The Pentagon’s Brain: An Uncensored History of Darpa, America’s Top Secret Military Research Agency 
by Annie Jacobsen.
Little, Brown, 560 pp., £12.99, September 2015, 978 0 316 34947 5
Show More
Show More

The development​ of a nuclear explosive device and two air-deliverable fission bombs by the Manhattan Engineering District of the US Army Corps of Engineers cost $1.845 billion, equivalent to the cost of a mere nine days of war. A much happier, and infinitely cheaper piece of research that also turned out to have world-historical impact was the development of a digital network between computers with TCP/IP communications protocols, better known as the internet. When the student-programmer Charley Kline sent the first instantaneous message (you had to print it out to keep it) on 29 October 1969, he inaugurated a new era; some months later Ray Tomlinson invented the first email program, using an @ address where such messages could linger.

That first computer network was funded by Arpa, the Advanced Research Projects Agency of the US Department of Defence. The same outfit, now known as Darpa, with ‘Defence’ tacked onto the start of the acronym, has achieved many other startling things over the years, from its support for the development of America’s first plastic and aluminium rifle in the early 1960s – actually a heroic struggle against the US army’s obdurate use of heavy steel and wooden stocks – to the development of Transit, the direct predecessor of the GPS satellite system that allows pilots and car drivers to find their destinations automatically, and enables the existence of the new self-driving cars, in addition to its many military uses. Another Darpa project was the Aspen Movie Map, a virtual tour of Aspen provided by the first hypermedia system (four cameras rigged on the top of a car taking pictures every ten feet). This technology was used to develop large-scale combat simulators which familiarise entire battalions with the place where they’re about to be deployed, and is now in everyday use. Other Darpa creations were more straightforwardly military, from penetration aid decoys that help ballistic missiles defeat interception attempts, to all manner of remotely piloted and robotic vehicles, including reconnaissance aircraft small enough to resemble large mosquitoes.

You might think there’s nothing surprising in all this: all sorts of scientific advances should be possible given the ample funds provided to this programme by the Department of Defence. But that’s the problem: the funds available are not ample, they are very modest by Pentagon standards. In 2015 $2.87 billion were allocated to Darpa, 0.47 per cent of the year’s total defence spending, and its staff of 220 are a tiny band among the 700,000 civilian employees of the Department of Defence. How can technology-crazy America allow such a miserly allocation of funds and people?

The short explanation is that most of the money is reserved for the pseudo-innovations pursued by the uniformed services: the navy’s supposedly ultra-new aircraft carrier that retains an unchanged 1960s configuration; the F-35 jet fighter that offers thirty-year-old ‘stealth’ as its cutting-edge novelty; the new army tank that still looks very much like the 1944 German Tiger. The dominance of this sort of pseudo-innovation is a direct result of the composition of the US armed forces as an alliance of proudly separate services, each with its own traditions, institutional culture, career paths and – most important – iconic weapons.

The Pentagon’s RDT&E (research, development, test and evaluation) money is not allotted to Darpa, or some other defence-wide organisation that might develop new weapons, systems or platforms capable of delivering really major advantages, such as remotely piloted vehicles (drones). Instead, the money is parcelled out to the separate services, each of which spends almost all of its share on enhancing its own military role, and its own identity, by expensively updating the weapons traditionally associated with it. The US army spends most of its RDT&E funds ($6.5 billion in 2015) on armoured vehicles, as well as the armed helicopters that it has been using for decades. The navy’s RDT&E funds ($16 billion in 2015) are spent on aircraft carriers and submarines, except for the portion controlled by the Marines, whose favourites are the landing ships and beach-crossing vehicles associated with its amphibian vocation – even though there has not been one opposed amphibious landing in all the wars Americans have fought in the 65 years since the Inchon landing in Korea. Most of the Marines’ money goes on aviation: their separate identity requires them to use vertical take-off and landing jet fighters, which are unwanted by the US air force or navy and are exceptionally expensive. (The travails of the F-35 joint-strike fighter are caused largely by its vertical take-off version, which distorts the entire design.) As for the air force’s $23 billion, they are spent primarily on the quest for a new manned strategic bomber, whose crew must be recovered safely even after a nuclear strike, and on the F-35, whose unit costs are as phenomenal as its shortcomings.

