Opinion: Robots will soon rule the world!

Featured Comment

Neil W Parker -

I've encountered people who openly admitted that their job was to automate and reduce human employment in specific industry segments.

"University of Oxford researchers Carl Benedikt Frey and Michael Osborne estimated in 2013 that 47% of total US jobs could be automated and taken over by computers by 2033. Even jobs in medicine, law, and education are not immune. Remarkably though, even as employment is shrinking, productivity is growing, thanks to the technological advancements."

Related Opinions

Recent Comments

Opinion Leaders

Recently Joined


The pro­duc­tiv­ity para­dox strikes

As the No­bel lau­re­ate econ­o­mist Robert Solow noted in 1987, com­put­ers are ‘‘ev­ery­where but in the pro­duc­tiv­ity sta­tis­tics’’. Since then, the so-called pro­duc­tiv­ity para­dox has become ever more strik­ing. Au­toma­tion has elim­i­nated many jobs. Ro­bots and ar­ti­fi­cial in­tel­li­gence now seem to prom­ise (or threaten) yet more rad­i­cal change. Yet pro­duc­tiv­ity growth has slowed across the ad­vanced economies; in Bri­tain, labour is no more pro­duc­tive to­day than it was in 2007. Some econ­o­mists see low busi­ness in­vest­ment, poor skills, out­dated in­fra­struc­ture, or ex­ces­sive reg­u­la­tion hold­ing back po­ten­tial growth. Oth­ers note wide dis­par­i­ties in pro­duc­tiv­ity be­tween lead­ers and laggards among in­dus­trial man­u­fac­tur­ers. Still oth­ers ques­tion whether in­for­ma­tion tech­nol­ogy is re­ally so dis­tinc­tively pow­er­ful. But the ex­pla­na­tion may lie deeper still. As we get richer, mea­sured pro­duc­tiv­ity may in­evitably slow, and mea­sured GDP per capita may tell us ever less about trends in hu­man wel­fare. The growth of dif­fi­cult-toau­to­mate ser­vice ac­tiv­i­ties may ex­plain some of the pro­duc­tiv­ity slow­down. In the United States, the Bureau of Labour Sta­tis­tics re­ports that eight of the 10 fastest­grow­ing job cat­e­gories are lowwage ser­vices such as per­sonal care and home health aides. The growth of ‘‘zero-sum’’ ac­tiv­i­ties may, how­ever, be even more im­por­tant. Look around the econ­omy, and it’s strik­ing how much high-tal­ent man­power is de­voted to ac­tiv­i­ties that can­not pos­si­bly in­crease hu­man wel­fare, but en­tail com­pe­ti­tion for the avail­able eco­nomic pie. Such ac­tiv­i­ties have become ubiq­ui­tous: le­gal ser­vices, polic­ing, and pris­ons; cy­ber­crime and the army of ex­perts de­fend­ing or­gan­i­sa­tions against it; fi­nan­cial reg­u­la­tors try­ing to stop mis­selling and the grow­ing ranks of com­pli­ance of­fi­cers em­ployed in re­sponse; the huge re­sources de­voted to US elec­tion cam­paigns; real-es­tate ser­vices that fa­cil­i­tate the ex­change of al­ready-ex­ist­ing as­sets; and much fi­nan­cial trad­ing. Much de­sign, brand­ing, and ad­ver­tis­ing ac­tiv­ity is also es­sen­tially zero-sum. It is cer­tainly good that new fash­ions can con­tin­u­ally com­pete for our at­ten­tion. Choice and hu­man cre­ativ­ity are valu­able per se. But we have no rea­son to be­lieve that 2050’s de­signs and brands will make us any hap­pier than those of 2017. In­for­ma­tion tech­nol­ogy may im­prove hu­man wel­fare in ways not cap­tured in mea­sured out­put. Bil­lions of hours of con­sumer time pre­vi­ously spent fill­ing in forms, mak­ing tele­phone calls, and queu­ing are elim­i­nated by in­ter­net-based shop­ping and search ser­vices. Valu­able in­for­ma­tion and en­ter­tain­ment ser­vices are pro­vided for free. Much that de­liv­ers hu­man wel­fare ben­e­fits is not re­flected in GDP. In­deed, mea­sured GDP and gains in hu­man wel­fare even­tu­ally may become en­tirely di­vorced. Imag­ine in 2100 a world in which so­lar-pow­ered ro­bots, man­u­fac­tured by ro­bots and con­trolled by ar­ti­fi­cial in­tel­li­gence sys­tems, de­liver most of the goods and ser­vices that sup­port hu­man wel­fare. All that ac­tiv­ity would ac­count for a triv­ial pro­por­tion of mea­sured GDP, sim­ply be­cause it would be so cheap. Con­versely, al­most all mea­sured GDP would re­flect ze­ro­sum and/or im­pos­si­ble-toau­to­mate ac­tiv­i­ties – hous­ing rents, sports prizes, artis­tic per­for­mance fees, brand roy­al­ties, and ad­min­is­tra­tive, le­gal, and po­lit­i­cal sys­tem costs. Mea­sured pro­duc­tiv­ity growth would be close to nil, but also ir­rel­e­vant to im­prove­ment in hu­man wel­fare. We are far from there yet. But the trend in that di­rec­tion may well help ex­plain the re­cent pro­duc­tiv­ity slow­down. The com­put­ers are not in the pro­duc­tiv­ity sta­tis­tics pre­cisely be­cause they are so pow­er­ful. Adair Turner is chair­man of the In­sti­tute for New Eco­nomic Think­ing.

