In response to yesterday’s expressed disappointment with a Guardian column that challenged the original post on the NSA controversy here, Clay Shirky, who authored that essay, showed up in the comments section this morning. That’s a mensch. All credit there. I’m going to highlight his comments and my answer to continue to focus my argument here. And perhaps maybe move the argument off of who is Simon, or Greenwald, or Andrew Sullivan, or Thomas Friedman, or Janet Mayer, or Michael Moore, or Glenn Beck, or certainly, Mr. Shirky himself, where they come from or what they might or might not know given what we think we understand of their background. That stuff does not improve argument; it weakens argument. Content matters. The arguments themselves matter, not who we think made them or why. Come to think of it, everyone just now arguing about who Edward Snowden is might similarly inoculate themselves against the great viral scourge of argumentum ad hominem, and focus, instead, on what Mr. Snowden has actually revealed — both good and bad for the republic.
Anyway, from Mr. Shirky:
I’m sorry I’m just getting to this now.
Let me apologize for over-reading “Allow for a comparable example…” from your earlier piece, which I presumed to be a more general comparison between the Baltimore cops and the NSA than you obviously meant it to be. Duly noted, and I won’t refer to that comparison in the future.
Let me also agree with you that I remain “…mute to the great questions here, which is how do we reconcile privacy with security, personal liberty with collective responsibility.” That is true; those were not the questions I meant to address.
I also understand that your belief that if what the NSA is doing is legal, it should also be acceptable to the American people, is a political assertion, and thus not one that can be right or wrong. Though it is one I disagree with, I know your views are likelier to carry the day than mine. (Indeed, my preferences on civil liberties have not been terribly well represented in the national debate any time in these last dozen years.) Such is life in a democracy.
And I wholly concur with your recent statement in this thread:
“The computers are big and powerful and capable of doing things — good and bad — that we once couldn’t do. No shit.”
This is of course true. However, I can’t find any version of this sentiment in “We are shocked, shocked…”.
I also can’t imagine any plain English reading of…
“Because that is tens of billions of phone calls and for the love of god, how many agents do you think the FBI has? How many computer-runs do you think the NSA can do — and then specifically analyze and assess each result?”
…that suggests anything other than a belief that the NSA’s surveillance capabilities can be usefully likened to those of the FBI, and that those capabilities are constrained by technological capacity or manpower in a way that should make us worry less about the potential for uses of which the American people would not approve.
Even to speak of “computer-runs”, with its sense of batch processing and mainframes is off the mark, and the FBI is famously a technological laggard (google ‘virtual case file’), not a problem the NSA has. Neither the tools the FBI has access to, nor the constraints they operate under, make for good comparisons. (The NSA is only directly comparable, in my view, to the NRO, Air Force Cyber-command, Secret Service, and CIA in terms of tools and techniques, and the scale of NSA resources is larger than those organizations.)
Similarly, the change in what constitutes meta-data, not just for Prism but for phone calls, means that the upper limits on human attention are quite different between law enforcement and the intelligence community. If you say “Here are 100 hours of David Simon’s phone calls. How long do you think it will take to find out if he said “Brooklyn Bridge” in them?”, the answer from traditional law enforcement will be “50 hours, on average.”
Once you have a machine that can search human speech, however, the same question question can be answered in seconds. The model you seem to hold out — first the computer works, then the human — doesn’t describe the way the tools can be used today. Computing power serves as a partial substitue for human attention, so that the amount of material a human agent can cover is far larger, in a place like the NSA, than it is in a place like the FBI.
The great questions are elsewhere, as you note, but this narrow point intersects with those questions in one critical way: the balance between power and oversight is affected by how significant the powers are. The executive branch of the US government has a poor record of using its powers against political dissent, from union busting, through the civil rights movement, to COINTELPRO, to surveillance of political movements like #Occupy today.
In “We are shocked, shocked…”, I think you under-represented those powers, in a way that would make it harder to understand what sorts of checks and balances the NSA should be subjected to.
Dear Mr. Shirky,
Pleased to make your acquaintance. Thank you for venturing here for a direct discussion as to my disappointment with your essay.
Thank you, too, for acknowledging that that the Baltimore example in my original post actually had a different substantive purpose than the one you claimed for it in your essay. Certainly, no sentient being would think that what police anywhere did with a Title III wiretap in the 1980s could be used to speak to the computing capabilities of American counterintelligence thirty years later. That would have been just silly. And of course, I was distressed to be considered so marginally on such little cause. Your application of the example to your own purposes in your essay — that of chronicling the power of the NSA computing capability — is at least suggestive of a certain propensity on the part of all of us to misshape the argument of others into a different one, an argument that we very much wish to make.
When confronted with such a dynamic, my uncle, the esteemed and learned Irving Ligeti, used to remark that he knew for sure who shot President McKinley and who shot President Garfield. And goddamit, he was going to work it into the conversation come hell or high water.