That’s how $45.5 billion are being spent this year but it takes a huge amount of research and development money to make just a little progress when all the parameters are set from the start. It’s for this reason that car manufacturers spend more than a billion dollars on turning out a new model only marginally superior to its predecessor: laws, regulations, tax regimes, the shape of the human body, the convention that cars have four wheels – all of this constrains innovation, just as the classic military platforms do. It’s only the advent of hybrid, electrical and fuel-cell propulsion that has allowed any innovation at all after decades of stasis. The US air force recently allocated $21.4 billion of its RDT&E funds, nine times this year’s total Darpa budget, to developing a manned bomber. Unsurprisingly, Northrop won the contract to replace the Northrop B-2 flying-wing stealth bomber designed in the 1970s (its only precursor was the 1946 Northrop XB-35 flying-wing bomber) with yet another manned flying-wing stealth design, as if all the intervening innovations starting with unmanned aircraft had never happened, and as if stealth-defeating radar techniques didn’t exist.

The US air force has never funded any innovative research: not ballistic missiles, since Wernher von Braun, who aimed at the moon but hit London, was employed by the US army after the war; not air-to-air tactical missiles, another German idea developed by the US navy; and not remotely piloted aircraft, at least not until the US navy successfully used some imported from Israel, which was also the proximate source of the only truly major innovation in jet-fighter design in decades: the helmet-mounted display, whereby all the critical data – airspeed, heading, altitude, targeting information and warnings – along with real-time imagery from infrared cameras mounted around the aircraft, are projected onto the visor of the helmet, allowing pilots to look through their own airframe with 360 degree vision.

Platformitis – the fixation on retaining existing configurations – is very costly and generates increasing vulnerabilities. Cars are still useful even if their basic layout has not changed in decades, but car drivers don’t have to fight their way past enemies armed with anti-car weapons developed to exploit car-specific vulnerabilities. The continued reliance on the cherished platforms of each service has resulted in the development of more and more lethal platform-killers, often missiles that are much cheaper than the platforms they can destroy. Ask the US navy how vulnerable its aircraft carriers are and you’ll be told not at all, because carrier task-forces have anti-submarine weapons against submarine attacks, anti-air weapons against air attack, and of course anti-ship weapons against ship attacks; but if Darpa were given the task of sinking aircraft carriers, it wouldn’t bother with any of those, but would instead use ballistic missiles to launch warheads that would plunge down at Mach 5 or 6 to cut right through the carriers, from flight deck to keel. That, by the way, is what the Chinese would do too.

Annie Jacobsen’s book sets the stage by reviewing the major pre-Arpa innovations, notably nuclear weapons, computers and ballistic missiles, before telling Arpa’s story from its 1957 foundation. Early on she gets one episode wrong – and it’s an interesting mistake. She dates the Mutual Assured Destruction concept of the 1960s to the 1950s, and then gives credit to the once celebrated and now deservedly forgotten Albert Wohlstetter for the ‘second-strike’ concept: the insight that what deters isn’t a country’s inventory of ballistic missiles but the portion of it that can reliably survive a surprise attack. Actually, Wohlstetter worked on the basing of bombers, not missiles, and the essential insight wasn’t his but his wife Roberta’s: she re-examined the Pearl Harbor attack (her book Pearl Harbor: Warning and Decision has been in print since 1962) and came up with two findings of enduring importance. The first is that surprise attacks succeed not because of impenetrable secrecy, but because the ‘signals’ generated by their preparation are obscured by the ‘noise’ of outdated, irrelevant and misleading information, amplified by wishful thinking. The remedy isn’t the mirage of ‘better analysis’, but deploying would-be deterrent forces in ways that are inherently resilient to surprise attack, even if they’re less cost-effective. The second finding followed directly from this: in Washington it was thought that keeping the US fleet at Pearl Harbor, halfway across the Pacific, would deter the Japanese from attacking US, British and Dutch possessions in South-East Asia, whereas the Japanese viewed the fleet’s relative proximity as an opportunity to destroy US naval strength with one blow. In other words, to deter Japan the fleet should have been kept in San Diego, well beyond Japanese reach. Not coincidentally, Albert Wohlstetter became famous for his 1954 Rand study, which taught the US air force that its nuclear bombers should be moved back to the US to deter the USSR, instead of being forward-based in Europe.

Arpa​ itself was the creation of Eisenhower’s Secretary of Defence Neil McElroy, who had sold soap door to door before rising ever higher at Procter & Gamble: his innovations there ranged from brand management via the competition of rival in-house brands to the now global phenomenon of soap operas. On 20 November 1957, only five weeks into his tenure, and only six weeks after the shock of the Soviet leap into space with Sputnik, he came up with the idea of creating Arpa to carry out ‘advanced research’ into space exploration among other things. He carefully avoided saying that the purpose was to circumvent the deficiencies of the services: the army, navy and air force all had their own space programmes, none sufficiently central to their concerns to warrant much money or attention, which is one reason the USSR won the space race. McElroy’s tact didn’t prevent the top brass – in rare unity – from trying hard to choke Arpa at birth. They failed only because the toppest brass of them all, Eisenhower, happened to be president.

