Andrew Cockburn is one third of a famous Irish journalist trio of brothers (along with Patrick and the late Alexander), and an intrepid reporter in his own right. He’s covered Iraq, both before and after the 2003 invasion, the Colombian cocaine cartels, the Red Princes that rule China, the rise of modern slavery, blood diamonds in Africa, and the covert U.S.-Israeli defense and intelligence connection. He co-produced the 1997 DreamWorks feature movie The Peacemaker, and American Casino, the 2009 feature documentary on the Wall Street crash; and, for the Gawkerverse, he’s the father of actress Olivia Wilde.

For historical buffs, he’s also reputedly related to Sir George Cockburn, the British admiral who ordered the Burning of Washington during the War of 1812. Andrew is Washington Editor of Harper’s magazine and has a new book out on drones. I’ve known him since the 1980s and asked him to answer some questions. He will be in the comments at 1 pm eastern under the username “andrewmylescockburn” to answer your questions.

Can you tell us a little about why you wrote Kill Chain and what you did to write it? Where’d you go? Who did you talk to? Did you start with any preconceptions that were challenged by what you learned?

I wrote Kill Chain because there was so much about the drone program that wasn’t being said. Most of the discussion seemed to be about legal issues: “what legal right does the President have to do this?” and humanitarian concerns – “what about the civilians being killed?” I’m not saying these aren’t important, they are, but no one seemed to be looking at whether drones and the systems of which they are part actually work as advertised.

For example: do they really have that 20/20 vision as displayed in the officially released videos you see on Youtube, and so on? Answer: no, more like 20/200, at least for the IR sensor – the legal definition of blindness for drivers according to most state DMVs. Nor was anyone looking at the basic strategy of assassination, “high value targeting” if you prefer, enemy leaders, and whether it was really helping us win the war and make the threat go away.

Also, there was this common idea that everything about drones and the strategy they serve is totally revolutionary and new, which is not true. As I relate in the book, the dream of having total intelligence about what is going on in the battlespace and therefore being able to select and strike exactly the right targets goes back at least to Vietnam and the fiasco of Igloo White – the electronic fence of sensors that was supposed to detect movement on the Ho Chi Minh trail, identifying targets that could then be hit. Igloo White was a total bust - half the North Vietnamese army marched through it in 1972 without anyone being any the wiser. Today’s sensors, electro-optical, infra-red, radar, are carried on drones, among other things, and so move over the battlespace to look and listen (for signals intelligence or SIGINT). But the basic concept is the same: looking at the enemy from a distance, assuming you can thereby understand what’s going on, and thus hit precisely the right target to achieve a desired effect.

One of the first people I spoke to was Timothy McHale, who as an army two-star general had done a fantastic job of investigating the February 21, 2010 incident in Uruzgan, Afghanistan, in which 23 civilians died. His report, largely declassified, gives us extraordinary insight into the way the system works, and doesn’t work. People had already commented on the callous attitude of the Predator crew responsible for the strike, (McHale himself was shocked by this) as revealed in the transcript of their conversations, but McHale’s report went much deeper than that, revealing that the vast electronic network linking drone crews with other components of the system, such as the armies of drone video analysts sitting in the center processing stations of the Distributed Common Ground System (DCGS), as well as multiple headquarters in the chain of command, with everyone looking at the same pictures and communicating with each other via internet relay chat or whatever and nobody having the faintest idea of the reality on the ground.

It was a perfect example of how complex technology constricts us. One example of this really hit me: reports of the fatal strike with multiple civilian dead took four and a half hours to make it through the military’s secure email channels from Kandahar to Kabul. The Taliban had their version all over Afghanistan in an hour.