My comment as to what you remain mute to was, of course, meant in corroboration of the aforementioned dynamic. Meaning, my original post was addressing other things that, in changing the subject, you did not address or wish to address. It is certainly your perogative to write about what interests you, but given the fundamental ethical and political conflict between privacy and security that is at the heart of this controversy, it is at least fair to note what matters of interest went unaddressed in your own effort.
With regard to your next point, you cannot find a sentence as such as “the computers are big and powerful and capable of doing things — good and bad — that we once couldn’t do” in my original essay because such a sentence would be, at this late date in the history of the computing revolution through which we have all lived and experienced, worthy of Candide’s first trip to Paris. What you would have found, in my original post, is this simple premise, acknowledged:
“Privacy is in decline around the world, largely because technology and big data have matured to the point where it is easy to create a net that monitors many daily interactions.
That’s there in black and white, italics now mine because you’ve ignored it twice now. It is not a paragraphs-long aside on just how wicked powerful the NSA computers are, but I certainly deem it sufficient to invoke a rhetorical legalism that all of us recognize: Stipulated.
In my original essay and elsewhere, it is stipulated as to what the NSA’s computers can do. It is one of the givens in this ethical, legal and policy problem, and yes, a notable one. But with so much on our collective plate to discuss in terms of what the NSA program might or might not mean to privacy or civil rights, with its potentialities for counter-terrorism, a long-winded dissertation along the lines of no-I-mean-really-really-powerful, such as you offered in non-rebuttal rebuttal of my essay seems, well, absurd.
Which brings us to your next disembodied shard that to you suggests my Luddite ignorance — the fact that I asked how many computer runs the NSA can do? There is an ambiguity in that sentence taken alone, I do agree. But backed hard against the paired question of how many FBI agents do you think we have, I feel that the crux of my argument is clear. If you understand law enforcement — whether it’s legitimate law enforcement or an effort by agencies to unfairly target and violate the civil rights of citizens — you know that the computer run is not the resource issue. The time and manpower are. And with every computer result, if Americans are going to be targeted off the basis of billions and billions and billions of calls — there will have to be a corresponding allocation of finite investigative resources. Taken as a whole — what a concept, I know — my argument is more than an opening for more dissertation on computing capability. My argument is this: Why would the government use this huge metadata pile for that?
Go back to the real point, again, of my Baltimore example and that question is already answered for us: It is legal already for this data to be obtained, without warrants, without standards of probable cause required for warrants. Metadata has been ruled legal for ordinary law enforcement purposes from the administration of Jimmy Carter forward. The point here is obvious and from your subsequent reflections on my cite of the legal standard, it’s clear that it has either eluded you or you have ignored it: If they want you, in the manner that you want Americans to worry about,<em> they’ve already got you.</em> There is no blessed reason for an investigator to use this particular datapile at this particular facility in Utah to target individual Americans — let’s say, the phones of news reporters, for example. They already have ready capability to grab such records whether this data pile exists or not. And in fact, given the project’s counter-terror value and the level of national security secrecy, there is actually a practical disincentive to expose this program by using its data to undertake casework in which the chain of evidence could, ultimately, lead any future controversy back to this top-secret project. No, if the government wants to use your phone metadata to mess with you, then a run-of-the-mill FBI agent walks his request over to any run-of-the-mill U.S. District Judge and gets a court order for material that, again, has never had constitutional protection. I understand that the FBI doesn’t have the computer capacity to navigate a mass of megadata as the NSA can, but such capability isn’t required to purposefully target large numbers of Americans using phone metadata. That stuff is obtainable from Verizon in minutes once you walk into their security office with a subpeona or court order. And if you think the data pile opens us up to fishing in metadata on a scale of 300 millions of citizen-targets, well then, you’re back to the questions of priority and manpower, which are the ones I raised in my essay. For every computer outcome, regardless of how easily arrived, a human being has to take the time to do the work to mess with someone.
I understand sufficiently well the linguistics science that has gone into NSA computer capture technology. The NSA language capture capability has been well reported by my old newspaper — Ft. Meade lies just below Baltimore — and in numerous publications for decades now. We’ve all read “The Puzzle Palace” and we all know who Mr. Binney is. Please, Mr. Shirky, no more stooping to lecture on the givens. But your “Brooklyn Bridge” example begs two questions: Is the NSA computer-searching recorded content, because that is not what the FISA-signed court order suggests. Recorded content on phone does have Fourth Amendment guarantees of privacy, and if that judge was signing off on content, I would think he is not signing a court order for metadata, he’s signing a wiretap warrant. Moreover, you are the first person I have heard suggest that the NSA data pile contains recorded conversations. Please resolve that. I do not believe that Verizon actually has the capability to even record all of our conversations, much less turn over recorded content of all its phone calls, even if the government asks. I believe the government needs to achieve its own telephonic intercept through wiretap for such, which is again a higher standard of probable cause. And to further point out where your “Brooklyn Bridge” example leads: Yes, the NSA can singularly and quickly determine who is talking about the bridge more than ordinary. But then, still, a human being has to go out on the street and research the why. The why is where the resources become more finite. How long for human resources to assess the data, confirm its connection to a human target, assess that target, determine its interest in the Brooklyn Bridge and move on to the next computer outcome? It’s Orwellian, sure. Scary, even. But while that imaginary scenario is going on, there’s also an awful lot of actual phone traffic between the United States and overseas areas of concern to which the system must attend.