Al­phaGo de­feats Lee again

Google’s ar­ti­fi­cial in­tel­li­gence (AI) pro­gram Al­phaGo de­feated world go cham­pion Lee Se-dol again in their se­cond game, Thurs­day, on the back of su­pe­rior cal­cu­la­tions that al­most seemed to ex­ceed hu­man brains. The game, which seemed to be a close one in the middle, dras­ti­cally in­clined to the AI’s fa­vor to­wards the end af­ter its at­tack on Lee’s ter­ri­tory in the cen­ter of the board. Lee, who chose rel­a­tively safer moves than in the pre­vi­ous round, tried to over­turn the game even af­ter he en­tered over­time count­ing but ended up re­sign­ing again af­ter 211 moves. Af­ter the match, Lee calmly ad­mit­ted a per­fect de­feat. “I was al­ready sur­prised enough yes­ter­day and I think I have noth­ing to say now,” Lee said. “It was re­ally a per­fect de­feat. I have never thought I was in the lead for a mo­ment.” He also said Al­phaGo played a flaw­less game. “I could not find any­thing par­tic­u­larly strange (from Al­phaGo). Yes­ter­day I thought there might be some­thing strange but to­day Al­phaGo played a per­fect game,” Lee said. De­spite the two con­sec­u­tive losses, Lee pledged to con­tinue to do his ut­most in the re­main­ing games. “Now that the score is 2-0 and (the vic­tory) is not ex­pected to be easy,” he said. “But I will do my best to win at least one round. “Re­flect­ing on to­day’s match, it is dif­fi­cult to over­turn the game af­ter the middle of it. I will need to try to con­clude the game be­fore then to have a bet­ter chance of win­ning.” A match com­men­ta­tor Michael Red­mond said, “Un­like in Oc­to­ber, Al­phaGo was very im­pres­sive, played in­no­va­tively and ad­ven­tur­ously and led dan­ger­ously-look­ing moves to suc­cess.” Chang­ing f rom the game, Lee played se­cond with white stones, re­ceiv­ing 7.5 com­pen­sa­tion points as Al­phaGo did in the pre­vi­ous game. Play­ing first with black stones, Al­phaGo be­gan the game at the 3-4 point of top left cor­ner on the board in­stead of the flower point where it started in the pre­vi­ous round and in all five wins against Euro­pean go cham­pion Fan Hui in Oc­to­ber last year. In the 13th move, the AI abruptly moved from the bot­tom right cor­ner and opened a new l ay­out in the up­per edge, which the com­men­ta­tor said was “sur­pris­ing and un­prece­dented.” From this mo­ment, Al­phaGo con­tin­ued with ir­reg­u­lar and un­pre­dictable po­si­tions while Lee pro­ceeded with rel­a­tively sta­ble moves to coun­ter­act the AI. In the mean­time, Lee re­frained from his unique ag­gres­sive and cre­ative play­ing style and spent twice as much time as Al­phaGo around an hour into the game, think­ing deeply about his moves. Two hours into the game, the AI started com­bat moves in the bot­tom left cor­ner. But Lee broke through and suc­ceeded in se­cur­ing ter­ri­tory. But when Al­phaGo at­tacked Lee’s group in the cen­ter, Lee turned to the AI’s ter­ri­tory in the top right in ex­change, which the com­men­ta­tors said was a bad move for Lee. Leav­ing only one minute and 15 sec­onds, Lee made a key move to turn the ta­bles. But the ma­chine did not fal­ter and drove Lee to over­time count­ing. In this spe­cial com­pe­ti­tion be­tween the world’s top go player and the AI sys­tem de­vel­oped by Google’s Lon­don-based sub­sidiary Deep­Mind, each player was given two hours per match with three lots of 60 sec­onds over­time count­ing af­ter they have fin­ished the al­lowed time. Each 60-se­cond lot is re­freshed if it is not used up. Lee strived for break­through but chose to re­sign four hours and 25 min­utes into the game. On Wed­nes­day, Al­phaGo beat Lee by a res­ig­na­tion in 186 moves af­ter three-and-a-half hours. Af­ter Al­phaGo’s moves that seemed by com­men­ta­tors as “ob­vi­ous and crit­i­cal” mis­takes in the lat­ter half of that game, Lee tried to push fur­ther but failed to se­cure enough points to over­come the 7.5 com­pen­sa­tion points and re­signed. Lee’s de­feat in the first game stunned the world as many have pre­dicted his win un­der as­sump­tion that go still re­mains dif­fi­cult area for ma­chines to out­per­form hu­mans as the an­cient game re­quires both ac­cu­rate cal­cu­la­tions and in­tu­ition. Last month, Lee showed con­fi­dence and said he would clinch a vic­tory with a score of 5-0 or 4-1. But in a press con­fer­ence on Tues­day, one day be­fore the first game, he said Al­phaGo’s im­proved abil­ity to nar­row down op­tions for next moves could be more in­tim­i­dat­ing than he orig­i­nally ex­pected. The five-game match be­tween Lee and the AI sys­tem will con­tinue on Satur­day and Sun­day be­fore wrap­ping up with the fifth-round on March 15. The Google Deep­Mind Chal­lenge Match, which was or­ga­nized by Google and the Korea Baduk As­so­ci­a­tion, will con­tinue even af­ter three wins by ei­ther side as the event aims at pro­vid­ing the ma­chine with more data.

Drone do that! Ro­bots killing jobs in Amer­ica

Wash­ing­ton: There was a ro­bot in­vented in Amer­ica that catches thieves. In the US, it caught 10 thieves; in Aus­tralia, it caught 100. In China, it caught 1,000. In In­dia, some­one stole the ro­bot. You may want the joke to come true be­cause ro­bots are com­ing to steal your jobs, par­tic­u­larly if your work is mostly repet­i­tive, me­chan­i­cal, mo­torised; some­thing that is pro­gram­mable. It’s no se­cret that au­to­ma­tion is tak­ing over low-wage jobs, but as ro­bots and drones get in­creas­ingly so­phis­ti­cated, a White House eco­nomic re­port re­leased on Mon­day has put num­bers on a trend that should give pause to any­one who thinks low-wage man­u­fac­tur­ing is the panacea to eco­nomic sal­va­tion in In­dia, or any­where else for that mat­ter. In the US, there’s an 83% chance that au­to­ma­tion will take a job with an hourly wage below $20, a 31% chance au­to­ma­tion will take a job with an hourly wage be­tween $20 and $40, and just a 4% chance au­to­ma­tion will take a job with an hourly wage above $40, a re­port by the White House Coun­cil of Eco­nomic Ad­vis­ers (CEA) warned. The risk of hav­ing your job ef­fec­tively taken over by a ro­bot, CEA chair­man Ja­son Furman told re­porters on Mon­day, “varies enor­mously based on what your salary is.” In other words, the more skilled, cre­ative, and high-earn­ing you are, less likely your job will be taken over by au­toma­tons. Me­chan­i­cal and as­sem­bly­line tasks, from lay­ing bricks to mak­ing cars and even driv­ing them, are al­ready be­ing taken over by drones, ro­bots, and other “in­tel­li­gent” forms even as sci­en­tists are try­ing to in­fuse them with in­tu­ition and emo­tions, their mem­ory and com­put­ing power long hav­ing sur­passed that of hu­man be­ings. Furman said the risk of many cur­rent jobs be­ing per­formed by ro­bots is an­other ex­am­ple of why it is im­por­tant to in­vest in education that helps peo­ple have skills that com­ple­ments au­to­ma­tion. De­vel­op­ments in US in­dust- ry serve as a warn­ing against over-re­liance on low-grade, as­sem­bly-line man­u­fac­tur­ing that can be re­placed by ro­bots. Even some of the low-level ser­vice jobs, such as dis­pens­ing food or gas or money, have been de­stroyed in the US with the ad­vance­ment of au­to­ma­tion — from ATMs to self-serve kiosks for pho­to­graphs. Man­u­fac­tur­ing jobs have de­clined by more than 7.2 mil­lion, or 37%, since em­ploy­ment in man­u­fac­tur­ing peaked in 1979. In 1965, man­u­fac­tur­ing ac­counted for 53% of the US econ­omy; by 1988 it only ac­counted for 39%, in 2014, it ac­counted for less than 9%. Not all of the de­struc­tion is ac­counted for by flight of jobs to China or Mex­ico. Lurk­ing in the back­ground, tech­no­log­i­cal ad­vance­ment. Univer­sity of Ox­ford re­searchers Carl Benedikt Frey and Michael Os­borne es­ti­mated in 2013 that 47% of to­tal US jobs could be au­to­mated and taken over by com­put­ers by 2033. Even jobs in medicine, law, and education are not im­mune. Re­mark­ably though, even as em­ploy­ment is shrink­ing, pro­duc­tiv­ity is grow­ing, thanks to the tech­no­log­i­cal ad­vance­ments. So, while the GDP is soar­ing, me­dian in­come in many jobs is fall­ing even as the job mar­ket it­self is be­com­ing smaller.

