TechnologyQuarterly

Artifi cial muscleschallenge motors

Brainwave control: sci-fi no longer

Marc Andreessen’s second act

September 3rd 2011

Changes in the airThe emerging technologies that willdefi ne the future of fl ight

TQCOV-September4-2011.indd 1 22/08/2011 15:42

Monitor

2 Computational photography, anew approach to desalination,monitoring yacht performance,spotting fakes with lasers,guiding nanoparticles to �ghtcancer, mopping up oil withwool, smaller military drones,keeping barnacles at bay andthe religious overtones ofcomputing programming

Di�erence engine

9 Worrying about wirelessConcerns about the health risksof mobile phones are misplaced

The future of �ight

10 Changes in the airThe technologies that will de�nethe development of aviation

Modelling behaviour

13 Game theory in practiceHow software can make forecastsand transform negotiations

Inside story

15 Muscling in on motorsArti�cial muscles could givemotors a run for their money

Brainwave controllers

17 Put your thinking cap onControlling things with thoughtis no longer science �ction

Brain scan

19 Disrupting the disruptersA pro�le of Marc Andreessen,programmer turned investor

2 Monitor The Economist Technology Quarterly September 3rd 2011

PHOTOGRAPHY can trace its roots tothe camera obscura, the optical princi-

ples of which were understood as early asthe 5th century BC. Latin for a darkenedchamber, it was just that: a shrouded boxor room with a pinhole at one endthrough which light from the outside wasprojected onto a screen inside, displayingan inverted image. This, you might think,is a world away from modern digitalcameras, brimming with fancy electronicswhich capture the wavelengths and inten-sity of light to produce high-resolutiondigital �les. But the basic idea of focusingrays through an aperture onto a two-dimensional surface remains the same.

Now a novel approach to photographicimaging is making its way into camerasand smartphones. Computational photo-graphy, a subdiscipline of computergraphics, does not simply capture singleimages. The basic premise is to use mul-tiple exposures, or multiple lenses, tocapture information from which photo-graphs may be derived. These data con-tain myriad potential pictures whichsoftware then converts into what lookslike a conventional photo. More computeranimation than pinhole camera, in otherwords, though using real light refractedthrough a lens rather than the virtual sort.

The best known example of computa-tional photography is high-dynamic-range(HDR) imaging, which combines multiplephotos shot in rapid succession, and at

di�erent exposures, into one picture ofsuperior quality. Where a single snap maymiss out on detail in the lightest and dar-kest areas, an HDR image of the samescene looks preternaturally well lit (seeabove). HDR used to be a specialisedtechnique employed mostly by profes-sionals. That changed when Apple addedit as an option in the iPhone 4. (PreviousiPhones lacked the oomph to crunchrelevant data quickly enough to be practi-cal.) Other examples include cameras andapps that stitch together panoramas fromoverlapping images shot in an arc arounda single point or as a moving sequence.

But HDR and panoramas are just twoways to splice together images of the samesubject, says Marc Levoy of StanfordUniversity, who kickstarted computation-al photography in a paper he wrote in 1996with his colleague Pat Hanrahan. Sincethen aspects of the �eld have moved fromacademia into commercial products. This,Dr Levoy explains, is mainly down to theprocessing power of devices, such ascamera-equipped smartphones, growingfaster than the capacity of sensors whichrecord light data. �You are getting morecomputing power per pixel,� he says.

To show o� the potential of some newtechniques, Dr Levoy created the Synth-Cam app for the iPhone and other Appledevices. The app takes a number of suc-cessive video frames and processes theminto a single, static image that improves on

Cameras get cleverer

Consumer electronics: New approaches to photography treat it as a branch ofcomputing as well as optics, making possible a range of new tricks

On the coverFrom lightweight componentsand drag-reducing painttoday, to holographicentertainment systems andhypersonic aircraft tomorrow,researchers are devising theemerging technologies thatwill de�ne the future of�ight. What can tomorrow’stravellers expect? Page 10

Contents

1

The Economist Technology Quarterly September 3rd 2011 Monitor 3

2

1

the original in a variety of ways. He andhis colleagues have also built severalmodels of Frankencamera, using bits of kitfound in commercially available devices�both low-end, like inexpensive cellphones, and high-end, such as priceysingle-lens re�ex cameras. The Franken-cameras use a host of clever algorithms tocapture sequences of images and turnthem into better photos. SynthCam andFrankencameras can improve the qualityof pictures taken in low-light conditions,which are usually quite grainy. They canalso create an arti�cial focus that is absentfrom the original images, or render aforeground �gure in crisp focus against ablurred backdrop.

Point and shoot

Still, for all the superior results that com-putational photography o�ers, Dr Levoylaments, camera-makers have been loth toembrace its potential. This is about tochange. In June this year Ren Ng, a formerstudent of Dr Levoy’s at Stanford,launched a new company called Lytro,promising to start selling an a�ordablesnapshot camera later this year.

Rather than use conventional tech-nology, as the Frankencamera does, tomeld successive exposures, Dr Ng has�gured out a way to capture lots of imagessimultaneously. This approach is knownas light-�eld photography, and Lytro’scamera will be its �rst commercial in-carnation. In physics, a light �eld de-scribes the direction of all the idealisedlight rays passing through an area. DrLevoy’s and Dr Hanrahan’s seminal paperdescribed a new way to model this �eldmathematically. Now, 15 years later, Dr Nghas worked out how to implement thetechnique using o�-the-shelf chips.

Dr Ng’s camera uses an array of severalhundred thousand microlenses insertedbetween an ordinary camera lens anddigital image sensor. Each microlens func-tions as a kind of superpixel. A typicalcamera works by recording where lightstrikes the focal plane�the area onto

which rays passing through a lens arecaptured. In traditional cameras the focalplane was a piece of �lm; modern onesuse arrays of digital sensors. In Lytro’scase, however, a light ray passes throughthe main lens�which uses a wide aper-ture�and then through one of the micro-lenses. Only then does it hit a sensor. Bycalculating the path between the micro-lens and the sensor, the precise directionof a light ray can be reconstructed. That inturn means it is possible to determinewhere the ray would have struck if thefocal plane had been moved. Moving thefocal plane is equivalent to refocusing thelens, so any point in the light �eld canthen, in e�ect, be brought into sharp focus.In other words, images can be refocusedafter they have been taken (see below).

It is also possible to �ddle with animage’s depth of �eld�photographicjargon for the space between the closestand most distant points in an image thatare in focus�or even create an image inwhich every point is in focus. And thelight-�eld approach can produce a com-pelling simulation of a stereoscopic im-age. Lytro’s website lets visitors �ddle withexisting images to see how some of thesefeatures will work.

Unlike the information it records, thecamera itself is simple. The main lens is�xed in place; there is no auto-focus, auto-aperture, or other machinery which needsto be activated every time a photo is taken.Such adjustments take time, causing a lagbetween pressing the shutter-releasebutton and actually capturing the image.Lytro’s snaps, by contrast, will be trulyinstantaneous, just as they were in oldsnapshot cameras. And, since the lens ispreset always to capture the greatestamount of light possible, exposure timecan be short, even in poorly lit conditions.

One potential downside is the cam-era’s low resolution, which is de�ned bythe number of microlenses, because theprocessing software treats each microlensas a single pixel. The sample images onLytro’s website measure 525 by 525 pixels,

which works perfectly well online, butwill not pass muster in print. This mightnot matter, though. Nowadays peoplemake fewer photographic prints, especial-ly large-format ones where resolutioncounts. By contrast, billions of photo-graphs are shared online each year. Profes-sional photographers may still seek higherpixel counts, but there is no reason whyfuture versions of the device could noto�er more microlenses.

For now, though, Lytro is targetinginternet photo-sharers. It will let ownersof its camera upload the image data andthe processing tools to Facebook and othersocial networks. The �rm has reportedlyalready raised $50m. Investors must behoping that consumers �nd all the irri-tants that Lytro’s camera removes, likeblurred or dim pictures, niggling enoughto want them eliminated once and for allfrom their holiday snaps. 7

Shoot �rst, refocus later

SINGAPORE’S average annual rainfall ismore than double that of notoriously

soggy Britain, so the casual observer mightbe surprised to learn that the place has ashortage of drinking water. Yet witharound 7,000 people per square kilo-metre, Singapore is the third most denselypopulated country in the world. Its landmass is not large enough to supply its 5minhabitants with water.

One answer is to desalinate seawater.That, though, is expensive, so the Singa-porean government is keen to �nd cheap-er ways of doing it. And, in collaborationwith Siemens, a German engineeringconglomerate, it may have done so, forSiemens says its demonstration electro-chemical desalination plant on the islandcan turn seawater into drinking waterusing less than half the energy required bythe most e�cient previous method.

To make seawater �t for human con-sumption its salt content of approximate-ly 3.5% must be cut to 0.5% or less. Existingdesalination plants do this in one of twoways. Some employ distillation, whichneeds about 10 kilowatt-hours (kWh) ofenergy per cubic metre of seawater pro-cessed. Brine is heated, and the resultingwater vapour is condensed. Other plantsemploy reverse osmosis. This uses molec-ular sieves that pass water molecules

Drops to drink

Desalination: A technique calledelectrodialysis may provide acheaper way to freshen seawater forhuman consumption

4 Monitor The Economist Technology Quarterly September 3rd 2011

2

1

while holding back the ions, such as sodi-um and chloride, that make water salty.Generating the pressure needed to do thissieving consumes about 4kWh per cubicmetre. The Siemens system, by contrast,consumes 1.8kWh per cubic metre, andthe �rm hopes to get that down to 1.5kWh.

It works using a process called electro-dialysis, in which the seawater is pumpedinto a series of channels walled by mem-branes that have slightly di�erent proper-ties from those used in reverse osmosis.Instead of passing water molecules, thesemembranes pass ions. Moreover, themembranes employed in electrodialysisare of two types. One passes positivelycharged ions and the other passes nega-tively charged ones. The two types al-ternate, so that each channel has one wallof each type. Two electrodes �anking thesystem of channels then create a voltagethat pulls positively charged ions such assodium in one direction and negativelycharged ions such as chloride in the other.

The result is that the ions concentrate inhalf of the channels, creating a strongbrine, while fresher water accumulates inthe other half. As the brine emerges, it isthrown away. The fresher water is putthrough the same process twice more andeventually has its salt concentration re-duced to 1%. That is not bad, but is stilldouble what is potable. There is thereforeone further step. This is to employ anion-exchange resin in addition to themembranes. Such resins increase theelectrical conductivity of the system andallow one more passage, bringing the saltconcentration below 0.5%.

A demonstration plant has been oper-ating since December, and a full-scale pilotplant should be completed by 2013. If allgoes well, then, Singapore’s inhabitantswill soon no longer feel like Coleridge’sancient mariner�that there is water, wa-ter, everywhere, nor any drop to drink. 7

DURING the European Grand Prix inValencia on June 26th, Lewis Hamil-

ton discovered that his tyres were over-heating. It was not, however, a whi� ofburning rubber that gave it away. Instead,the news came from the pits, where agroup of engineers spend the entire raceglued to a bank of monitors replete withnumbers and graphs streaming in fromtheir teams’ cars. Every second, sensors onthe vehicles take hundreds of di�erentmeasurements�the engine, suspension,or the drivers’ well-being�and relay themto the pits. Besides highlighting problems,this stream of data lets team strategistsadvise Mr Hamilton and his rivals ontactics and on how to optimise vehiclesettings with the help of the dozen ormore switches on the steering wheel.

Such data logging and telemetry havemade what used to be more of an art intoan exact science. As a result, drivers areable to shave fractions of a second o�their lap times. These, aggregated over atypical Formula 1 race’s 50 or so laps, canmake the di�erence between winning andlosing. Now the technology, pioneered inmotor racing, is being applied in anotherdiscipline where split seconds provide anedge: sailing.