But of course when you cite my computer-run language in isolation, carefully eschewing the manpower question that precedes it — this allows for a return to your one-trick pony. Simon knows nothing of the brave new digital world. Ergo, why would he opine as he does? And how easy it is for us to disregard the rest of what this abacus-toting ass actually argues. It’s worse than a red herring. Call it red flounder.
As to the remainder of your post here, I am in agreement that legality is not morality. My point in raising legality is not to equate it with what is necessarily just or moral. You are mistaken if you think so. Again, I am sentient. Segregation was once legal. So is drug war, of which you may be aware, I am a rather implacable opponent who urges jury nullification in non-violent offenses every chance he gets. My point in citing the legality is to say, in the wake of the Guardian’s headless-chicken act, that Americans have in fact been tolerant of court-interpretations on the use of phone metadata going back through three decades of law enforcement use. And to point out that this stuff has never required a warrant from law-enforcement. Again, that is me stating a given, not using the legality for some sort of circular argument.
In truth the two unique aspects of this data haystack that the NSA constructed are these:
1) It is vast and includes, I am going to assume from the Verizon court order alone, all American phone traffic, and…
2) It pre-exists the discovery of specific, suspected phone numbers that are to be run through the haystack in search of a needle.
Elsewhere on this site (this is long enough already) there are detailed arguments — arguments that couldn’t actually be made by someone living in ignorance of NSA capabilities or methodology — as to why these unique aspects are present and, perhaps (we will soon see), legally justifiable. Briefly, by creating the biggest haystack possible, by NOT excluding anyone from the data base, this program is actually, in practice and with regard to individual Americans, one of the least invasive and most discriminate acquisitions and uses of metadata I have ever encountered in all my years of considering, observing and writing about law enforcement. If used as the counter-intelligence community has asserted in seeking the FISA approval, and in briefing all branches of government as to that purpose, then the fact that all Americans — grandma, the fourteen year old, the Republican god-fearing farmers of Nebraska — are included in the data pile and that no human knows the identity behind any single call, or looks into any shard of data, or segregates the data by ethnicity, by politics, by religion or whatever, it is, in practice, much less invasive of individual privacy. Only when the wrong number in Chechnya calls someone is the identity of a citizen suddenly germaine; the identities of all others remain unknown, unlike the reality of every other numbers dump or DNR that law enforcement has ever done.
As to the pre-existing nature of the data pile: I am guessing that if we ever see the material in support of the request for a court order, the government will have offered the following: The average shelf-life of a cellphone in the echelons of a competent terror organization is, at best, days. And the time and bandwidth required for us to have Verizon access this much data, ship it over to us in Utah, and for us to get up and running is, well, a bit longer. By then the cell phone is dumped and the opportunity for pro-active, real-time intelligence disappears. Legally, it’s called an exigent circumstance and it is an effective and germaine legal point.
To your simple statement that while legal, Americans must agree with it, I say precisely so. Would that the Guardian’s initial reporting, rather than editorializing on the “indiscriminate and sweeping” nature of the NSA program had contained even as much context as the above two paragraphs, I’m actually pretty damn sanguine as to the result. We’re a generally sensible people once all the facts and context are laid out, and full-blooded arguments and counter-arguments have their day. Most of us are, anyway.
As to your final point about the government’s history for abuse of such data, I again stipulate readily. In fact, I concur fully. And in the original post, I acknowledged the potential for misuse, even noting such misuse at some point, by someone — as with any law enforcement asset — is likely. In fact, I ventured further in the original post further to express concerns about how the secrecy of the entire FISA-process is problematic in preventing the evidence of such abuse from coming forward for systemic redress. You quoted none of that, regrettably, but mischaracterized my even-handed language as being this: Simon thinks we should “trust” the NSA. I said nothing of the sort.
“All governments lie,” said I.F. Stone, and he was dead right, then and for now and for all time. But Mr. Stone also understood that governance is nonetheless necessary, and so while trust is perhaps too indulgent for any thinking citizen, a measured and discerned attention to all of the costs and risks of any and all action and policy becomes the real intellectual currency. That’s what I’m trying to spend here. Trust was off the table a long time ago. But again, your simplification of my stance certainly would suit the agenda of someone determined to paint someone as a political and technological naif, while eschewing any interest in his actual arguments.
And therein lies all of my disappointment with your response to my original post in the Guardian.
Let me add, that I can only assume that Mr. Shirky understood himself to be replying in a comments section and so was more brief than he might have otherwise wished. Comments are generally more circumscribed than what can be accomplished in a full post, and so, indeed, my reply is longer. To that end, whatever else I receive from Mr. Shirky in further reply will be appended below this, to any length he requires. And further, I will accord Mr. Shirky the last word here on this post. He earns that by engaging openly and directly, and I credit him again.