Eisenhower denounced petty ‘jurisdictional’ concerns, successfully asked Congress for funds, and Arpa was in business. McElroy’s choice as its head was Roy Johnson of General Electric, who left his $160,000 job there for $18,000 at Arpa. The chief scientist was Herb York, part-Mohawk, son of a railway baggage man, a Manhattan Project nuclear physicist with a practical bent, utterly unpretentious and fun to be with. Space exploration was soon taken away by the creation of Nasa, so Johnson and York focused on what was known as Project Vela. This had three parts: Vela Hotel concentrated on the development of high-altitude satellites that would be able to detect nuclear explosions from space; Vela Uniform on seismic sensors to detect underground nuclear explosions; and Vela Sierra on detecting nuclear explosions in space. These were truly sinister programmes from the viewpoint of York’s erstwhile mentors in the nuclear weapons business, among them Edward Teller, because they indicated the nefarious purpose of negotiating the prohibition of nuclear tests with Khrushchev – which was exactly Eisenhower’s aim.

By the summer of 1958, Johnson and York were ready to take a much broader look at possible projects. A gathering of 22 defence scientists was charged with identifying promising opportunities. One was the Christofilos effect: if a nuclear warhead was detonated in the upper atmosphere, charged particles would create a radiation belt, which could in theory be used to destroy incoming ballistic warheads by frying their fuses. True, thousands of nuclear explosions would be needed each year to renew the shield, but the idea itself was brilliant, as was the inventor, a Greek lift mechanic called Nicholas Christofilos who taught himself physics under German occupation. After the war, Christofilos started sending letters to US nuclear laboratories describing high-energy accelerators of his own design. The letters were eventually read, and he was hired by York to help build a really big accelerator.

The hypothesis that the Earth’s magnetic field would trap charged particles was tested in the Argus detonations of August and September 1958, which were set up with a rapidity unimaginable today, especially given that the test involved assembling in the far South Atlantic an aircraft carrier, a seaplane tender, a fleet oiler, three destroyers, eight helicopters, 21 fixed-wing aircraft, a dozen missiles, three nuclear warheads and thousands of servicemen, technicians and scientists. Argus was a partial failure (the missiles malfunctioned; the detonations weren’t high enough), but the tests confirmed Christofilos’s hypothesis. Similar tests carried out the same year demonstrated the damaging effects of the electro-magnetic pulse (EMP) caused by a nuclear explosion above the atmosphere on the communications and electrical infrastructure below. These were also shown inadvertently by a Soviet high-altitude nuclear test on 22 October 1962, which set fire to the Karaganda electrical generating plant 180 miles below, and overloaded lead-and-steel-shielded trunk power lines almost a metre below ground. EMP, in other words, has a much larger reach than blast, heat or radiation, which was a matter of great importance to weaponeers, until the post-nuclear age came along to kill off battle-planning with nukes. Christofilos soon got bored with his own ‘effect’ and diverted himself by trying to solve the riddle of communicating with ballistic-missile submarines without forcing them to compromise their security by floating antennae to the surface. He came up with the conceptual answer – extremely low frequency radio – and worked out all the engineering in full detail. It was thanks to him that the US navy was able to communicate with its Polaris submarines and their successors.

Annie Jacobsen’s book doesn’t tell that story in much detail but she surveys many more stories on her way to the present-day Darpa. Natural scientists and engineers were joined by social scientists who tried to help the fight against the Viet Cong by way of psychological operations, success metrics and such. The engineers came up with a lot of things that worked in Vietnam, including almost silent scout aircraft, swamp boats, and the laser-guided bombs that made a large-scale impact in the Gulf War in 1991 – but the social scientists could only fail, because they had no way of contending with ideology. Computer engineers eventually empowered the social scientists, albeit in the guise of intelligence analysts, not to predict the behaviour of the members of al-Qaida, Islamic State, Boko Haram and so on, but to monitor networks of people. Their capability is impressive, but as one who works in this business professionally, I can testify that the loss-of-privacy/terrorist-finding trade-off is roughly a million to one: good enough at a specific time/place, but not everywhere all the time. So while I would be happy for the US services to hand over 90 per cent of their RDT&E money to Darpa, I would also shut down its social science stuff altogether.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences