal, and warrant-less spying on the American people (and peoples around the world) by the NSA has been defended in the name of defense against terrorists attacks—creating the strangest of bedfellows in the progress.
But those planes really did hit those buildings. And that bomb did indeed blow up at the finish line of the Boston marathon. And we really are in a continuing, low-intensity, high-risk conflict with a diffuse, committed and ideologically-motivated enemy. And for a moment, just imagine how much bloviating would be wafting across our political spectrum if, in the wake of an incident of domestic terrorism, an American president and his administration had failed to take full advantage of the existing telephonic data to do what is possible to find those needles in the haystacks.
The massive NSA surveillance program predates the September 11th attacks. The whole ugly mess was in place well before the Boston bombings, the 7/7 attacks in London, the terror attack in Mumbai, and so on. There is not one drop of credible evidence that massive surveillance of the public has done anything to preserve the safety of the public.
There is a good reason to suspect why. Collecting massive amounts of data, particularly automatically collected massive amounts of data, makes the problem of finding meaningful bits of information harder, not easier. The problem goes from finding a needle in a haystack, to finding a needle in a needlestack. The more one collects, the more meaningless connections one discovers.
The tightest metaphor is with earthquakes. Both terrorist attacks and earthquakes follow a power-law distribution—rare very severe events, more common minor events. Despite decades of data collection, we still cannot accurately predict when the big earthquakes are coming—only a rough probabilistic sense of where we should be prepared. There is no rational reason to expect the sort of massive, limitless, data collection by the NSA and other authoritarian governments will be any more able to predict (and prevent) terror attacks than seismologists could precisely predict when the next earthquake will hit Seattle.
A massive pot of data, lazily collected, allows one to make any connection you'd wish. Instead of discovering the next terror plot, you end up discovering your own biases, your own preconceived notions of what will happen. The result of the destruction of privacy (even the idea of privacy) in the United States has not been a safe public, it's been the Iraq War.
Things like warrants, judges (real judges, accountable and open to the public, not the secret kangaroo FISA court) act as a filter, limiting focus, and increasing the chance that the connections detected with surveillance will mean something. That human element, slow, ugly, clunky, and a security risk in itself, is what separates useful intelligence from a mountain of crap.
To borrow from Mr. Simon: The planes did hit the towers. It wasn't the military, the NSA, the president, the congress, the air marshals that stopped the attack. It was a random collection of ordinary Americans, the passengers on United 93, who reacted properly, and brought the attack to an end—at the cost of their lives. The bombs did go off at the marathon, but it wasn't the comically militarized police in Boston who brought the post-Marathon bombing terror to an end. (To the contrary, the ridiculous shutdown of a major US city, and the inept chase after two teenagers added to the terror.) No, it was again a random collection of Americans, volunteers, paramedics, doctors who ran towards the blast, at risk of their own life and limb, to save others, who ended the terror.
I'll take the American attitude, and put my faith in my fellow citizens to deal with the inevitable (the next terror attack, the next earthquake).