« The (Opportunity) Cos… | Home | Even More Ways to Use… »

Two Supercomputers Diverged in the Woods

(Click here to read the entire “The (Opportunity) Costs of Security” series.)

Quick question: who, in the whole world, has the most powerful supercomputer?

Answer: The U.S. Government.

Second question: What do they do with it?

Answer: Collect, store and analyze internet traffic.

For details, read this prescient Wired article by author James Bamford that describes how the NSA is building a gigantic facility in Utah to track, store, crack and read a huge chunk of the communications passing through the U.S interwebs. To help decrypt and understand all this information, the NSA is trying to build the world's most powerful supercomputer. (If you had read this article pre-Snowden leaks, well, Prism and XKeyscore weren’t stunning revelations. Those leaks, though, confirmed what many of us suspected but couldn’t prove.)

As I said in the intro to this series, I don’t want to frame the NSA’s meta-data collection in terms of privacy and civil liberties, but in the terms of economics...and loss. The computing power purchased and harnessed by the NSA has opportunity costs. By using the world’s most powerful computer to track terrorist communications, we’re not using it on other things. Things that could save more lives.

For instance, when the NSA and Congress choose to spend billions on designing and building supercomputers for eavesdropping on our phone calls and emails, we choose to not use it on cracking the human genome. Genetics scientists have become much better at decoding individual human genomes. However, doing extensive analysis requires--you guessed it--massive amounts of computing power.

(For the sake of total honesty, the U.S. may not currently hold the title of “World’s Fastest Supercomputer”. The super-computer rankings are constantly shifting. Either way, the U.S. still has several very, very powerful super-computer systems dedicated to stopping terrorism.)

So I ask you, what would be a better use of the world's largest supercomputer, trying to prevent a handful of attacks a year by terrorists, or developing medical innovations that could come by feeding the world's fastest supercomputer human genome data?

To be safe, let's run the numbers.

Assume that without the supercomputer, Al Qaeda would have the ability to conduct three 9/11-sized attacks every year. (This is, of course, a ridiculous exaggeration.) That's around 9,000 people a year saved by the supercomputer. (Again, this is a ridiculously high figure, approximately twice as many people as the actual number of Americans who have died by international terrorism since 1969 or roughly 44 times more than the average number of people who have died of terrorism since 1969). The average age in the U.S. is 36.8. The life is expectancy is around 77. That means that the computer saved around 361,000 U.S. life years. (I mean, it didn’t, but go with us here.) Go NSA supercomputer!

Let’s assume that instead of stopping terror attacks, we use all that computing power on decoding the human genome. Let’s say we target it at breast cancer alone. Let’s say this develops a cure for one form of breast cancer, or approximately 12% of the 232,000 newly diagnosed cases every year. The average person lives 5 years after a breast cancer diagnosis. Let’s say this adds on an additional 12 years. This would assume the average diagnosis age is 60, and the average survival is now normal life expectancy. (It’s a cure, remember?) This supercomputer, on a very limited and conservative estimate, just saved 334,000 U.S. life years.

Don’t think too hard; these are just rough back of the envelope calculations, filled with assumptions. Just kicking the tires a tiny, tiny bit, though, shows my thesis holds up: we are wasting computing power on stopping terrorism. That assumption of 9,000 U.S. lives saved by terrorism is ridiculously high. Ludicrously high. If the government’s computer could save even a fraction of lives by decoding the human genome (the break-even in life years), then we should use it for that purpose.

(And I just used it for breast cancer, not other cancers, or heart disease or diabetes or infectious diseases or a host of other illnesses where genetics play a part. We didn’t even mention unfolding proteins, which is both medically useful and the perfect task for a supercomputer.)

The moral? Don’t let counter-terrorism advocates fool you into thinking spending money stopping terrorism saves your life. Wasting money on supercomputers to stop terrorism is killing you.

four comments

Great points, and not the first time they’ve been raised:

http://harpers.org/blog/2007/11/eisenhow..

Cost-effective resilience to low probability, high impact attacks is a much better course of action, and it seems like more and more thoughtful people are giving credence to this idea that was advocated by John Robb in his book Brave New War.


To put your numbers in better perspective, a study mentioned by CNN (http://www.cnn.com/2014/04/14/opinion/bergen-sterman-kansas-shooting/index.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed2Fcnn_latest+3A+Most+Recent%29) noted that “since 9/11 extremists affiliated with a variety of far-right wing ideologies, including white supremacists, anti-abortion extremists and anti-government militants, have killed more people in the United States than have extremists motivated by al Qaeda’s ideology. According to a count by the New America Foundation, right wing extremists have killed 34 people in the United States for political reasons since 9/11. . . By contrast, terrorists motivated by al Qaeda’s ideology have killed 21 people in the United States since 9/11.” It also notes that left wing groups have carried out a number of attacks on infrastructure, but haven’t killed anyone.

So:
Your figures are absurdly high (as you note).
The focus on Islamic terrorism is actually making the US more vulnerable to other, more probable, violent threats.
Given that the right wing attacks cited include the recent Kansas attack that targeted Jews and killed Christians, and the Wisconsin attack on a Sikh temple in the mistaken belief that it was a Muslim site, it seems the lack on spending on effective education is proving as dangerous as terrorism (on a related note, SO has a link on his blog to a great piece about where people think Ukraine is. Sarah Palin is in for a nasty surprise.).
If a country can tolerate 33K traffic fatalities a year and 14K murders a year, perhaps it can accept a very small amount of additional random violence and simply get on with business. But that would fly in the face of the CNN-effect on politicians. It would also probably require a year of zero-based budgeting, and hell hath no fury like a lobbyist scorned.


Jim – Thanks for the links. Yeah, we aren’t hte first to raise this point (other good thinkers on it are Micah Zenko and John Mueller). And a true story is that when we drove to Kentucky from Arizona for my fourth duty station, we stopped at the Eisenhower library and I bought a copy of “The Chance for Peace” speech. It is fantastic.

F- Yep, agree with all your points. The figures are ridiculously high, and even then they aren’t worth it. CNN-effect is a great name for it.

As for SO, I have meant to email him for weeks to talk about doing a link drop of his fantastic pieces on national security and economics.


It use to be a sign of paranoia to believe the government is watching you. That’s not true anymore. The popular news is fill with examples of our government forced to admit that yes, they lied about it and they are documenting our activities.

This has become so mainstream we are even entertained by a television show whose premise is the government is watching. One finds examples of traffic camera at minor intersections that seem to have little impact on the tide of daily traffic. Why? Fiction and daily life are on a collusion course.

One of the biggest comments in the 1960’s on Orwell’s 1984 was “Who watches the watchers?” This pre-computer analysis suggest this fictional society would be impossible as increasingly larger and larger numbers of people would be required to supervise each layer of watchers. Soon everyone would be working for the Ministry of Truth. Who would make consumables?

Clearly computers and software eliminate this criticism.

So it seem, to me it’s not the opportunistic cost of not being able to solve the riddle of breast cancer that we should be concerned with but the continual and often unjustified intrusion into our private live.

stay safe……..
Frank