Leading the way is Cosworth, a Britishcompany best known for making racing

engines, but which also provides many ofthe Formula 1 teams with their data-acqui-sition and analysis equipment. Theirmarine systems work in much the sameway as they do on racing cars. But insteadof measuring cornering forces and sus-pension movement, they look at windspeed, yaw, rudder angles and sundryother factors that e�ect the performanceof a racing yacht or dinghy. Some sailingteams training for the Olympics haveadopted the technology, as have severalcompetitors in the America’s Cup, sail-ing’s most prestigious event.

As with motor racing, a reliable gaugeis needed to ensure that successfulmanoeuvres are repeated consistently. Theheart of Cosworth’s Pi Garda is a blackbox that logs data from sensors inside it,such as accelerometers to measure g-forces and a satellite positioning system todetermine position and speed. It also takesinformation from sensors which measurehow the boat is behaving on the water, aswell as standard marine instruments, suchas a wind wand on the mast that measureswind speed and direction.

Some sensors are copied straight fromthe racetrack. A small laser sensor, forinstance, is mounted underneath a racingcar’s chassis to bounce a beam o� thetrack surface in order to calculate thedown-force being exerted on the vehicle. Itis accurate to within 0.03mm. The samesensor is used on a boat to shine a beamo� the tiller bar, using the re�ection tomeasure the angle of the rudder. Othersensors are more bespoke. For instance,strain gauges calculate the stretch in thewires or ropes in the rigging and sails.

The information can be displayed tothe crew on the yacht or transmitted tocoaches on a chase boat, or to a supportteam on shore. Crunching the numbersallows accurate predictions of when the

Formula 1 goessailing

Performance analysis: Technologyused to assess and improve theperformance of racing cars is nowtaking to the water

It says the tyres are overheating, captain!

The Economist Technology Quarterly September 3rd 2011 Monitor 5

2

1

yacht will get to the next buoy, how manytacks it will require to get there and howbest to tackle the necessary turns, saysSimon Holloway of Cosworth’s marinedivision. A similar system might ease thepressure on the solitary round-the-worldyachtsmen, he adds.

In some competitions, including theOlympics, such equipment is not allowedduring the event itself. Nevertheless, itremains a valuable training aid. Smaller,self-contained systems have also beendeveloped for enthusiastic amateurs, evenwindsurfers. This echoes what has hap-pened to motor-racing telemetry. Manymodern cars now employ devices �rstused in Formula 1 to let the driver knowhow the vehicle is faring. It remains to beseen whether the Oxford punt will get asimilar makeover. 7

FROM banknotes to bottles of Bordeauxand Vans shoes to Viagra, good forger-

ies can be hard to detect�even for experts.Yet if you look closely enough with amicroscope, the surface of almost anymaterial shows a naturally occurringrandomness. The wood �bres in a piece ofpaper look like a layer of noodles; smoothplastic resembles a mountain range. Thedetails of these patterns are unique toeach item and thus could be used like a�ngerprint, to provide an almost foolproofmeans of identi�cation.

The trouble is that employing a micro-scope powerful enough to record surfacefeatures at the required level of detail (afew microns) would be an expensive andcumbersome business, and not at allpractical on a production line. But if youshine a laser at the surface of an object, thecharacteristic way in which the light isre�ected back can be used to gather infor-mation about the same features. And afast, low-cost way of doing just that hasnow been commercialised by IngeniaTechnology, a company based in London,to provide what it calls a tamper-proofmethod of �laser surface authentication�.

The process was developed initially atImperial College, London, and is based ona phenomenon known as laser speckle�the scattering of light caused by micron-sized ridges and grooves on an object’ssurface. By analysing this speckle, it is

possible to chart the texture of the surface.Ingenia’s machines use a scanning

head consisting of three small lasers andsix detectors to examine part of an object.The strip that is scanned is predetermined;the top left-hand corner of a credit card,for instance. Variations in the speckle arethen digitised to produce a code that isunique to the scanned item. This code islogged in a database, along with the pro-duct’s serial number or bar code. It canalso be encrypted into the bar code. Whenwhat purports to be the same item isre-scanned at some later date, it shouldshow the same pattern of speckle.

According to Andrew Gilbert, one ofIngenia’s directors, the probability of twosurfaces generating the same code is lessthan one in a million trillion trillion. Thatis far more accurate than �ngerprints, forexample. Nor is the system easy to fool. Apiece of paper such as a banknote can becrumpled, soaked in water, scorched andscribbled on but will still be readable.Even torn, scratched and partially missingsurfaces can be read. This is because,during the original scan, the detectors pickup such a large amount of informationthat a re-scan need provide only part ofthe speckle pattern for a reliable compari-son to be made. Too much damage to ortinkering with an item would, of course,raise suspicions anyway.

With scan times of less than a second,the system is fast enough to be used on aproduction line. Nor does it involve hav-ing to make changes to a product or itspackaging to incorporate security features,such as adding watermarks, �tting holo-grams or implanting microchips. Ingeniahas tested the system on the packagingused for various luxury goods, along withthe security seals used on dangerous orvaluable substances, and on passports,postage stamps and documents such as�nancial instruments. Indeed, should abank want to, it could match every note itissued against its printed serial number.However clever a counterfeiter was, forg-ery would then be all but impossible. 7

Zapping fakeswith lasers

Security technology: Microscopicsurface features can provide proof ofidentity, allowing counterfeit goodsto be identi�ed more easily

Is it real or fake?

FOR decades doctors have attackedcancer with drugs that kill malignant

cells. Unfortunately, such chemotherapykills a lot of healthy cells as well. In recentyears, the use of drug-carrying particles afew nanometres across has improvedmatters. Such particles can be tailored torelease their payloads only when thesurrounding environment indicates thatthey are near a tumour, thus reducingcollateral damage. Even that, however, hasnot proved perfect. Typically, only about1% of the drugs packaged up in nanoparti-cles this way make it to their destination.

Sangeeta Bhatia and Geo�rey vonMaltzahn of the Massachusetts Institute ofTechnology, however, hope to change that.As they report in Nature Materials, theybelieve that by granting nanoparticles theability to communicate with one another,the success of drug delivery can be in-creased 40-fold. They were inspired byone of the body’s natural communica-tions systems: the way that injured tissuecalls for help to stem bleeding. They won-dered if they might be able to piggybackon this system to deliver drugs to tu-mours�and they found that they could.

When the body sustains an injury,molecules called noti�cation proteins areproduced at the site. These proteins com-

Particle physic

Medicine: Taking advantage of anatural signalling system enablesnanoparticles to deliver cancer drugsfar more accurately

6 Monitor The Economist Technology Quarterly September 3rd 2011

2

1

municate with clotting agents in theblood. They round up cellular fragmentsknown as platelets, along with moleculesof a soluble protein called �brinogen,both of which circulate in the blood-stream. The �brinogen turns into �brin, aninsoluble, �lamentous protein whichtraps the platelets and causes them to linkup into a quilt that helps stop bleeding.

The two researchers wondered if theycould subvert this system to gather drug-carrying nanoparticles into the right place.To do so, they realised that they wouldneed two types of nanoparticles. �Signal-

ling nanoparticles� would function likenoti�cation proteins, marking the spotwhere action was required. �Receivingnanoparticles� would then be recruited asplatelets, but instead of stanching awound they would deliver the drugs.

For the signalling nanoparticles theteam used tiny golden rods. These tend tocollect at the locations of tumours be-cause the blood vessels which serve tu-mours often have unusual pores in them.These pores are between 100 and 200nanometres in diameter�perfect for trap-ping the rods and thus marking the tu-

mour. Once the rods were in place, theteam �red a burst of laser light in thegeneral location of the tumour. This lightwas tuned to be absorbed by gold and wasthus converted into heat only in placeswhere the rods had accumulated. Thatdamaged the surrounding tissue enoughto activate the coagulation system.

The clever bit was that the receivingnanoparticles, which carried the pharma-ceutical payload, were doped with proteinfragments that bind to �brin�and thus tothe wound-stanching quilts that formwhen the heat from the gold nanorodsdoes its work. Only then do the nanoparti-cles release their cargo. The result, DrBhatia and Dr von Maltzahn report, is adelivery system that is 40 times moree�ective than using nanoparticles bythemselves. Moreover, in mice at least, thisapproach shrinks cancers more e�ectivelythan other nanoparticle-based treatments.Work to evaluate the treatment in humanpatients should follow soon. 7

BIELLA, in north-west Italy, is thecentre of a cluster of wool manufac-

turers and the home of ErmenegildoZegna, a luxury clothing brand. A groupof the town’s businessmen have, how-ever, come up with a scheme far fromthe catwalks and seasonal collections.They plan to use wool, which is good atrepelling water and absorbing oil, tosoak up oil spills. They had the idea afterthe Deepwater Horizon disaster and itwould, they reckon, have worked betterthan the containment booms, chemicaldispersants and other methods de-ployed last year in the Gulf of Mexico.

Earlier this year, TecnomeccanicaBiellese, an engineering �rm that makesmachinery for the woollens industry,carried out experiments using greasywool to see how good the �eece was atgathering oil. It turned out to be verygood. Coarse wool (the cheapest sort,with a �bre diameter of between 25 and40 microns) was able to absorb tentimes its own weight of heavy fuel oil, are�nery product similar to crude. More-over, the oil could be squeezed out andthe wool reused. Indeed, even after adozen immersions in oil, for between 15and 20 seconds each time, the wool’sabsorptive capacity did not decline.

Moving out of the laboratory andonto the water, with a working oil-collection system, is the next step. InMarch the businessmen, who havecalled their project Wool Recycle EcoSystem, obtained patents for a contain-erised kit that can be set up in boats todeal with small spills, and for a biggership-based system to tackle large ones.

Mario Ploner, the managing directorof Tecnomeccanica Biellese, says theship-based system will use external

booms running parallel with the ves-sel’s sides to channel oil onto wool thathas been spread over the surface of thesea. As the ship moves through a spill,the oil-impregnated wool will be gath-ered mechanically up ramps and takeninto the ship. As the wool is transportedup these ramps any droplets of waterattached to it will be shaken o�. Onceon board the wool will be pressed torecover the oil and then reused.

Mr Ploner estimates it would costabout �1m ($1.4m) to equip a 50-metrevessel to carry ten tonnes of wool. Thatwould be sucient, in optimum circum-stances, to recover more than 1,000tonnes of oil. In practice, he reckons,cleaning up the Deepwater Horizon spillof almost 5m barrels would have need-ed around 7,000 tonnes of wool. At acurrent market price of less than $1akilo, that does not add up to a huge sumfor an industry as big as Big Oil. Itwould, on the other hand, be a nice littleearner for sheep shearers.

A golden �eece?

Environmental technology: It sounds low-tech, but wool could provide ane�ective way of mopping up oil spills at sea

THE future of air power is likely to beunmanned. It may also be surprisingly

small. Reapers and Predators grab theheadlines, but such big, well-knowndrones are already outnumbered by small,cheap and capable aircraft.

One good example is the RQ-11B Raven,made by AeroVironment of Monrovia,California, and widely used by America’sarmed forces. It looks like a model aircraft.When disassembled it �ts into a backpack.Launching it is a matter of snapping theparts together and throwing it into the air,whence it is carried aloft by an electricpropeller. It weighs two kilograms. TheAmerican army’s entire annual purchaseof almost 1,300 Ravens is thus lighter thana single fully armed Reaper. Pilots mightdismiss Ravens as radio-controlled toys,but they are popular with soldiers. Moreare being rushed to Afghanistan.

At its simplest, a Raven acts as a �yingpair of binoculars that can look over thenext hill, or escort a convoy from above.Being small and quiet, it can get close totargets unobserved, for a good look. Un-like bigger drones, whose limited num-bers mean that ocers in the �eld are inconstant competition for their services,Ravens are abundant and thus generally

Joining thedrones club

Military technology: When it comesto unmanned aircraft, lots of smalldrones are cheaper, and in some waysbetter, than a few big ones

The Economist Technology Quarterly September 3rd 2011 Monitor 7

2 available to provide instant video imagery,day or night. Special display softwareoverlays the images they send back on amap to produce a moving picture of whatis going on on the ground. An operator canthus call down artillery �re with lethalprecision without having to see the targetdirectly. For extra accuracy, Raven can alsomark targets with a laser illuminator.

Another reason for Raven’s popularityis that it is easy to use. The controllinghardware is a tablet computer with but-tons on the side, rather like a portablegames console, and most people can getthe hang of it in a couple of days. Preda-tors, by contrast, were originally �own byreal, albeit ground-based, pilots�and,though high demand for operators has ledto a new rapid-training course for ground-lings, it still takes 22 weeks to complete.Ravens are now being upgraded to use acommunications system that providesenough bandwidth for 40 of them to �y inthe same area, instead of the current four.

The American army has experimentedwith turning the drones into miniaturebombers, capable of delivering a grenade-sized weapon, to destroy a small vehicleor take out the occupants of a particularroom with high precision and little collat-eral damage. For greater punch, Aero-Vironment has a prototype version of alethal drone called Switchblade. Thisresembles Raven, but is a �ying bomb,packed with explosives. Its guidancesoftware enables it to lock on to and fol-low a rapidly moving target.

At the moment, Ravens cost around$56,000 each, and economies of scaleshould bring this down. By contrast, ma-chines like the Predator cost at least $5m,and another $5,000 an hour to �y. Fromthe army’s point of view, small is de�-nitely beautiful. 7

It’s certainly not a toy

IN THE decades-long battle for navalsupremacy that was fought between

Britain and France at the end of the 18thcentury, the British �eet had a secret weap-on. It was, as secret weapons often are,hugely expensive. But it paid o�, givingBritish ships more speed, manoeuvrabil-ity and staying power than their Frenchrivals. It was copper.

By covering the underwater parts oftheir ships’ hulls with copper plates�which slowly dissolved, releasing toxiccopper ions as they did so�the Britishadmirals stopped barnacles, mussels andburrowing clams from taking up resi-dence. In �eets that were otherwise well-matched the result was decisive. Francelost. The British empire became the globalsuperpower of the 19th century. And theworld speaks English, not French.

Ship-fouling, then, can have ratherfar-reaching consequences. Even now,when naval supremacy is less of an issue,the problem is rife. The drag imposed by aheavy infestation of barnacles may push aship’s fuel consumption up by as much as40%. The solution usually adopted issimilar to the Royal Navy’s: poison. Cop-per is still used, though in the form ofcopper-laced paint. Another popularchemical is tri-butyl tin. But releasing toxicheavy metals into the sea is frowned uponthese days�indeed, tri-butyl tin is nowillegal in many parts of the world�so thesearch is on for alternatives.

One possibility is to use one of a groupof chemicals called avermectins. Theseare antiparasite agents (the most familiarof which is called ivermectin) that arewidely used against �eas and gut worms.They also, according to Hans Elwing of theUniversity of Gothenburg, prevent barna-cle colonies from taking hold by stuntingtheir growth. A barnacle that runs into thechemical �nds it cannot bind as stronglyto the surface. Only a tiny amount, aboutone part in a thousand of the paint byweight, is needed, and other marine spe-cies are not, as far as can be ascertained,a�ected by it.

Another way to discourage barnacles isto confuse them. A formula developed byGiancarlo Galli of the University of Pisauses polymer molecules that are water-attracting on one side and water-repellenton the other. When they are painted onto

a surface, this arrangement forces theminto a kind of checkerboard pattern whichmakes it much harder for barnacles andmussels to stick, according to David Wil-liams, who is in charge of commercialisingthe idea at AkzoNobel, a multinationalchemical company.

If checkerboarding does not work out,AkzoNobel has an alternative: create asurface so smooth that barnacles cannothold on to it. This is done with a �uoro-polymer�a chemical similar in structureto Te�on. The paint does not stop theanimals attaching themselves to a hull inthe �rst place, but once the vessel is mov-ing faster than ten knots, the water sweepsthem away. That is no problem for com-mercial vessels, which are always on thego. But for pleasure boats, which mayspend a lot of time idle, Dr Williams’steam is trying to improve the formula sothat a boat need not be moving so fastbefore the paint does its job.

Small boats, particularly on inlandwaterways, are also the focus of work byJohn Schetz of the University of NorthTexas and Robert McMahon of the Univer-sity of Texas. They have been experi-menting with a mixture containing amolecule similar to capsaicin (the activeingredient of hot peppers) and anothersimilar to THC (the active ingredient ofcannabis). Fouling is less of a problem forboats in fresh water, as barnacles arepurely marine. But recently the inlandwaterways, docks and fresh-water intakesof North America have been overrun byzebra and quagga mussels�species thatoriginate from the area around the BlackSea. The mixture Dr Schetz and Dr McMa-hon have come up with seems particular-ly e�ective against these animals, thoughthey have yet to commercialise it.

Their aim is not just to help boat own-ers, but also to stop them unwittinglyspreading the mussels still further. Accord-ing to Dr McMahon, a big part of the pro-blem is that both species can survive outof the water for several days, so transport-ing a boat overland from one river basin toanother, a common practice in NorthAmerica, will not necessarily kill them.Also, boat owners are not always as dili-gent as they might be when it comes toinspecting their vessels for signs of in-festation. And even if they do look, themussels can be hard to see, especiallywhen they are young and therefore small.

There is thus a lot to play for. Reducingfouling will save fuel, in turn saving mon-ey, as well as cutting shipping’s contribu-tion to greenhouse-gas emissions. Andstopping the spread of invasive musselswill make life easier for those who navi-gate the waterways of America and Cana-da. The prize may not, this time round, beworld domination. But whoever comesup with the winning formula is likely tomake a fortune. 7

Reducing thebarnacle bill

Anti-fouling technology: Ships’ hullsare kept clean using poisonouschemicals, but a number of cleaneralternatives are being pursued

8 Monitor The Economist Technology Quarterly September 3rd 2011

�THE kingdom of heaven belongs tosuch as these,� Jesus said of little

children. But computer hackers might givethe kids some competition, according toAntonio Spadaro, an Italian Jesuit priest.In an article published earlier this year inLa Civiltà Cattolica, a fortnightly magazinebacked by the Vatican, entitled �Hackerethics and Christian vision�, he did notmerely praise hackers, but held up theirapproach to life as in some ways divine.Mr Spadaro argued that hacking is a formof participation in God’s work of creation.(He uses the word hacking in its tradition-al, noble sense within computing circles,to refer to building or tinkering with code,rather than breaking into websites. Suchnefarious activities are instead known as�malicious hacking� or �cracking�.)

Mr Spadaro says he became interestedin the subject when he noticed that hack-ers and students of hacker culture used�the language of theological value� whenwriting about creativity and coding, so hedecided to examine the idea more deeply.The hacker ethic forged on America’s westcoast in the 1970s and 1980s was playful,open to sharing, and ready to challengemodels of proprietary control, competi-tion and even private property. Hackerswere the origin of the �open source�movement which creates and distributessoftware that is free in two senses: it costsnothing and its underlying code can bemodi�ed by anyone to �t their needs. �In aworld devoted to the logic of pro�t,�wrote Mr Spadaro, hackers and Christianshave �much to give each other� as theypromote a more positive vision of work,sharing and creativity.

He is not the only person to see ana�nity between the open-source hackerethos and Christianity. Catholic open-source advocates have founded a groupcalled Elèutheros to encourage the churchto endorse such software. Its manifestorefers to �strong ideal a�nities betweenChristianity, the philosophy of free soft-ware, and the adoption of open formatsand protocols�. Marco Fioretti, co-founderof the group, says open-source softwareteaches the �practical dimension of com-munity and service to others that is al-ready in the church message�. There arealso legal motivations. Commercial soft-ware such as Microsoft Word is widelypirated in many parts of the world, byCatholics as well as others. Mr Fioretti

advocates the use of open-source softwareinstead, because he doesn’t want people�to violate a law without any real reason,just to open a church document�.

Although the Vatican has yet to encour-age the faithful to live like hackers, it haspraised the internet as �truly blessed� forits ability to connect people and shareinformation. The pope has even joinedTwitter. But praise has always been tem-pered by warnings. As early as 2002, forinstance, the Vatican’s �Church and In-ternet� document cautioned that �thereare no sacraments on the internet� andworried about the solipsistic appeal oftechnology. Moreover, hackers in partic-ular have problematic traits from theperspective of the Catholic church, suchas a distrust of authorities and scepticismtoward received wisdom. And the idea oftweaking source materials to �t one’sneeds doesn’t mesh well with the Catho-lic emphasis on authority and tradition.

Cathedrals and bazaars

Mr Spadaro recognises these tensions but�nds them manageable. Not everyoneagrees. Eric Raymond, author of a classicessay on open-source software, �TheCathedral and the Bazaar�, �nds it hard tobelieve that some Christians want tocanonise the hacker mindset. After beingquoted in Mr Spadaro’s paper, Mr Ray-mond took to his own website to note thathe had deliberately equated cathedralswith proprietary, closed-source softwaredirected from above, by contrast with themore chaotic bazaar of equals whichproduces open-source code. �Cathedrals�

vertical, centralising religious edi�cesimbued with a tradition of authoritar-ianism and ‘revealed truth’�are the polaropposite of the healthy, sceptical, anti-authoritarian nous at the heart of thehacker culture,� Mr Raymond declared. Asfor Mr Spadaro’s ideas, they possessed a�special, almost unique looniness�.

But Mr Spadaro is merely the latest tolink coding with Christian attitudes to-wards creativity and sharing. Don Parris, aNorth Carolina pastor, wrote an article inLinux Journal in 2004 in which he arguedthat �proprietary software limits my abili-ty to help my neighbour, one of the cor-nerstones of the Christian faith.� LarryWall, the creator of Perl, an open-sourceprogramming language, said in an in-terview a decade ago that God expectshumans to create�and to help others doso. Mr Wall said he saw his popular lan-guage as just such a prod to creation, say-ing, �In my little way, I’m sneakily helpingpeople understand a bit more about thesort of people God likes.�

More recently Kevin Kelly, co-founderof Wired magazine and author of �WhatTechnology Wants�, published last year,has argued that creation can go further incode. Whereas a novelist can craft a newworld, coders can build worlds completewith arti�cial agents that exist and evolveoutside the creator’s mind. Mr Kelly takesliterally the words of his friend StewartBrand, whose �Whole Earth Catalog�quipped, �We are as gods and might aswell get good at it.� Mr Kelly, a Christian,says the ability to create arti�cial life willcome with great parental responsibilityand suggests that arti�cial worlds willneed to be imbued with moral value.�This causes a kind of revival of religion,�he says, �because religion has been think-ing about this issue.�

From the outside, hacking computercode has largely been viewed as a tech-nical discipline, not as a theologically richvision of how to live. But some see adivine aspect to programming�at leastwhen looking with the eye of faith. 7

What would Jesus hack?

Cybertheology: Just how much does Christian doctrine have in common withthe open-source software movement?

The Economist Technology Quarterly September 3rd 2011 Di�erence engine 9

ALTHOUGH the myth that mobilephones cause cancer has been laid to

rest, an implacable minority remainsconvinced of the connection. Their fearshave been aggravated of late by bureau-cratic bickering at the World Health Orga-nisation (WHO). Let it be said, once andfor all, that no matter how powerful aradio transmitter�whether an over-the-horizon radar station or a microwavetower�radio waves simply cannot pro-duce ionising radiation. The only possiblee�ect they can have on human tissue is toraise its temperature slightly.

In the real world, the only sources ofionising radiation are gamma rays, X-raysand extreme ultra-violet waves, at the far(ie, high-frequency) end of the electromag-netic spectrum�along with �ssion frag-ments and other particles from within anatom, plus cosmic rays from outer space.These are the sole sources energeticenough to knock electrons out of atoms�breaking chemical bonds and producingdangerous free radicals in the process. It ishighly reactive free radicals that can dam-age a person’s DNA and cause mutation,radiation sickness, cancer and even death.

By contrast, at their much lower fre-quencies, radio waves do not pack any-where near enough energy to produce freeradicals. The �quanta� of energy (ie, pho-tons) carried by radio waves in, say, theUHF band used by television, Wi-Fi, Blue-tooth, cordless phones, mobile phones,microwave ovens, garage remotes andmany other household devices haveenergy levels of a few millionths of anelectron-volt. That is less than a millionthof the energy needed to cause ionisation.

All of which leaves doctors more thana little puzzled as to why the WHO shouldrecently have reversed itself on the ques-tion of mobile phones. In May the organi-sation’s International Agency for Researchon Cancer (IARC) voted to classify radio-frequency electromagnetic �elds (ie, radiowaves) as �a possible carcinogenic tohumans� based on a perceived risk ofglioma, a malignant type of brain cancer.

A year earlier, after a landmark, de-cade-long study undertaken by teams in 13countries, the IARC had reported that noadverse health e�ects associated with theuse of mobile phones could be found. Asfor the heating e�ects of radio waves, the

increase in temperature of the skin causedby holding a phone close to the ear wasfound to be an order of magnitude lessthan that caused by direct sunlight.

The Group 2B classi�cation the IARC

has now adopted for mobile phonesdesignates them as �possible�, rather than�probable� (Group 2A) or �proven�(Group 1) carcinogens. This rates thehealth hazard posed by mobile phones assimilar to the chance of getting cancerfrom co�ee, petrol fumes and false teeth.

That has not stopped the tinfoil-hat

brigade from continuing to believe thatdeadly waves in the ether are frying theirbrains. Lately the paranoia has spread tothe smart meters being introduced byelectrical utilities in various parts of theworld. Smart meters are designed to relaywireless messages to the power companyabout a household’s pattern of electricityuse. Such real-time data could help util-ities manage their generating capacitymore intelligently.

But a backlash among homeowners innorthern California, who fear they areabout to be drenched in dangerous radiowaves, has forced a handful of municipal-ities to slap moratoriums on the smartmeters being introduced by Paci�c Gas &Electric. Customers will be given theoption to keep their old analogue meters,but will be charged for having someonecome to read them every month.

Actually, smart meters are just aboutthe last thing that people need worryabout. In an independent study releasedin April, the California Council on Scienceand Technology, an advisory arm of thestate legislature, concluded that wirelesssmart meters produce much lower levelsof radio-frequency exposure than manyexisting household devices�especiallymicrowave ovens. The council noted that,to date, it had not been possible to identifyany health problems resulting from poten-tial non-thermal e�ects of radio waves(should such e�ects exist). But nor had itbeen possible to show categorically thatthere weren’t any.

You can’t prove a negative

The latter is next to impossible. Indeed, byclassifying mobile phones as a Group 2B

risk, what the IARC was e�ectively saying(and the California Council on Scienceand Technology implying) was that, evenif such a health risk exists, there is no wayof ever ruling out bias, chance or otherconfounding circumstance with anyreasonable degree of con�dence. So, tohedge bets, protect careers and guaranteefuture funding, the obvious thing to sug-gest is yet more research on the long-term,heavy use of mobile phones. The mostlikely result is that the results will beequally inconclusive.

And equally irrelevant. The Twittergeneration tweets and texts rather thantalking. Older people are catching up fast.According to Nielsen, a market-research�rm, the number of text messages sentand received by Americans aged between45 and 54 rose by 75% in the year to thesecond quarter of 2010. Over the sameperiod, the number of phone calls madeand received by adults of all ages fell by25%. Meanwhile, for those who still insiston yakking, hands-free is fast becomingthe norm, thanks to sti�er penalties forusing handsets while driving and thespread of Bluetooth headsets.

The whole brouhaha over mobilephones causing brain cancer is a monu-mental irrelevance compared with sco�-laws who insist on using their handsets totext or talk while driving. Regretfully, thatis a far more likely cause of death or dis�g-urement than some inexplicable form ofradio-induced glioma. 7

Di�erence engine l Worrying about wireless

Technology and society: Concerns about the danger posed to human health by radio waves are misplaced�andincreasingly irrelevant. The use of phones while driving is far more likely to cause harm

ON THE evening of July 23rd 1983, AirCanada Flight 143 ran out of fuel after

a series of human errors. The new Boeing767, lightly loaded with 61 passengers andeight crew, became a glider with only 8,686metres (28,500 feet) of altitude to reach thenearest airport at Winnipeg, around120km (75 miles) away. Ten minutes later itbecame clear that the plane was losing alti-tude too fast to make it.

The pilots changed course, hoping toreach a former air-force base near the townof Gimli. They were unaware that its run-ways had been converted into a drag-rac-ing track. Spectators scattered when theysaw the silent approach of the aircraft. Asthe aircraft’s wheels hit the racetrack, thefront landing gear collapsed and the noseslammed into the tarmac, sending sparks�ying. Scraping against a guard rail that di-vided the track in two, the aircraft skiddedto a stop. No one was killed.

Had the pilots been �ying one of to-day’s more aerodynamic airliners, theycould easily have reached Winnipeg’s air-port, says Carl Holden, a recreational-glid-er instructor and head of Holden Dynam-ics, a consultancy based in Sydney thatadvises Australia’s Civil Aviation Safety

Authority. Today’s airliners would glideabout 25% farther, he says, and the nextgeneration promises additional gains.

Gliding distance is an imperfect mea-sure of an airliner’s aerodynamic e�cien-cy, since it is not designed for gliding. Butthe Gimli Glider incident, as it becameknown, helps illustrate the magnitude ofadvances in aviation technology. Im-proved e�ciency means that Boeing’s new787 airliner consumes about 40% less fuelper passenger than its 1970s aircraft. Airbusand other manufacturers have achievedsimilar results.

Not all improvements in aircraft tech-nology are incremental. As myriad tech-nologies mature, new sorts of aircraft be-come possible. Unmanned aircraft have�own at more than �ve times the speed ofsound. Last year a lightweight, pilotedSwiss aircraft, Solar Impulse, capturedenough solar energy during the day to �ythroughout the night. Small drones are be-ing developed with highly e�cient wing-bottom infra-red cells that scavenge radia-tion energy re�ected up from the ground.Boeing is developing unmanned spy air-craft capable of staying aloft using hydro-gen power for �ve years without refuell-ing. Drew Mallow, the project’s leader, callsPhantom Eye, a prototype with a 46-metrewingspan, a �poor man’s satellite�. The fu-ture of �ight will involve gradual changesin the near term, with the prospect of moreradical shifts in the decades to come.

Much research has been driven by ef-

forts to save jet fuel. Having more thandoubled in price in recent years, it now ac-counts for about half of airlines’ operatingcosts. Even slight gains in e�ciency quick-ly pay o��as a rule of thumb, a 1% im-provement knocks more than $1m o� a air-liner’s fuel bill over its lifetime of roughly20 years, says Ihssane Mounir, Boeing’svice-president of sales for China. Thesesavings snowball. Fuel-sipping planes aremore pro�table, so banks will �nancethem at lower interest rates.

Time to lose weight

In the push to improve e�ciency, wing�aps are now operated with lightweightelectrical systems instead of hydraulics. Atleast one airline, Australia’s Jetstar Air-ways, is replacing in-�ight entertainmentkit with Apple iPads, which are muchlighter. Flight Sciences International, a con-sultancy based in Santa Barbara, Califor-nia, has found that fuselage-insulationblanketing costs airlines unnecessarily: itabsorbs humidity and becomes heavierover the years. This is typical of the zealwith which savings are being sought. Theonly area where technologists have failedto improve e�ciency is in reducing theweight of passengers, says John Corl ofFlight Sciences. He is only half joking.

Aircraft engineers have for years soughtto replace metal components with light-weight plastics reinforced with carbon �-bres. Such materials, known as compos-ites, are generally 20-40% lighter according

Changes in the air

Aviation: Emerging technologies areushering in more fuel-e�cient,comfortable and exotic aircraft. Getready for the future of �ight

10 The future of �ight The Economist Technology Quarterly September 3rd 2011

1

The Economist Technology Quarterly September 3rd 2011 The future of �ight 11

2

1

to ATK, an aerospace company based inUtah that makes them for aircraft manu-facturers. Composites account for as muchas 15% of today’s airliners, but some next-generation aircraft will be more compositethan metal, including the Boeing 787(which enters service this year) and AirbusA350 (due in 2013).

New, lightweight ceramics will furtherreduce the need for metals in aircraft, saysJoy de Lisser, vice-president of ATK Aero-space Structures, the division developingthem. Ceramic composites can also with-stand hotter temperatures than metal al-loys can. Accordingly, they are beginningto replace some metal parts in jet enginesdeveloped by Snecma, a French engine-maker, and General Electric, an Americanmanufacturer. GE says it has shaved 136kg,or 3%, o� the weight of an engine that pro-pels the Boeing 787 using a ceramic-com-posites fan case and blade, a world �rst.

GE has also found a way to lighten met-al components, including some for en-gines, by �printing� rather than forgingthem. Known as 3D printing or additivemanufacturing, the process involves build-ing components by zapping a successionof thin layers of powdered metals with alaser or electron beam which melts andbonds the material. Precision is measuredin microns. Designers leave empty spacesinside some components, reducing theirweight by a �fth. The process is lessexpensive than hollowing outforged parts, says Luana Iorio, headof manufacturing technologies atGE Global Research’s lab in Niskay-una, New York. She reckons thatGE’s printed, hollow parts will beused in passenger aircraft withinabout three years.

In 2009 GE, working withNASA, America’s space agency,picked up work it had largely setaside in the 1980s on a radically dif-ferent sort of engine called an un-ducted fan. It combines the fuel ef-�ciency of a propeller engine withthe greater power and accelerationof a jet by using two rings of short,propeller-like rotors that spin inopposite directions in open air be-hind the jet housing. GE says theengine consumes almost a thirdless fuel than other designs. But it isloud, and if a rotor breaks it couldsmash into the fuselage.

Pratt & Whitney, another Amer-ican engine-maker, has devised adi�erent design that is far closer towidespread use. Called a �geared

turbofan� engine, it uses a gearbox, ratherthan a shaft, to transmit power from theturbine (which spins as hot gases blast outof the back) to the fan (which sucks in air atthe front). This allows the turbine to spinfaster than the fan, which is more e�cient.Called the PurePower PW1000G, it cutsfuel consumption (and noise) by about15%, says Paul Finklestein of Pratt & Whit-ney, saving about $400 per �ight hour.More than 1,200 of the engines have beenordered at an estimated $13m each. Deliv-eries begin in 2013. The �rm’s president,David Hess, has said the new engine coulddouble the size of the company, which hadsales of nearly $13 billion last year.

On most passenger jets, the wings andfuselage generate about 90% and 10% ofthe lift respectively. Working with fundingfrom NASA, aerospace engineers at theMassachusetts Institute of Technology(MIT) have designed an aeroplane with abody so fat, and wings so narrow, that thefuselage provides about a �fth of the air-craft’s lift. Its cross-section resembles thatof two partially joined bubbles. The �Dou-ble Bubble�, as it is called, looks awkward,but the team estimates that its designwould reduce fuel consumption by about70%. This is only partly because it would�y about 10% slower than today’s airliners.

Tail wings push the back of an aircraft

down, increasing drag, in order to lift itsnose up. The Double Bubble sports a wide,downward-sloping nose which air�owpushes up, so its tail wings can be muchsmaller. Conventional airframes requireheavy structural material to transfer the fu-selage’s weight laterally to landing gearand wheels under the wings. The MIT

team reduced the plane’s weight about 1%by fattening the aircraft’s body��essential-ly running the fuselage to the landinggear�, in the words of Mark Drela, theteam’s leader. The engines are mounted atthe back of the fuselage, rather than underthe wings. Air slipping along the fuselagemoves slower, so the engines ingest lessoxygen and burn less fuel.

Making planes more slippery

Changing the shape of an aircraft can bedone at a microscopic as well as a macro-scopic level. Aircraft paint, viewed with amicroscope, �looks like the Pyrenees�, saysPaul Booker, managing director of tripleO,a �rm based in Poole, England. His �rm hasdeveloped a way to reduce drag on aircraftby smoothing the painted surfaces with avery thin layer of acrylic resin that �lls intiny cracks. Britain’s easyJet, the �rst com-mercial carrier to use the product, hadthree airliners coated about 16 months ago.The airline has since coated �ve moreplanes and two other airlines have also

given it a go. Mr Booker says the ex-tra slipperiness cuts fuel consump-tion by around 1%, so that the coat-ing treatment pays for itself withina few months.

There may also be a way to cutaircraft drag by making some sur-faces less slippery. In research fund-ed by the European Union, Ales-sandro Bottaro of the University ofGenoa in Italy has devised smallkeratin bristles that mimic thesmallest type of bird feathers,known as coverts. Vibrating in thewind, the bristles create some drag.But they also reduce the wing’s slip-stream, an area of low-pressure tur-bulence that pulls back on thewing, and hence reduce drag. Afuzzy tennis ball �ies faster than abald one for the same reason, MrBottaro explains.

However perfectly an aircraft isbuilt, its full potential cannot beharnessed without a perfectly cal-culated trajectory. At most airports,tra�c controllers organise the ap-proach and landing order of incom-ing planes in their last half-hour or Airbus imagines in-�ight entertainment in 2050

�Changing the shape of an aircraft can be done at amicroscopic as well as a macroscopic level.�

so of �ight. As a result, pilots waste fuelslowing down and speeding up as they de-scend in staircase fashion. This and otherine�ciencies�such as circling whileawaiting a landing slot�will soon be great-ly reduced, thanks to a new sort of �ight-management software.

Such software crunches data on eachaircraft’s performance and other tra�c inthe air or at airports to determine the opti-mal �ight plan. The software can work out,for example, the exact rate at which a planeshould rise into thinner air (to reduce drag)as fuel burn makes it lighter. Aircraft cancollect and exchange atmospheric data tohelp each other �ne-tune trajectory andspeed. Crucially, the technology harnessesairliners’ ability to glide. With a favourablewind, a new airliner’s engines can be idledmore than 150km from an airport for a glid-ing descent to the runway.

A single such �green approach�, as it isknown, saves about 100kg of fuel, saysTorbjorn Henriksen, head of airline negoti-ations at Avinor, the operator of Norway’s19 commercial airports. Steve Fulton ofNaverus, a subsidiary of GE that designsand installs such systems, likens them to arailway track: aeroplanes do not deviatemore than a wingspan from their chartedcourses and touch down within ten sec-onds of the predicted time.

The airport at Brisbane, Australia, is theonly one that fully uses the system so far. Ithas reduced delays and cut noise in sur-rounding neighbourhoods by nearly athird. If adopted across Europe, fuel costs(and pollution) for internal �ights woulddrop by more than 8%, says Mr Fulton. Doz-ens of airports are adopting the technol-ogy, but the process requires a lot of instal-

lation and training. Avinorsays it will take at least anoth-er �ve years to deploy thetechnology widely in Norway.

Saving fuel is all very well,you may be thinking, butwhat can technology do to im-prove conditions for passen-gers? In the run-up to thisyear’s Paris Air Show in June,Airbus released its vision ofcreature comforts for the air-liner of 2050. Cabin wallshave been replaced with askeletal structure and trans-parent membrane. �Vitalis-ing� swivel seats mould to,and massage, each passen-ger’s body while harvesting itsheat to power individualsound pods, mood lighting

and holographic entertainment units. Itsounds great, even if Airbus’s vice-presi-dent of engineering, Charles Champion,acknowledges that much of the kit cannotbe built with today’s technology. He pointsout that in recent years the industry hasplaced a far higher priority on making air-craft more e�cient and comfortable than ithas on making them go faster.

Yet despite the withdrawal from serviceof Concorde in 2003, the dream of super-sonic �ight has not died. Dassault Avia-tion, a French �rm, and Aerion and Gulf-stream Aerospace, two Americancompanies, are among the �rms develop-ing technologies for private supersonicjets. Breaking the sound barrier generates asonic boom, so supersonic travel is heavilyrestricted over land. Tests by NASA with amodi�ed �ghter jet have shown that novelairframe shapes can reduce the boom. ButAerion reckons that a far better approach isto abandon e�orts to reduce the sonicboom and �y supersonic only over water.The company’s 8-to-12-seat SupersonicBusiness Jet, designed but not yet built,sports thin but broad �knife edge� wingsand other aerodynamic features thatproduce less drag than compet-ing designs, says DouglasNichols of Aerion.

Around 50 potential customers have putdown a $250,000 deposit for the $80m jet,which would �y at 1.6 times the speed ofsound (Mach 1.6). Aerion does not yet havea manufacturing partner, however.

America’s armed forces see potential inhypersonic aircraft, which �y at Mach 5 orfaster using a type of engine known as ascramjet. HTV-2, an unmanned hyperson-ic aircraft designed to travel at Mach 20,failed during a test �ight last month. An-other hypersonic craft, the X-51A Wave-Rider developed by Boeing, has fared littlebetter. Of the two WaveRiders tested, bothfor short distances over the Paci�c, onefailed. But Joe Vogel, the project manager atBoeing, says the technology has �crossedover the threshold� into hypersonic �ight.He reckons that scramjets might one daypower civilian aircraft.

Some military types have enthusedthat, before then, hypersonic troop carrierscould be built. But Robert Mercier, a seniortechnology o�cial in the Air Force Re-search Laboratory’s aerospace propulsiondivision, notes with understatement thatparachuting into the trailing vortices ofsuch an aircraft would make for a roughride. It is more likely, he says, that a hyper-sonic aircraft would be used as a high-speed cruise missile, to deliver a surprisehammer-blow behind enemy lines. Usinga long-range ballistic missile to do the jobwould be risky, as its launch could be mis-taken for an imminent nuclear strike.

Might the idea of near-hypersonic pas-senger aircraft, which has lain dormant fora few years, be coming back? At this year’sParis Air Show, EADS, the parent companyof Airbus, revealed a concept design for anaircraft called the Zero Emission High Su-personic Transport (ZEHST), devised inconjunction with Japanese researchers. Ithas three separate kinds of engine: ordin-ary jet engines (running on biofuels madefrom seaweed or algae) for take-o�, rocket

engines to accelerate to Mach2.5, and ramjets to reach Mach 4.

The aircraft would carry 50-100passengers and would travel from

Paris to Tokyo in around 2.5 hours, rath-er than the 11hours it takes today.

Even its designers admit that theZEHST is unlikely to be �ying before 2040.But �the future of air travel will look some-thing like the ZEHST,� declared Jean Botti,director-general for technology and inno-vation at EADS. It sounds fanciful. But sotoo, not that long ago, did rapid and routineintercontinental air travel. In aviation,what sounds outlandish today may becommonplace tomorrow. 7

Aerion’s design (top) and the Double Bubble (below)

ZEHST the job for travellers in a hurry

2

12 The future of �ight The Economist Technology Quarterly September 3rd 2011

FOR a man who claims to lack expertisein the �eld, Bruce Bueno de Mesquita,

an academic at New York University, hasmade some impressively accurate politicalforecasts. In May 2010 he predicted thatEgypt’s president, Hosni Mubarak, wouldfall from power within a year. Ninemonths later Mr Mubarak �ed Cairo amidmassive street protests. In February 2008Mr Bueno de Mesquita predicted that Paki-stan’s president, Pervez Musharraf, wouldleave o�ce by the end of summer. He wasgone before September. Five years beforethe death of Iran’s Ayatollah Khomeini in1989, Mr Bueno de Mesquita correctlynamed his successor, and, since then, hasmade hundreds of prescient forecasts as aconsultant both to foreign governmentsand to America’s State Department, Penta-gon and intelligence agencies. What is thesecret of his success? �I don’t have in-sights�the game does,� he says.

Mr Bueno de Mesquita’s �game� is acomputer model he developed that uses abranch of mathematics called game the-ory, which is often used by economists, towork out how events will unfold as peopleand organisations act in what they per-ceive to be their best interests. Numericalvalues are placed on the goals, motivationsand in�uence of �players��negotiators,business leaders, political parties and orga-nisations of all stripes, and, in some cases,their o�cials and supporters. The comput-er model then considers the options opento the various players, determines theirlikely course of action, evaluates their abil-ity to in�uence others and hence predictsthe course of events. Mr Mubarak’s in�u-ence, for example, waned as cuts in Ameri-can aid threatened his ability to keep cro-nies in the army and security forces happy.Underemployed citizens then realised thatdisgruntled o�cials would be less willingto use violence to put down street protestsagainst the ailing dictator.

Mesquita & Roundell, Mr Bueno deMesquita’s company, is just one of several

consulting out�ts that run such computersimulations for law �rms, companies andgovernments. Most decision-making ad-vice is political, in the broadest sense of theword�how best to outfox a trial prosecu-tor, sway a jury, win support from share-holders or woo alienated voters by shuf-�ing a political coalition and makinglegislative concessions.

But feeding software with good data onall the players involved is especially trickyfor political matters. Reinier van Oosten ofDecide, a Dutch �rm that models politicalnegotiations and vote-trading in EuropeanUnion institutions, notes that forecasts goastray when people unexpectedly give into �non-rational emotions�, such as ha-tred, rather than pursuing what is appar-ently in their best interests. Sorting outpeople’s motivations is much easier, how-ever, when making money is the main ob-ject. Accordingly, modelling behaviour us-ing game theory is proving especiallyuseful when applied to economics.

Follow the money

Modelling auctions has proved especiallysuccessful, says Robert Aumann, an aca-demic at the Hebrew University of Jerusa-lem who received a Nobel prize in 2005 forhis work in game-theory economics. Bids,being quanti�ed, facilitate analysis, andpredicting the right answer can be very lu-crative. Consulting �rms are popping up tohelp clients design pro�table auctions orwin them less expensively. In the run-up toan online auction in 2006 of radio-spec-trum licences by America’s Federal Com-munications Commission, Paul Milgrom, a

consultant and Stanford University profes-sor, customised his game-theory softwareto assist a consortium of bidders. The re-sult was a triumph.

When the auction began, Dr Milgrom’ssoftware tracked competitors’ bids to esti-mate their budgets for the 1,132 licences ono�er. Crucially, the software estimated thesecret values bidders placed on speci�c li-cences and determined that certain big li-cences were being overvalued. It directedDr Milgrom’s clients to obtain a patchworkof smaller, less expensive licences instead.Two of his clients, Time Warner and Com-cast, paid about a third less than their com-petitors for equivalent spectrum, saving al-most $1.2 billion.

Advances in game theory have �pickedup dramatically� in recent years as it hasbecome apparent that failing to do aproper analysis can be costly, says SergiuHart, a colleague of Dr Aumann’s at He-brew University. For example, a few yearsago Israel’s government added a noveltwist to an auction of oil-re�nery facilities.To encourage more and higher bids, thegovernment o�ered a $12m prize to the sec-ond-highest bidder. It was an expensivemistake. Without the incentive, the highestbid would have been about $12m higher,an analysis showed�participants bid lowbecause the loser would strike it rich. Com-bine that sum with the prize payout, andthe government’s loss amounted toroughly $24m. The conclusion, then, is�don’t presume you know what the sol-ution is� without help from modellingsoftware, says Brad Miller, senior modellerat Charles River Associates, a consultancy

Game theoryin practice

Computing: Software that modelshuman behaviour can makeforecasts, outfox rivals andtransform negotiations

The Economist Technology Quarterly September 3rd 2011 Modelling behaviour 13

1

14 Modelling behaviour The Economist Technology Quarterly September 3rd 2011

2 in Boston. It designs game-theory softwareto model industrial auctions and the plot-ting of corporate mergers and acquisitions.

Software is not always needed. A stu-dent at Hebrew University, for example,demonstrated the Israeli government’s$24m loss using pen and paper. It took himabout two days, however, according to aprofessor there. Software, naturally, is farfaster. But gathering and handling the nec-essary data can require expensive exper-tise or training. Decide, the Dutch consul-tancy, usually charges �20,000-70,000($28,000-100,000) to solve a problem us-ing its software, called DCSim, because itmust �rst conduct lengthy interviews withexperts. Its clients include government bo-dies in the Netherlands and abroad, andbig companies including IBM, a computergiant, and ABN AMRO, a Dutch bank.

PA Consulting, a British �rm, designsbespoke models to help its clients solvespeci�c problems in areas as diverse aspharmaceuticals, fossil-fuel energy andthe production of television shows. Britishgovernment agencies have asked PA Con-sulting to build models to test regulatoryschemes and zoning rules. To give a simpleexample: if two shrewd, competing ice-cream sellers share a long beach, they willset up stalls back-to-back in the middle andstay put, explains Stephen Black, a model-ler in the �rm’s London headquarters. Un-fortunately for potential customers at thefar ends of the beach, each seller preventsthe other from relocating�no other spotwould be closer to more people. Introducea third seller, however, and the sti�ingequilibrium is broken as a series of market-energising relocations and pricing changeskick in. The use of modelling makes busi-ness clients more inclined to adopt longer-term strategies, Dr Black says.

But game-theory software can alsowork well outside the sphere of econom-ics. In 2007 America’s military providedMr Bueno de Mesquita with classi�ed in-formation to enable him to model the po-litical impact of moving an aircraft carrierclose to North Korea (he will not reveal the�ndings). Game-theory software can evenhelp locate a terrorist’s hideout. To run sim-ulations, Guillermo Owen of the NavalPostgraduate School in Monterey, Califor-nia, uses intelligence data from the US AirForce to estimate on a 100-point scale theimportance a wanted man attaches to hislikes (�shing, say) and priorities (remain-ing hidden or, at greater risk of discovery,recruiting suicide-bombers). Such factorsdetermine where and how terrorists de-cide to live. Game-theory software played

an important role in �nding Osama binLaden’s hideout in Abbottabad, Pakistan,says Mr Owen.

Where is all this heading? Alongsidethe arms race of increasingly elaboratemodelling software, there are also e�ortsto develop software that can assist in nego-tation and mediation. Two decades agoClara Ponsatí, a Spanish academic, cameup with a clever idea while pondering thearduous Israeli-Palestinian peace process.As negotiators everywhere know, the �rstside to disclose all that it is willing to sacri-�ce (or pay) loses considerable bargainingpower. Bereft of leverage, it can be pushedback to its bottom line by a clever oppo-nent. But if neither side reveals the conces-sions it is prepared to make, negotiationscan stall or collapse. In a paper publishedin 1992, Dr Ponsatí described how softwarecould be designed to break the impasse.

Di�cult negotiations can often benudged along by neutral mediators, espe-cially if they are entrusted with the secret

bottom lines of all parties. Dr Ponsatí’sidea was that if a human mediator was nottrusted, a�ordable or available, a comput-er could do the job instead. Negotiatingparties would give the software con�den-tial information on their bargaining posi-tions after each round of talks. Once posi-tions on both sides were no longermutually exclusive, the software wouldsplit the di�erence and propose an agree-ment. Dr Ponsatí, now head of the Instituteof Economic Analysis at the AutonomousUniversity of Barcelona, says such �media-tion machines could lubricate negotia-tions by unlocking information that wouldotherwise be withheld from an opponentor human mediator.

Such software is now emerging. Barry

O’Neill, a game theorist at the Universityof California, Los Angeles, describes howit can facilitate divorce settlements. A hus-band and wife are each given a number ofpoints which they secretly allocate tohousehold assets they desire. The wifemay inform the software that her valua-tion of the family car is, say, 15 points. If thehusband puts the car’s value at 10 points,he cannot later claim that he deservesmore compensation for not getting the carthan she would be entitled to.

Predicting an end to con�ict

Participants need to be sure that such me-diation technology is fully neutral. Forlarge deals, audit �rms closely monitor thedevelopment and use of such software toensure that no party secretly obtains infor-mation about another’s bargaining posi-tions, says Benny Moldovanu, a gametheorist at the University of Bonn. He ad-vises �rms that design negotiation soft-ware for privatisation schemes and whole-sale-electricity markets. This approachwill spread to other utility markets, such aswater, he believes.

Could software-based mediationspread from divorce settlements and utili-ty pricing to resolving political and mili-tary disputes? Game theorists, who con-sider all these to be variations of the samekind of problem, have developed an in-triguing conceptual model of war. The�principle of convergence, as it is known,holds that armed con�ict is, in essence, aninformation-gathering exercise. Belliger-ents �ght to determine the militarystrength and political resolve of their op-ponents; when all sides have �convergedon accurate and identical assessments, asurrender or peace deal can be hammeredout. Each belligerent has a strong motiva-tion to hit the enemy hard to show that itvalues victory very highly. Such a modelmight be said to re�ect poorly on humannature. But some game theorists believethat the model could be harnessed to makediplomatic negotiations a more viablesubstitute for armed con�ict.

Today’s game-theory software is notyet su�ciently advanced to mediate be-tween warring countries. But one day op-ponents on the brink of war might betempted to use it to exchange informationwithout having to kill and die for it. Theycould learn how a war would turn out,skip the �ghting and strike a deal, MrBueno de Mesquita suggests. Over-opti-mistic, perhaps�but he does have ratheran impressive track record when it comesto predicting the future. 7

�The use of modelling makes business clients moreinclined to adopt longer-term strategies.�

IT IS di�cult to imagine a world withoutelectric motors. They are everywhere:

spinning discs inside DVD players, power-ing washing machines and fans, keepingtrains and escalators moving, starting carengines, waving windscreen wipers andmaking phones vibrate when new mes-sages arrive. If you want to transform elec-trical energy into physical motion, you usea motor. But that could be about to changeas a lighter, more �exible and less power-hungry alternative starts to muscle in,quite literally, on the motor’s patch.

Researchers have been experimentingwith arti�cial muscles�electroactive po-lymers that can expand or contract when avoltage is applied�for decades. But untilrecently researchers assumed that theywould mainly be used, as their name sug-gests, to move the limbs of robots. It turnsout, however, that with a few design twiststhey can be applied to many other tasks.Arti�cial muscles can provide force-feed-back wobbles and clicks in mobile devices,for example, or move components insideautofocus camera lenses. With the rightcon�guration of muscles it is even possibleto produce rotary motion. And, just likemotors, arti�cial muscles can also be used

in reverse as generators, transformingmovement into electricity.

There are two main types of electro-active polymer (EAP), which are usually re-ferred to as ionic and dielectric (see dia-gram on next page). An ionic EAP consistsof a sponge-like polymer, soaked in a liq-uid electrolyte containing free-�oatingpositive ions, and sandwiched betweentwo electrodes. When a voltage is applied,the positively charged ions move towardsthe negative electrode, carrying electrolyt-ic �uid with them. The sponge swells upon one side and shrinks on the other,which causes it to move. This kind of EAP

can bend a lot, producing a large move-ment when a voltage is applied. A singlemuscle of this kind can be used to power awindscreen wiper, for example, by makingit bend one way and then the other. Two ormore such muscles can be con�gured as agripper to pick up objects. But the electro-lyte can evaporate, causing the muscle todry out when exposed to the air. And ionicEAPs revert to their original shape as soonas the applied voltage is removed.

Dielectric EAPs, by contrast, are easierto work with and have much more pullingand pushing power. They consist of a �ex-ible polymer sandwiched between twoelectrodes. When a voltage is applied, posi-tive charge builds up on one electrode andnegative charge on the other, creating an at-tractive force that squeezes the polymer.As the polymer contracts in one direction,it expands in another, and these move-

ments can do useful work. Unlike a rotarymotor, which requires a crank arrange-ment to produce linear motion, an arti�-cial muscle can do it directly, for examplein a smartphone’s autofocus mechanism,which requires tiny, precise linear move-ments. Moreover, a stack of several dielec-tric EAPs can generate substantial forcesusing little power. They are also light,which is one reason why America’s spaceagency, NASA, is so interested in them, saysYoseph Bar-Cohen, a physicist and pioneerin the �eld at NASA’s Jet Propulsion Lab-oratory in Pasadena, California.

Brain and brawn

Using arti�cial muscles in place of motorscould reduce the size and weight of spaceprobes and robots. �The next-generationspace telescope will have to be signi�cant-ly bigger,� says Dr Bar-Cohen. But doublingthe diameter of the telescope’s mirror,without changing its design, would in-crease its weight fourfold. So one idea be-ing explored is to make in�atable spacetelescopes that use arti�cial muscles to un-furl the mirror and then, once it is fully ex-tended, to adjust its optics, much like themuscles in an eye, he says.

Just how strong are these muscles? In2005 Dr Bar-Cohen decided to �nd out bypitting a human against a trio of roboticarms, based on EAPs, in an arm-wrestlingcontest. In theory it should be possible tomake EAPs as strong as human muscle, hesays. (By weight, EAPs are 40 times strong

Muscling in on motors

Materials science: Electroactivepolymers, also known as arti�cialmuscles, could replace electricmotors in some applications

The Economist Technology Quarterly September 3rd 2011 Inside story 15

1

16 Inside story The Economist Technology Quarterly September 3rd 2011

2 er than human muscle, according to re-searchers at the Auckland BioengineeringInstitute.) Even so, the person representinghumanity in the contest�a 17-year-oldhigh-school student from San Diego calledPanna Felsen�took a matter of seconds tocrush each of her three robotic opponents.To Dr Bar-Cohen this humiliating defeat isa clear sign that it is not how big your mus-cles are that counts, it is how you use them.

And it is precisely by using dielectricEAPs in a clever way that, earlier this year, agroup of researchers in New Zealand man-aged to get muscles to do something thatno living creature has managed to pull o�:turn a wheel. With the possible exceptionof the propeller-like �agella of bacteria,there are no examples of organisms thatuse muscles to create continuous rotarymotion. But a team led by Iain Anderson,the head of the Auckland BioengineeringInstitute’s Biomimetics Lab, has built anentire menagerie of muscle-powered mo-tors. Although there are di�erences be-tween them, the basic principle for each isthe same. �The muscles manipulate theshaft of the motor like a �nger and thumbwould hold a pencil,� says Dr Anderson.

These rotary motors (see picture on pre-vious page) resemble bicycle wheels thathave had their spokes replaced with thin,black slivers of EAP. With at least six permotor, working as opposing pairs, thesemuscles are positioned between the outerrim and the central driveshaft. To make theshaft turn, the muscles work in concert,rhythmically contracting, one pair after an-other. As they do so, each pair applies pres-sure to a soft ring around the driveshaft.The pulsating muscles collectively andcontinuously pinch the shaft and apply arotational force, causing it to turn.

Strictly speaking this is not the �rst timearti�cial muscles have been used to turn awheel, says Dr Anderson. But it was previ-ously done using a ratcheting mechanismwhich required heavy solid parts such as aclutch. His team’s new design eliminatesthe need for anything rigid, such as bear-ings or gears. The breakthrough idea was togrip the shaft from both sides, which madebearings redundant, he says.

Dr Anderson concedes that these mus-cle motors are unlikely to start poweringcars or trains any time soon. They cannotcome close to electric motors in revolu-tions per minute, for a start. �But it hasopened up a new design space,� he says.For example, by using multiple motors onthe same shaft it is possible to get morethan one degree of freedom. �Not only canthey turn the shaft but they can lift it as

well,� says Dr Anderson. This is hard toachieve with rigid bearings and could leadto strange new modes of locomotion, al-lowing robots to walk or roll, dependingupon which is more appropriate.

As well as working on rotary motion,Dr Anderson’s group is also investigatingthe use of EAPs as electrical generators.The pioneering work in this �eld was doneby researchers at SRI International, former-ly known as the Stanford Research Insti-tute, a non-pro�t research out�t spun outof Stanford University that is based inMenlo Park, California, and holds many ofthe patents on EAPs.

Although SRI researchers have shownthat it is possible to use arti�cial muscles toconvert kinetic energy into electricity, get-ting the process started has always been abit of a problem, because dielectric EAPsneed a small initial charge in each cycle.SRI solved this problem using an externalpriming circuit, but Dr Anderson’s teamhas found a way to integrate the primer

into the EAP itself, making the device bothsoft and self-contained. These could beused to produce energy on a large scale, inwave-power generators, or on a smallscale, by integrating small, lightweight de-vices into the heels of shoes. Previous re-search by SRI estimated that this arrange-ment could generate around one watt ofpower from normal walking.

Pressing forward

Using your shoes to recharge your phoneas you walk has obvious appeal, thoughquite how the two would be connected isunclear. But the �rst product intended toput arti�cial muscles into consumers’hands will be aimed at video gamers. Acompany called Arti�cial Muscle based inSunnyvale, California, which was spunout of SRI International, is about to launcha product called the Mophie Pulse, a �hap-tic� case for the iPod touch. It uses a tech-nology called ViviTouch which is designedto provide gamers with realistic tactilefeedback, letting them feel every tremorand explosion in a game or conveyingmore subtle sensations, like a heartbeat,says Marcus Rosenthal of Arti�cial Muscle.

Existing haptic devices use rotary mo-tors with o�-centre weights to generatesuch tremors. But this approach only al-lows vibrations within a narrow band offrequencies to be generated, says Mr Ro-senthal. By using a pair of arti�cial musclesto move a weight and generate tremors, his�rm’s technology provides far more subtlecontrol. �It’s the di�erence between a door-bell and a speaker,� he says. And it usesone-third of the energy of a conventionalhaptic device. Beyond gaming, the technol-ogy can also be used to provide better sen-sory feedback when using touch screens.Existing devices can emulate a clickingsensation when the user touches an on-screen button, for example, but Arti�cialMuscle’s approach can create many di�er-ent types of click, depending on the kindof control being pressed.

Arti�cial muscles seem unlikely ever tobe able to compete with hydraulic actua-tors for strength, or combustion enginesfor speed and torque. But in some applica-tions, it seems, they could give the venera-ble electric motor a run for its money. Andalthough they may yet �nd uses in Marsrovers and space telescopes, in the shortterm they would appear to have more po-tential inside mobile phones, cameras,game controllers and other consumer-electronics products. Having originated inthe �eld of robotics, arti�cial muscles’ realstrength, it seems, may lie elsewhere. 7

+– +–

Electrodes

Electrolyte

Positive ions

Off On

+–

+– – – – – – – –

+ + + + + + +

Electrodes

Flexible polymer

On

Off

How artificial muscles workAn ionic artificial muscle consists of a permeablepolymer membrane, immersed in a positivelycharged electrolyte and sandwiched between twoelectrodes. Positive ions can move freely withinthe membrane, but negative ions cannot.

When a voltage is applied, positive ions in theelectrolyte drift towards the negative electrode,carrying electrolytic fluid along with them. Thepolymer expands at the negative electrode andshrinks at the positive electrode, causing it to move.

In a dielectric artificial muscle, a flexible polymeris sandwiched between two electrodes.

When a voltage is applied, positive charges buildup on one electrode, and negative charges on theother. Because unlike charges attract, theelectrodes move towards each other, squeezingthe polymer and causing it to move.

Pumping ions

Sources: Auckland Bioengineering Institute; The Economist

�Using arti�cial muscles instead of motors could cutthe size and weight of space probes and robots.�

THE idea of moving objects with thepower of the mind has fascinated man-

kind for millennia. At �rst it was the prov-ince of gods, then sorcerers and witches. Inthe late 19th century psychokinesis, as thetrick then came to be known, became a le-gitimate object of study, as part of the na-scent �eld of parapsychology, before fall-ing into disrepute in the arch-rationalist20th century. Since the 1990s, however, ithas seen something of a revival, under amore scienti�cally acceptable guise.

There is nothing particularly magicalabout moving things with thoughts. Hu-man beings perform the feat every timethey move a limb, or breathe, by sendingelectrical impulses to appropriate muscles.If these electrical signals could be detectedand interpreted, the argument goes, thereis in principle no reason why they couldnot be used to steer objects other than thethinker’s own body. Indeed, over the pasttwo decades brain-computer interfaces(BCIs) which use electrodes implanted inthe skull have enabled paralysed patientsto control computer cursors, robotic armsand wheelchairs.

Now, though, non-invasive BCIs, whereelectrodes sit on the scalp instead of bur-rowing through it, are �nally becoming arealistic alternative to the complicated sur-gical procedure that implants necessitate.

Electroencephalography (EEG), whichmeasures electrical activity along thescalp, has long been used clinically to diag-nose epilepsy, comas and brain death, andas a research tool in neuroscience and cog-nitive psychology. Last year an Austriancompany called Guger Technologies re-leased a system that uses EEG to allow par-alysed patients to type. The system high-lights letters one by one on a grid. Whenthe desired letter comes up an EEG headsetpicks up the brain activity associated withrecognising it. At �ve to ten characters perminute the process is slow and laborious,but it o�ers patients a way to communicatewith others. The device can also be used toattract a minder’s attention, to get a com-puter to read out a text or to send com-mands to external devices such as a TV.

Follow the jumping blue dot

A team at the University of Zaragoza, inSpain, led by Javier Minguez, has devel-oped some promising prototypes usingsimilar principles. One is a wheelchair thatgenerates a 3D map on a screen. A blue dotjumps between places on the map andwhen it hits the desired destination, theEEG electrodes pick this up and steer thewheelchair towards it. Another lets severe-ly disabled patients use a real-time videocamera connected to the internet, allowing

them, for instance, to attend a family re-union virtually. Dr Minguez is also part of agroup developing a robot arm that couldhelp patients reach for a drink or feedthemselves. The University of Zaragozahas set up a spin-o� to commercialise theresearch and Dr Minguez expects a fully�edged device by 2014.

The number of electrodes needed to getuseful EEG results is steadily decreasing.Several EEG devices have been released inrecent years sporting fewer sensors thanthe high-end kit, notably for use in videogaming. In May an organisation calledThought-Wired, based in New Zealand,showed o� a system that uses a commer-cially available 14-sensor headset calledthe Emotiv EPOC, which costs around$300 and was designed with gaming inmind, as a smart-home controller. It en-ables disabled people to control lights,make phone calls, operate air conditionersand perform other simple tasks.

Early e�orts to use rudimentary head-sets for computer games were not exactlyriveting. Two headset-based toys releasedin 2009�Uncle Milton’s Star Wars ForceTrainer and Mattel’s Mind�ex�used play-ers’ concentration and relaxation levels,gauged by comparing brain waves of dif-ferent frequencies, to make a ball rise or fallusing air�ow from a fan.

Since then, however, headset priceshave tumbled, leading to several niftier of-ferings. For instance, researchers in Ger-many have used Emotiv’s headset to builda device which allows a driver to tell a carto go left or right, accelerate or slow downby thinking about it. In July Toyota un-veiled a concept bicycle, the PXP, with athought-controlled gear-shift mechanism.

Meanwhile, a British company calledMyndplay is creating short interactive�lms in which viewers step into the shoesof the protagonists and in�uence the plotby focusing or relaxing. MohammedAzam, Myndplay’s boss, thinks the tech-nology has potential beyond entertain-ment. For example, he has been workingon an app in which users must keep theircool in various dating scenarios. Otherprojects include training for job interviewsand a virtual-therapy app to help treatstress and phobias. Myndplay will unveilits latest product at the Tokyo Games Showon September 15th.

In April NeuroSky, a Californian makerof EEG headgear, launched its $100 Mind-Wave headset, which reduces the numberof sensors to just one. This is possible usingalgorithms that �lter out interferencecaused by external electrical activity. The

Put your thinking cap on

Consumer electronics: Once the stu� of fables, hoaxes and science �ction,controlling things via thought alone is fast becoming a reality

The Economist Technology Quarterly September 3rd 2011 Brainwave controllers 17

1

18 Brainwave controllers The Economist Technology Quarterly September 3rd 2011

2 American Olympic archery team has usedthe company’s technology to help athletesgain and retain composure. B-Bridge Inter-national, NeuroSky’s partner, haslaunched BrainAthlete, a golf visorequipped with two EEG sensors. It letsplayers plot their concentration levels atimportant moments in a game�during agolf swing, say�on a chart and, in so doing,helps them learn to ignore distractions.

There are some decidedly frivolous ap-plications for EEG technology, too, such asa pair of cat ears on a hairband that standup when the wearer is concentrating andlie down when he or she is relaxed. Videosof the one-o� toy proved such a hit onlinethat its Japanese creator, Neurowear, isconsidering mass production.

Not all fun and games

On a more serious note, earlier this yearMyndplay games and interactive �lmswere used as part of a scheme to reduce re-o�ending rates and anti-social behaviouramong prison inmates aged 18 to 25. �Wefound it to be an excellent tool to showthem how they can develop greater con-trol over their thought processes,� saysDavid Apparicio, a magistrate who spear-headed the programme, called Chrysalis.He hopes to commission customised con-tent to test young o�enders’ reactions inscenarios such as being challenged to a�ght or invited to commit a crime.

SpeedMath, a game that comes withthe NeuroSky headset, shows studentswhich types of task require most mentale�ort on their part, with the aim of encour-aging more focused work. Jack Mostow, aresearcher at Carnegie Mellon Universityin Pittsburgh, has used NeuroSky’s head-sets to distinguish between the patterns ofbrain activity involved in reading easy anddi�cult sentences. This, he thinks, meansthat EEG could be used to tailor teaching to

converted EEG readings of people drinkingvodka cocktails into colourful, real-time vi-sual displays. Chris Meyer, a graphic de-signer from California, has used the tech-nology to modify his Nerf gun so that hecan shoot foam bullets using thoughtalone. Others have gone even further. Onegroup of artists from Toronto used a Neu-roSky headset connected to a gas canisterto generate seven-metre (23-foot) �ames.

As might be expected, not everyoneshares the enthusiasm for supplantingmankind’s traditional, arm’s length rela-tionship with technology with a deeper,BCI-mediated sort. Jens Clausen, a medicalethicist at Tübingen University, warns thatexcessive use of BCI for gaming could alterbrain activity in ways that conventionalgaming does not, and that as yet are poorlyunderstood. And blurring the distinctionbetween thinking about an act and actual-ly performing it raises some tricky moraland philosophical questions.

Yet as it stands, the technology seemspoised for a period of rapid developmentwhich both the needy and thrill-seekersare bound to greet with cheers. As Tan Le,co-founder of Emotiv, the headset-maker,told the TED conference last year, �We areonly really scratching the surface of what ispossible.� Those scratches are, however,getting deeper all the time. 7

pupils’ needs.Market-research com-

panies, too, havejumped on the BCI

bandwagon. Propo-nents of the newish�eld of neuromarket-ing trumpet non-inva-sive BCI as a way forcompanies to gain in-sights into how con-sumers perceive their

products, packaging oradvertisements. In

March NeuroFocus, anoth-er California-based out�t,

demonstrated what it claims is the �rstmedical-grade dry EEG headset. The com-pany, which was recently bought by Niel-sen, a big research �rm, has been usingconventional EEG devices (which requiregooey conductivity gels), in its neuromar-keting activities for around �ve years.Mynd, the �rm’s new wireless headset,will have between 32 and 64 sensors, andwill begin commercial use from Septem-ber. A.K. Pradeep, the company’s boss, saysthat by measuring attention, engagementand memory retention, its equipment cangauge purchase intent, novelty factor andcomprehension of an advertisement.

Others dismiss neuromarketing as junkscience, at least for now. Although there isevidence that EEG-based metrics can re-�ect cognitive processes of interest to mar-keters, says Phil Harris, who lectures onneuromarketing at the University of Mel-bourne, no published evidence so far indi-cates that portable dry-electrode systemso�er a reliable glimpse of these processes.The Advertising Research Foundation(ARF), an industry body, recently carriedout an evaluation of neuromarketing tech-nologies. NeuroFocus and EmSense, an-other company that uses dry EEG headsetsin market research, declined to take part.Dr Pradeep says he thinks the methods theARF used in its appraisal were �awed. Theeight companies that did participate,which were asked to analyse a campaignfor a toothpaste brand, came up withwidely varying conclusions.

Mind-blowing

Artists using BCI face no suchniggles. At the Future-Everything festival,which was held in Mayin Manchester, MarcosLutyens created an in-teractive multimediainstallation which

Simpler sensors: Headsets from Emotiv(above) and NeuroSky (below)

The Economist Technology Quarterly September 3rd 2011 Brain scan 19

�SOFTWARE is eating the world,�proclaims Marc Andreessen, the

40-year-old co-founder of AndreessenHorowitz, a venture-capital �rm in SiliconValley that has leapt to prominence sincehe set it up in mid-2009 with his partner,Ben Horowitz. That alimentary analogy isshorthand, in Andreessen-speak, for thephenomenon in which industry afterindustry, from media to �nancial servicesto health care, is being chewed up by therise of the internet and the spread ofsmartphones, tablet computers and otherfancy electronic devices. Mr Andreessenand his colleagues are doing their best tospeed up this digital digestion process�and make money from it as they do so.

Andreessen Horowitz has raised pilesof cash from investors�it has some $1.2billion under management�and has beeneagerly putting the money to work both inlarge deals, such as a $50m investment inSkype, an internet-calling service recentlyacquired by Microsoft, and in a host ofsmaller companies, such as TinyCo, amaker of mobile games. It has also takenstakes in several of the biggest social-networking �rms, including Twitter, Face-book and Foursquare (a service that letspeople broadcast their whereabouts totheir friends). Along the way it has attract-ed some prominent supporters. LarrySummers, a former treasury secretary, is aspecial adviser to the �rm and MichaelOvitz, a former Hollywood power-broker,is among its investors.

Some rivals argue that by making bigbets on relatively mature companies suchas Facebook, Mr Andreessen’s �rm isacting more like a private-equity �rm thanas a nurturer of �edgling businesses�andis contributing to a bubble in tech valua-tions too. �They’re behaving in ways thatwill not be helpful to them in the longrun,� gripes a �nancier at a competingventure �rm, who insists on anonymityfor fear of alienating Andreessen Horo-witz’s in�uential founders.

Pooh-poohing such criticisms, MrAndreessen argues that �growth� in-vestments, such as the one in Skype�which was sold to Microsoft for $8.5 bil-lion in May, netting Andreessen Horowitza return of over three times its originalstake�make sense because profound

changes in the technological landscapemean some relatively big companies canstill grow to many times their current size.He reckons that talk of overheated valua-tions among social-media �rms is beingdriven by people who got their �ngersburned in the dotcom bust and can’t seethat the world has changed since then.�All this bubble stu� is people �ghting thelast war,� he says.

Mr Andreessen is also frustrated withCassandras who occasionally predict thatinnovation in computer science is prettymuch over. Andreessen Horowitz’s part-ners believe there are still plenty more�black swans��ideas with the potential totrigger dramatic changes in technology�tocome in computing, which explains whythey have resisted the temptation to copyother big venture out�ts that have div-ersi�ed into new areas such as biotechand clean tech. �This is an evergreen area.Just when you think computer science isstabilising, everything changes,� he says.

For instance, he believes that net-working and storage technology is aboutto go through the same kind of funda-mental transition that the server businessexperienced in the late 1990s, when ex-pensive, proprietary servers were re-placed by much cheaper ones that usednew technology. That shift made possiblethe explosive growth of �rms such asGoogle and Facebook, who bought largenumbers of cheap servers to power theirbusinesses. Mr Andreessen reckons asimilar change in the networking andstorage world will lead to the creation ofmany more new companies.

He is also convinced that there will bedramatic changes in the realm of personaltechnology. One of the companies thatAndreessen Horowitz has invested in isJawbone. Best known for its Bluetooth-equipped headsets and portable speakers,the �rm is developing plans for a range ofwearable smart devices that operate on asingle software platform, or �body-areanetwork�. �Jawbone is the new Sony,�claims Mr Andreessen, who predicts thatits future products will prove wildly suc-cessful as people carry more and morenetworked gadgets around with them.

From boom to bust

It is tempting to discount such a grandioseclaim as typical venture-capital pu�ery.But Mr Andreessen is hardly a typicalventure capitalist. Raised in small townsin Iowa and Wisconsin, he started playingaround on the internet while at universityand co-created Mosaic, which became the

Disrupting the disrupters

Marc Andreessen made his nametaking on Microsoft in the browserwars. Now he is stirring things upagain as a venture capitalist

1

20 Brain scan The Economist Technology Quarterly September 3rd 2011

�rst widely used web browser. Aftermoving to Silicon Valley, he started Net-scape Communications when he was 22years old. Its stockmarket �otation in 1995marked the beginning of the dotcomboom and made Mr Andreessen a celebri-ty in the business world.

Having at �rst dismissed Netscape,Microsoft sought to crush the �edglingcompany, whose browser posed a threatto the dominance of Microsoft’s Windowsplatform. Mr Andreessen maintains thatmost big companies are painfully slow toreact to upstarts that might threaten theirbusiness�a point made in Clayton Chris-tensen’s book, �The Innovator’s Dilem-ma�, which is one of the few business-school texts Mr Andreessen thinks isworth reading. But he admits Microsoft�did a remarkably good job� in the 1990s.

After a bruising battle a much-dimin-ished Netscape was sold to AOL in 1999and Mr Andreessen went on to foundLoudcloud, a cloud-computing �rm, withMr Horowitz and other executives. ButLoudcloud was soon caught up in thefallout from the dotcom bust. To survive, itshed sta�, renamed itself Opsware andfocused on software development beforebeing sold to Hewlett-Packard for $1.6billion in 2007. Mr Andreessen then spentsome time as an angel investor beforelaunching Andreessen Horowitz.

Mr Andreessen’s own experience aspractising entrepreneur makes him ideallyplaced to counsel the bosses of start-upsthat his �rm has funded, including MarkZuckerberg of Facebook and Mark Pincusof Zynga, a social-gaming company. MrAndreessen seems especially fond ofwhat he calls �founder CEOs�, perhapsbecause he was once one himself. Manyventure �rms tend to back young entrepre-neurs for a period before replacing themwith professional managers. But Mr An-dreessen argues that founders who stuckwith their businesses for a long whilewere often the ones who created many ofthe biggest successes in technology, in-cluding Microsoft (Bill Gates), Amazon(Je� Bezos) and Oracle (Larry Ellison).

Another reason that Mr Andreessenhas become something of an entrepre-neur-magnet is his extensive network ofcontacts in Silicon Valley. He sits on theboards of Hewlett-Packard, eBay andFacebook, among others. This gives himan ideal perch from which to spot trendsforming. �Marc has a view of the entiretech ecosystem that very few peoplehave,� says David Lieb, the boss of Bump,a wireless start-up in which Andreessen

Horowitz has invested.His fans claim that Mr Andreessen’s

ability to draw insightful conclusions fromthese trends helps Andreessen Horowitzstand out from the crowd. �In a worldwhere there is a lot of dreaming, hopingand guessing, Marc takes a really analyti-cal approach,� says Mr Summers, whosigned on as a part-time adviser to the �rmafter meeting with a number of otherventure out�ts. Mr Andreessen’s symbiot-ic relationship with Mr Horowitz, a highlyexperienced manager, is also said to becentral to the �rm’s success. Tim Howes, aco-founder of RockMelt, a browser com-pany in which Andreessen Horowitz hasinvested, jokes that the two men haveworked together for so long that they arelike an old married couple who comple-ment one another perfectly.

The Hollywood treatment

Their partnership has spawned a boldnew approach to �rm-building in SiliconValley. Most venture �rms employ a skele-ton sta� of in-house experts in areas suchas recruiting and marketing to help advisestart-ups. Andreessen Horowitz, whichhas a total sta� of 36, has taken a di�erentapproach. In addition to its six generalpartners, the �rm has hired a bevy ofexecutives who are specialists in partic-ular areas, including 11dedicated to re-cruitment. This set-up, says Mr Andrees-sen, is inspired by Creative Artists Agency,which used to be run by Mr Ovitz. It andother Hollywood talent-managementcompanies spend a great deal of timenurturing directors and �lm stars, andhelping them to �nd jobs. AndreessenHorowitz wants to do the same thing fortalented tech folk, whose career pathsmight one day involve a stint at one of the�rms it backs.

This in-house entourage also re�ectsMr Andreessen’s �rm belief that manystart-ups today are damaging their pros-pects by starving areas such as sales andmarketing of investment on the oftenmisguided assumption the internet willmagically guarantee them a sizeable mar-ket. Mr Andreessen says he really wants toback �full-spectrum �rms� that aim to beoutstanding in every operational area,rather than just a few. Andreessen Horo-witz’s team will provide advice and guid-ance on how best to achieve this.

Many of these �rms will be Americanones. Mr Andreessen won’t rule out in-vesting in other countries�Skype, forexample, started in Estonia and is nowbased in Luxembourg�but says his �rm

has a preference for America because hebelieves it remains the best place in theworld to build companies. Still, manyinternet start-ups need to think globalearly on these days, which is one reasonwhy Andreessen Horowitz has engagedMr Summers to give it advice on every-thing from pricing strategies to geopolitics.

Yet Mr Andreessen is especially bullishabout Silicon Valley, where the process ofknowledge-sharing that drives innovationhas been greatly accelerated by the in-ternet and the rise of social networking. �Itnow feels like we were operating in theStone Age when I �rst came out here,� hesays. Another notable change is a demo-cratisation of entrepreneurialism in theValley. Entrepreneurs no longer simplyfollow a well-worn path to venture funds’doors from a handful of giant technologycompanies such as HP and Intel; todaythey come from a much wider range ofbackgrounds. In addition, says Mr An-dreessen, there has recently been �a mas-sive brain drain from Boston to the Valley,which has all but gutted Boston as a placefor high-tech entrepreneurship�.

This narrow geographic focus meansthat Andreessen Horowitz could be indanger of missing out on the fat pro�ts tobe made backing entrepreneurial out�tsfounded in some of the world’s largestand fastest-growing markets. But MrAndreessen likes to point out that it is noaccident that Silicon Valley has produceda string of success stories from Netscape toeBay, Google, Facebook and Twitter.China, India and other markets may beexciting places, but a large proportion ofthe software that is �eating the world� stillseems to come from California. 7

�Just when you think computer science isstabilising, everything changes.�

O�er to readersReprints of Technology Quarterly are availablefrom the Rights and Syndication Department.A minimum order of �ve copies is required.

Corporate o�erCustomisation options on corporate orders of100 or more are available. Please contact us todiscuss your requirements.For more information on how to order specialreports, reprints or any queries you may haveplease contact:

The Rights and Syndication DepartmentThe Economist26 Red Lion SquareLondon WC1R 4HQ

Tel +44 (0)20 7576 8148Fax +44 (0)20 7576 8492e-mail: rights@economist.comwww.economist.com.rights

2