One important source that I do name in the book is Rex Rivolo, who while working at the Institute for Defense Analysis (the Pentagon’s in-house think tank) had figured out that high-value-targeting is directly counter-productive. He’s a former F-4 pilot who flew over 500 combat missions in Vietnam and a brilliant statistician who devised empirical methods to prove the point. First, in the 1990s, he showed that targeting the leadership of the Colombian cocaine cartels, promoted by the DEA as the “Kingpin Strategy,” actually increased the supply of cocaine to the U.S. (by promoting competition.) Then in 2007 he was sent to Baghdad, working in a small, high-powered intelligence cell reporting directly to the ground commander and there he did the numbers and proved that targeting insurgent cell leaders actually increased attacks on Americans, for reasons I describe in Kill Chain.

A lot of things I discovered ran counter to preconceptions. One of them was the fallibility of the drone sensors. I had assumed that the drone-acquired videos released to the media, or featured on GWOT movies like Zero Dark Thirty, were probably accurate representations of the real thing. Not so.

I managed to lay my hands on the original DOT&E (Pentagon weapons testing office, loathed by services and contractors) report on the Predator. It had never been publicly released, apart from the executive summary, and when I read it I could see why. To grade the sensors, the testing team used something called the national imagery rating scale, which runs from 1, defined as the ability to pick out a big airplane like a 737, to 9, the ability to pick out a human face. The Predator, which was supposed to score a 6 (at 15,000 feet slant range), which is defined as the ability to identify vehicles, could only manage a 2.7, “detect bridges.” Of course, at lower altitudes and very good weather it could do better, but even in the famous Predator video of Tarnak Farms in Afghanistan taken in the fall of 2000, which allegedly revealed Osama bin Laden out for a walk, all you can see is a white dot, which becomes even less clear when you blow it up.

Naturally, the claim is that things have vastly improved since then, but it doesn’t seem to be that way. Sensors now in service still can’t tell children from adults when they are on their own, a problem known in the business as “slants.” In 2011, in Afghanistan, marine sergeant Jeremy Swift and corpsman Benjamin Rast were killed because the Reaper video couldn’t tell two Americans in combat gear, complete with helmets, vests, etc, from Taliban in turbans. Nor could they tell even from the muzzle flashes which way they were firing their weapons.

Two dead hostages and two dead American terrorists inadvertently killed in January. You wrote the book on this: How could the weapons be so good and the intelligence so bad?

What this tells us is that after all these years, we have very little human intelligence on the home territory of what is supposedly our number one threat, “core al-Qaeda.” Here was an American prisoner, captured while in the service of the U.S. government, and we had no idea where he was? So we rely on technology to deliver intelligence on “patterns of life,” i.e. human behavior as –necessarily– observed from a distance. The CIA has an enormous database of what are deemed to be tell-tale signs identifying the enemy, thus legitimizing “signature strikes” of the kind that killed the hostages and the American terrorists. One example I give in the book is that Arab males supposedly squat to urinate, while Pushtun men do it standing up. Since an Arab in Waziristan is probably al Qaeda, heaven help anyone who squats to piss! Obviously, or at least hopefully, they use other and more sophisticated indicators, but there’s plenty of room for error. Just look at the wedding parties that have been hit in Afghanistan in 2008 and Yemen in 2013, to name but two – convoys of tribesmen in vehicle, some of them armed, clearly a bad pattern of life, so they were hit and dozens of civilians died.

For years, the most useful intelligence technology in the entire counter-terror/insurgency campaign has been what is variously known as Triggerfish, Stingray, IMSI Catcher, which can be carried in a backpack, or mounted on a drone. Essentially this mimics a cell tower, getting a cell phone or cell phones even when they are miles away to connect, and thereby reveal their IMSI (International Mobile Subscriber Identity) number and location. In the Air Force at least, picking up a targeted cell phone this way was is known as a “touchdown.” As I understand it, a phone previously linked in some way to al Qaeda was tracked in this way to the house where Weinstein and Lo Porto were being held.

It has to be said that the intelligence in this case was partly right: the building was an Al Qaeda hangout. But it can go badly wrong. I describe a case in Kill Chain where JSOC tracked a Taliban leader via his phone for weeks and eventually attacked his convoy and killed him – zeroing in on his phone from a drone mounted IMSI-Catcher. Except it was the wrong phone. They’d somehow got numbers mixed up and had been tracking an innocent guy who was out campaigning very publicly in a parliamentary election. A glance at a newspaper or a call to the local police would have told JSOC that they were following the wrong person, but they were hypnotized by their own technology and so went on insisting that they’d got the right man while their actual target was alive and kicking and giving interviews in Pakistan.

We’ve been hearing about how the “heat sensors” (that sounds like plain old fashioned infra red sensors to me) on the CIA drones (which would actually be flown by air force crews at Creech AFB in Nevada) detected only four people in the house that was hit in January, and missed the two hostages. The suggestion is that the hostages were hidden deep enough underground to mask their body heat. But if that were so, how come they died in the missile strike? A Hellfire isn’t a bunker buster.

Why do you think this story leaked to the Wall Street Journal in the first place? Why weren’t the deaths just announced when they happened?

You’re kidding, right? The CIA would never have wanted to admit this if it hadn’t leaked, and it’s not clear how quickly they fessed up to Obama. The drone campaign had basically faded from the news, and this was obviously going to push it into the headlines again in the worst way. As it is, they successfully held out against any official admission that they were involved. Obama talked only about a “counter-terror operation” in the Afghan-Pakistani border area. No mention of CIA, or drones (though plenty of self-serving leaks after the fact, of course).

The CIA’s drone program, the President’s drone program, Congressionally approved, not approved, tacitly accepted: almost every description of the drone program makes it sound like it isn’t the United States and its foreign policy. Is that the consequence of something unique to drones?

It’s interesting, drones and covert foreign policy seem to go together. In Operation Menu, Nixon’s secret bombing of Cambodia, the B-52 flight paths were directed from the ground, as was the moment of bomb release. In other words, the B-52s were essentially drones. Maybe the drone campaign isn’t described as the foreign policy of the United States because there’s a tinge of embarrassment that we’re murdering people in foreign countries as a matter of routine.

Beyond that, maybe we should call it the drone program’s drone program, because it’s taken on a life of it’s own, a huge machine that exists to perpetuate itself. Just take a look at the jobs listed almost every day for just one of the Distributed Common Ground System stations at Langley AFB in the Virginia Southern Neck. On April 25, for example, various contractors (some of which you’ve never heard of) were asking for a “USAF Intelligence Resource Management Analyst,” a “Systems Integrator,” a “USAF Senior Intelligence Programs and Systems Support Analyst,” a “USAF ISR Weapons Systems Integration Support Analyst” a “DPOC Network Engineer,” whatever that is, and a few others. All high paying, all of course requiring Top Secret or higher clearances. Every so often we hear that the CIA drone program is going to be turned over to the military. I say, ‘good luck with that’ – is the CIA really going to obligingly hand over a huge chunk of its raison d’etre, and its budget, its enormous targeting apparatus? There’s a lot of talk about “autonomous drones,” which aren’t going to happen, but I think the whole system is autonomous, one giant robot that has become unstoppable as it grinds along, sucking up money and killing people along the way.

“The collateral damage has been extraordinarily low,” Sen. Dianne Feinstein, ranking member of the Senate Intelligence Committee said last week. “That is really a fact. I only wish I could tell you what it was. That is the irony here. This is all classified.” What do you make of that claim?