Nasa de­vis­ing hu­manoid ro­bots for fu­ture space mis­sion

NASA is de­vel­op­ing a sixfeet tall hu­manoid ro­bot that could as­sist as­tro­nauts in risky and ex­tremely haz­ardous deep space mis­sions to Mars and as­ter­oids in the fu­ture. The US space agency is con­sid­er­ing ush­er­ing new hu­manoid ro­bots that could of­fer as­tro­nauts a help­ing hand in fu­ture ex­pe­di­tions. “Nasa is count­ing on ro­bots to setup and care for deep space ex­plo­ration fa­cil­i­ties and equip­ment pre-de­ployed ahead of as­tro­nauts,” Sasha Congiu El­lis of Nasa’s Lan­g­ley Re­search Cen­tre, told Astrowatch.Net. “Ro­bots are also ex­cel­lent pre­cur­sors for con­duct­ing sci­ence mis­sions ahead of hu­man ex­plo­ration,” El­lis said. That is why the agency is de­vel­op­ing a six-feet tall hu­manoid ro­bot called R5, pre­vi­ously known as Valkyrie. The ma­chine weighs about 131 kg. It was ini­tially de­signed to com­plete disas­ter-re­lief ma­noeu­vres. In Novem­ber last year, Nasa awarded two R5 ro­bots to univer­sity groups — the Mas­sachusetts In­sti­tute of Tech­nol­ogy (MIT) and the North­east­ern Univer­sity in Bos­ton. Ac­cord­ing to Nasa, the teams have two years to per­form re­search and soft­ware de­vel­op­ment in or­der to im­prove the ro­bot’s au­ton­omy. They also have ac­cess to on­site and vir­tual tech­ni­cal sup­port from the agency.

Rise of ro­bots

In a glitzy ex­hi­bi­tion hall in Bei­jing, an ex­tra­or­di­nary Tai chi show is on. The per­form­ers are a mar­tial arts mas­ter in spot­less white robes and — wait for it — an in­dus­trial ro­bot. It’s the kind of ro­bot you would of­ten see hid­ing be­hind safety bar­ri­ers at fac­to­ries. No sooner had the mas­ter pushed his right hand against the ro­bot than the lat­ter sprung into ac­tion, cir­cled around and smoothly pushed him back in one fluid mo­tion. The 90- se­cond demo is more bal­let than com­bat as the man and the ma­chine en­gage in al­most in­ti­mate mo­tions. Ev­ery time the mas­ter comes into con­tact, the sen­sor- rich ma­chine, well, senses the touch in­stantly, de­ter­mines the amount of pres­sure in it and ‘ in­stinc­tively’ moves in the in­tended di­rec­tion. “This is China’s first home­grown seven- axis col­lab­o­ra­tive ro­bot,” said Qu Daokui, pres­i­dent of Siasun Ro­bot & Au­to­ma­tion Co, the man­u­fac­turer of the motorized arm. It can be trained by hands to per­form a string of in­dus­trial tasks like grind­ing, pack­ag­ing and feed­ing parts. “It is ready to work side by side with hu­mans on as­sem­bly lines, per­form­ing a string of in­dus­trial tasks like grind­ing, pack­ag­ing, as­sem­bling and feed­ing parts,” he said. The col­lab­o­ra­tive ro­bot, which was un­veiled in Novem­ber, is part of a broad ef­fort by Siasun, as well as its do­mes­tic peers, to cash in on the coun­try’s grow­ing ap­petite for in­dus­trial ro­bots as en­ter­prises are crank­ing up au­to­ma­tion of car and elec­tron­ics fac­to­ries. The world’s se­cond- largest econ­omy is al­ready the lead­ing mar­ket for in­dus­trial ro­bots, ac­count­ing for a quar­ter of global sales, ac­cord­ing to the In­ter­na­tional Fed­er­a­tion of Ro­bot­ics. Be­tween 2010 and 2014, to­tal sup­ply of in­dus­trial ro­bots in China in­creased by about 40 per­cent per year on av­er­age. “This ( China’s) rapid de­vel­op­ment is unique in the his­tory of ro­bot­ics. There has never been such dy­namic rise in such a short pe­riod of time in any other mar­ket,” IFR said in a re­port. But still, for ev­ery 10,000 em­ploy­ees, there are only 36 ro­bots in China, com­pared with 478 in South Korea, 292 in Ger­many and 164 in the United States in 2014. As the surg­ing la­bor cost is push­ing more en­ter­prises to em­brace ro­bots, IFR es­ti­mates China will ac­count for more than one- third of the in­dus­trial ro­bots in­stalled world­wide in 2018, more than dou­bling over the next two years — from 262, 900 now to 614,200. Nat­u­rally, the growth prospect of­fers a golden op­por­tu­nity for do­mes­tic ro­bot man­u­fac­tur­ers. In ad­di­tion to Siasun, the coun­try’s largest ro­bot maker by mar­ket value, a bunch of new play­ers are get­ting lured in and many ro­bot mak­ers are up­ping the ante with am­bi­tious de­vel­op­ment plans. “The past two years have seen an ex­plo­sion of do­mes­tic ro­bot mak­ers, partly stim­u­lat- ed by strong pol­icy sup­port,” said Luo Jun, chief ex­ec­u­tive of­fi­cer of the In­ter­na­tional Ro­bot­ics and In­tel­li­gent Equip­ment In­dus­try Al­liance, a Bei­jing- based in­dus­try as­so­ci­a­tion. The ro­bot in­dus­try is cen­tral to the coun­try’s Made in China 2025 strat­egy de­signed to pro­mote high- end man­u­fac­tur­ing. It is also high­lighted in Bei­jing’s new five- year plan, which guides na­tional eco­nomic de­vel­op­ment. One of the new en­trants is HIT Ro­bot Group, which was es­tab­lished in De­cem­ber 2014 with fund­ing from a pro­vin­cial govern­ment and the Harbin In­sti­tute of Tech­nol­ogy, a Chi­nese univer­sity which was ranked 7 th among the best global en­gi­neer­ing univer­si­ties by US News in 2015. The elite univer­sity has done years of cut­ting- edge re­search into ro­bots — it is the maker of China’s first space ro­bots and lu­nar ve­hi­cle. For its part, the com­pany is po­si­tioned to have a pres­ence in a wide range of ar­eas, from ro­bot com­po­nents, in­dus­trial ro­bots, ser­vice ro­bots to auto- mated equip­ment for nu­clear power plants and aero­space in­dus­tries. “We have ar­guably the best ro­bot en­gi­neers and re­searchers in China, which can give us an un­par­al­leled ad­van­tage,” wrote Han Jiecai, hon­orary chair­man of HIT Ro­bot and vice pres­i­dent of the Harbin In­sti­tute of Tech­nol­ogy, in a by­lined ar­ti­cle in the Eco­nomic Ob­server. But de­spite the ef­forts by lead­ing Chi­nese ro­bot man­u­fac­tur­ers to build a pre­mium brand, most play­ers are still locked in low- end com­peti- tion, ex­perts said. “Though the do­mes­tic sci­en­tific com­mu­nity has been re­search­ing ro­bots for many years, the ro­bot in­dus­try is still in its in­fancy,” said Hao Yucheng, deputy di­rec­tor of the China Ro­bot In­dus­try Al­liance. “The bulk of en­ter­prises have no in­tel­lec­tual prop­er­ties, tal­ents, and cash. They are just en­ter­ing the sec­tor with en­thu­si­asm and do­ing repet­i­tive jobs like as­sem­bling ro­bots in­stead of mak­ing ro­bots,” he said. In 2014, China bought over 57,000 in­dus­trial ro­bots, but less than 30 per­cent of them are from do­mes­tic sup­pli­ers, data from the China Ro­bot In­dus­try Al­liance shows. For­eign heavy­weights in­clud­ing ABB Ltd, KUKA Ro­bot­ics Corp and FANUC Corp are shar­ing the rest 70 per­cent of the mar­ket. “China’s ro­bot in­dus­try is like a tod­dler. But it is grow­ing in the world’s largest ro­bot mar­ket where many com­pet­i­tive for­eign en­ter­prises are scram­bling for a pie. Op­por­tu­ni­ties abound, so do chal­lenges. A sim­ple mis­take is likely to nip the in­dus­try in the bud,” Qu of Siasun said. Also, a wide tech­no­log­i­cal gap still ex­ists be­tween do­mes­tic ro­bot man­u­fac­tur­ers and their for­eign coun­ter­parts. China now has few en­ter­prises that can pro­vide masspro­duced and re­li­able in­dus­trial ro­bot com­po­nents such as speed re­duc­ers, drive and con­trol devices, as well as ser­vo­mo­tors. “Most of th­ese com­po­nents are still im­ported from for­eign coun­tries, whose steep tar­iffs in­crease the cost of ro­bots,” Hao of the China Ro­bot In­dus­try Al­liance said. For in­stance, speed re­duc­ers could ac­count for about 30 per­cent of do­mes­tic ro­bots’ cost, com­pared with only 12 per­cent for sim­i­lar Ja­panese ro­bots, an in­dus­try re­source said. But lo­cal en­ter­prises are al­ready ac­cel­er­at­ing steps to boost their ca­pa­bil­ity in re­search and de­vel­op­ment as well as small- scale pro­duc­tion of key ro­bot com­po­nents. Shaanxi Qinchuan Ma­chin­ery De­vel­op­ment Co Ltd, for in­stance, has poured 194 mil­lion yuan in 2013 into a ro­bot speed re­ducer pro­ject, which can now pro­duce 500 to 700 units per month. “We are work­ing on a mass pro­duc­tion as­sem­bly line pro­ject. When com­pleted, it will boost monthly pro­duc­tion ca­pac­ity to 5,000 units by the end of 2016,” the com­pany said in a press re­lease in Novem­ber. Siasun also ex­panded its pres­ence in the niche by set­ting up a unit in May 2015. It is plan­ning to ac­quire com­pet­i­tive do­mes­tic and in­ter­na­tional com­po­nent man­u­fac­tur­ers af­ter the Shen­zhen- listed com­pany raised 2.96 bil­lion yuan from five in­sti­tu­tional in­vestors in Novem­ber. “We have been in ne­go­ti­a­tions with po­ten­tial com­pa­nies for over a year and we hope to com­plete the ac­qui­si­tion by June,” Qu, pres­i­dent of Siasun said, de­clin­ing to of­fer more de­tails. “Our goal is not to catch up but to take the lead by in­no­vat­ing on the ba­sis of state- of- the- art tech­nolo­gies,” he said.

Why good jobs are too few, wages are poor

Amer­i­cans are jus­ti­fied to be an­gry about the eco­nomic re­cov­ery. As Pres­i­dent Barack Obama en­ters his fi­nal year, good-pay­ing jobs re­main scarce and fam­ily in­comes are down about $1,650 on his watch. Since Ron­ald Rea­gan ran the coun­try, the avail­abil­ity of at­trac­tive em­ploy­ment has been trend­ing down and slow­ing eco­nomic growth is of­ten blamed — dur­ing Obama’s re­cov­ery, gross do­mes­tic prod­uct has ad­vanced at a 2.2 per­cent an­nual pace, whereas the com­pa­ra­ble fig­ures for Rea­gan and Clin­ton were 4.6 and 3.7 per­cent. But that puts the story back­ward — the lack of work­ers ad­e­quately trained for a more tech­no­log­i­cal de­mand­ing work­place is slow­ing growth, not the other way around. Au­to­ma­tion has been an en­dur­ing theme through­out Amer­i­can his­tory. First, reapers and trac­tors con­sol­i­dated farms and sent work­ers to fac­to­ries. Then ma­chines re­placed work­ers in man­u­fac­tur­ing, push­ing them into more highly paid pro­fes­sions in medicine, education and tech­nol­ogy but also less well-paid oc­cu­pa­tions in restau­rants, re­tail­ing and other ser­vices. Un­til re­cently, com­puter-pro­grammed ma­chines could be taught stren­u­ous and repet­i­tive tasks like at­tach­ing a heavy, rigid fender onto an au­to­mo­bile. Go­ing for­ward ro­bots will in­creas­ingly re­place peo­ple in ac­tivi- ties re­quir­ing more sub­tle man­ual dex­ter­ity — like mak­ing shirts and har­vest­ing fruit — and those re­quir­ing more com­plex cog­ni­tive pro­cesses like ma­sonry con­struc­tion, driv­ing lim­ou­sines and build­ing new ro­bots that adapt to chang­ing en­vi­ron­men­tal con­di­tions. The drug­store I visit in Wash­ing­ton no longer has cashiers — just a group of check­out ma­chines and one clerk to as­sist tech­no­log­i­cally flum­moxed pa­trons. Over the next two decades, ro­bots will be ca­pa­ble of un­load­ing pal­lets, stock­ing shelves, fill­ing pre­scrip­tions, and gen­er­ally run­ning the store with min­i­mal hu­man in­ter­ven­tion. By 2030, it will be­come tech­no­log­i­cally pos­si­ble to re­place 90 per­cent of the jobs as we know them by smart ma­chines. The real chal­lenge will be train­ing most Amer­i­cans to en­gage in in­tel­lectu- ally de­mand­ing and cre­ative work, or the glob­al­iza­tion of tech­nol­ogy and com­pe­ti­tion will rel­e­gate most of us to very low pay­ing work bet­ter left to an­droids. In 2016, Amer­i­cans should be skep­ti­cal, not merely of false prom­ises to re­store pros­per­ity made by Bernie San­ders and Don­ald Trump but also out­raged by the hand­i­work of main­stream politi­cians. The lat­ter’s ef­forts to make a high school diploma uni­ver­sal have made it a nearly worth­less cre­den­tial. Less than 40 per­cent of 12th-graders are ready to read or learn math at the col­lege level, and many fewer have skills to en­ter tech­ni­cally de­mand­ing po­si­tions with­out post-sec­ondary train­ing. A col­lege diploma is not much bet­ter. Af­ter push­ing mil­lions of un­qual­i­fied stu­dents into univer­si­ties through af­fir­ma­tive ac­tion and govern­ment loan pro­grams, 4 in 10 grad­u­ates lack the com­plex rea­son­ing skills needed for white-col­lar work — as it ex­ists to­day, never mind as it will be af­ter ma­chines equipped with high-level ar­ti­fi­cial in­tel­li­gence can re­place armies of stock­bro­kers, in­sur­ance ad­justers and restau­rant man­agers over the next sev­eral decades. Mean­while, the pres­i­dent and his pre­sump­tive heir, Hil­lary Clin­ton, re­main ob­sessed with sex­ism in education and the work­place. That nearly 60 per­cent of col­lege de­grees are now awarded to women and fe­males of­ten earn more than males in com­pa­ra­ble po­si­tions are in­con­ve­nient facts when there are vot­ers to be mis­led to ex­tend a political dy­nasty. And con­ser­va­tives — in­clud­ing the likes of Ted Cruz and Marco Ru­bio — op­pose uni­ver­sal stan­dards for more aca­demic rigor like the Com­mon Core. The fu­ture lies in ed­u­cat­ing Amer­i­cans, not to be an­gry about false in­jus­tice or an om­nipresent state but, rather, to build and teach the ma­chines that will do the work that has bur­dened hu­man­ity since the first branch was shaped into a hunt­ing im­ple­ment. With­out young peo­ple trained and en­cour­aged to do that so­phis­ti­cated work, the lo­cus of pros­per­ity will per­ma­nently shift from Amer­ica to Asia, where prag­matic lead­ers urge chil­dren to study en­gi­neer­ing, not the su­per­sti­tions ped­dled by pi­ous academics and de­ceit­ful politi­cians.

China dreams of elec­tric sheep at robot con­fer­ence

— In a mar­tial artist’s white silk py­ja­mas, a man prac­ticed tai-chi in har­mony with a mo­tor­ized arm at a Beijing ex­hi­bi­tion show­cas­ing a vi­sion of ro­bots with Chi­nese char­ac­ter­is­tics. Ve­hi­cles with au­to­mated gun tur­rets sat along­side drink-serv­ing karaoke ma­chines at the World Robot Con­fer­ence, as man­u­fac­tur­ers sought new buy­ers for their “jiqiren” — “ma­chine peo­ple” in Chi­nese. The push has sup­port at the high­est lev­els of gov­ern­ment. Pres­i­dent Xi Jin­ping is­sued a let­ter of con­grat­u­la­tions for the con­fer­ence, and the in­dus­try is name-checked in the draft version of the coun­try’s new five-year plan, the pol­icy doc­u­ment that guides na­tional eco­nomic de­vel­op­ment. The world’s sec­ond-largest econ­omy is al­ready the lead­ing mar­ket for in­dus­trial ro­bots, ac­count­ing for a quar­ter of global sales, ac­cord­ing to the In­ter­na­tional Fed­er­a­tion of Ro­bot­ics. But ex­ec­u­tives at a con­fer­ence round­table said the real mar­ket op­por­tu­nity was in ser­vice ro­bots for the homes and of­fices of the world’s most pop­u­lous coun­try. “There are now less than 100,000 ro­bots in Chi­nese fam­i­lies, not in­clud­ing vac­uum clean­ers,” said Liu Xue­nan, chief ex­ec­u­tive of­fi­cer of Can­bot. In the fu­ture, said Yu Kai, the head of Hori­zon Ro­bot­ics, China’s au­to­mated helpers will do ev­ery­thing from build­ing cars to driv­ing them, pre­dict­ing that “each per­son might have 10 ro­bots” — nearly 14 bil­lion po­ten­tial tin men at cur­rent pop­u­la­tion lev­els. Planet of the Apps Ro­bots have cap­tured China’s imag­i­na­tion. From Trans­form­ers to Bay­max, the star of Dis­ney’s movie “Big Hero 6,” Chi­nese con­sumers have em­braced robot he­roes, spend­ing hun­dreds of mil­lions on re­lated movies and mer­chan­dise. In Chi­nese cities, busi­nesses try to at­tract cus­tomers with robot wait­ers, cooks, and concierges. In the coun­try­side, ru­ral Da Vin­cis cob­ble to­gether me­chan­i­cal men from scrap­yard junk. A panel at the con­fer­ence strug­gled with the ques­tion of how China would deal with the rise of ar­ti­fi­cially in­tel­li­gent ma­chines. But the tran­si­tion from the world of fan­tasy and nov­elty to a real robot econ­omy could be tricky, with the coun­try’s tech­nol­ogy still lag­ging far be­hind neigh­bors Korea and Ja­pan, the undis­puted king of the ro­bots. China should have more re­al­is­tic expectations for the near fu­ture, said Pin­pin Zhu, pres­i­dent of China’s voice con­trolled ser­vice Xiao I Robot, which was in­volved in a patent dis­pute with Amer­i­can tech gi­ant Ap­ple linked to its per­sonal dig­i­tal as­sis­tant Siri. The coun­try may de­scend from the peak of high expectations into a “trough of dis­il­lu­sion­ment,” said Zhu, who be­lieves a smart­phone-based “Planet of the Apps” is more likely than a world served by hu­manoid ro­bots.