On the face of it, weird. The row about the torture report suggested that Feinstein developed a healthy distrust of the CIA, especially as they spied on her committee (really, a crime just as momentous as the Watergate break-in, except that there was no need for a cover-up because no one cared, certainly not Obama). Asked why she believed what the CIA told her about civilian casualties when she knew they lied about torture, she replied “that’s a good question, actually.” Part of the answer may be that torture was a Bush policy, whereas drone strikes are very much an Obama policy, popular in congress and with the public (though the amp-up actually began in Bush’s last year, when he ok’d signature strikes. We also know that the CIA reduces civilian casualties by posthumously enlisting them as “militants,” even when they don’t know who they are, or were, and maybe she took that claim at face value. For example, the March 17, 2011 strike at Datta Khel, in Waziristan, killed 44 people. Intensive investigations by human rights researchers and journalists indicated that four of them had been Taliban, the others mostly local Pakistani government officials and tribal elders. Yet so far as the agency was concerned, they had all been militants and worthy of execution.

There’s another intriguing aspect to Feinstein’s defense of the drone program. Mike D’Andrea, described to me as “scary,” ran CIA’s Counterterrorism Center from 2006 to earlier this year, and was thus master and avid proponent of the drone program. A Shia Muslim convert who rarely left his office, he was powerful enough to resist efforts by various Directors of the CIA to get him to leave, though Brennan has now finally succeeded. One source of his power was his close relations with… Diane Feinstein, which might explain why she defended the party line on collateral damage.

We are at war, the administration and Congress says, even routine critics of Obama. Doesn’t that mean national sacrifice and something beyond unmanned vehicles sent out to hunt and kill?

This is why our leaders love precision airpower, manned and especially unmanned, not to mention covert Special Forces hit teams. With no national sacrifice, there’s no need to pay much attention to what the people think. It’s part of a trend stretching back to the abolition of the draft after it wrecked the possibility of keeping the Vietnam War going. Desert Storm was another step forward, or downhill, when the public was treated to a TV war with lots of fun video of bombs unerringly hitting their targets and almost no Americans getting killed. Later there was the 1999 Kosovo war, same thing, except no Americans at all got killed. That was the war that really sold liberals on this kind of intervention. Right after that, Bush, as a candidate, promised to “begin creating the military of the next century… Our military must be able to identify targets by a variety of means, then be able to destroy those targets almost instantly. We must be able to strike from across the world with pinpoint accuracy… with unmanned systems.” Sound familiar? Twelve years later, Obama unveils his “strategic guidance” and promises to “protect, and in some cases increase, our investments in special operations forces, in new technologies like ISR and unmanned systems…” In the meantime of course huge sections of the globe have descended into chaos, but who cares?

Last week’s drone fiasco, albeit involving the death of two Western hostages, one an American, is really just the latest in a long line, and isn’t the last. Am I wrong?

No, you’re right. I mentioned the 2010 incident in Uruzgan, but it’s not just drones, we have similar fiascos with manned planes. Remember the Granai massacre in Afghanistan, May 4, 2009? A B-1 bomber killed as many as 147 civilians, mostly women and children. The official report cited “the inability [of the bomber crew] to discern the presence of civilians and avoid and/or minimize accompanying collateral damage resulted in the unintended consequence of civilian casualties.” (I like the bit about “minimize” collateral damage – “killing civilians is OK, so long as you minimize it.”) Just last June a B-1 killed six soldiers – five Americans and one Afghan – because they lacked the means to see their IR strobes. What it comes down to is the widening gap between the operator who pushes the “kill” button and the outside world. The drone operator is sitting half a world away in Nevada, but the B-1 weapons officer at 15,000 feet might as well be twelve thousand miles from the target. He or she is sitting in a windowless compartment, looking at a video screen and marking targets with a cursor. In the F-35, allegedly a fighter plane, the pilot basically looks at the world though an electronic medium. Obama checks off victims on his classified iPad. So at every level we’re waging war through a video screen, and that’s got to end badly.

Update (2:35 pm): We’re going to wrap it up for today. Big thank you to Andrew Cockburn for taking the time to do this, and to you, for stopping by with your thoughtful questions.

Cockburn’s most recent book, “Kill Chain: The Rise of the High-Tech Assassins,” can be purchased here.