The dis­rup­tions re­shap­ing our world

The day af­ter a book called The Rise of the Ro­bots won the Fi­nan­cial Times McKin­sey Busi­ness Book of the Year award in Lon­don this week, a lit­tle com­pany called Fast­brick Ro­bot­ics listed on the ASX and promptly dou­bled in value. It was in­stant ful­fil­ment. Fast­brick Ro­bot­ics has in­vented a brick­lay­ing robot; the book, by Sil­i­con Val­ley en­tre­pre­neur Martin Ford, presents a grim view of a fu­ture in which ro­bots per­ma­nently re­place hu­man jobs. Re­plac­ing brick­ies with ro­bots is not that sur­pris­ing when you think about it — af­ter all, lay­ing bricks is a repet­i­tive, robot-like task. Talk to me when you’ve in- vented a robot builder, then I’ll be im­pressed (and grate­ful). Brick­ies might not agree, but the cor­po­rate videos of Fast­brick Ro­bot­ics present a bright new world of quick and ef­fi­cient con­struc­tion, in which a house can go up in two days. The Rise of the Ro­bots, on the other hand, fore­sees a dark dystopian fu­ture of fewer jobs, with dis­rup­tion from ma­chines and al­go­rithms on both man­u­fac­tur­ing and pro­fes­sional in­dus­tries. Ac­cord­ing to the re­views (I haven’t read the book yet), Ford es­sen­tially ar­gues that the cur­rent in­dus­trial revo­lu­tion will not be like the last one, when new jobs were cre­ated just as quickly as the old ones were elim­i­nated by tech­nol­ogy. He writes: “While hu­man­ma­chine col­lab­o­ra­tion jobs will cer­tainly ex­ist, they seem likely to be rel­a­tively few in num­ber and of­ten short-lived. In a great many cases, they may also be un­re­ward­ing and even de­hu­man­is­ing.” Ac­cord­ing to Martin Ford, ar­ti­fi­cial in­tel­li­gence is al­ready well on its way to making “good jobs” ob­so­lete: many par­ale­gals, jour­nal­ists, of­fice work­ers and even com­puter pro­gram­mers are poised to be re­placed by ro­bots and smart soft­ware. As progress con­tin­ues, blue and white col­lar jobs alike will evap­o­rate, squeez­ing work­ing and mid­dle class fam­i­lies ever fur­ther. But Ford does not only talk about job theft by ma­chines. One of his key points is that ro­bots weaken mid­dle-class de­mand by skew­ing the gains to a few — wors­en­ing in­equal­ity. As labour be­comes un­eco­nomic rel­a­tive to ma­chines, pur­chas­ing power falls. For ex­am­ple, the US econ­omy pro­duces a third more to­day than it did in 1998 with the same sized work­force. What he doesn’t men­tion is that macroe­co­nomic pol­icy is do­ing the same thing, to some ex­tent driven by dig­i­tal dis­rup­tion and ro­bot­ics. Just as fall­ing wage growth is re­duc­ing mid­dle-class de­mand, so are su­per low in­ter­est rates. It’s true that dis­pos­able in­comes of mid­dle-class mort­gagees are be­ing boosted by low in­ter­est rates, but that is be­ing can­celled by the op­po­site ef­fect on re­tirees who live off their sav­ings. To some ex­tent it’s a cir­cu­lar phe­nom­e­non: au­to­ma­tion and dis­rup­tion re­duce costs and prices, re­duc­ing in­fla­tion and there­fore in­ter­est rates. So mon­e­tary pol­icy it­self be­comes a dis­rupter. Not only is wages growth the low­est on record and unit labour costs not grown at all for three years, in­ter­est rates are the low­est they have ever been. In the US they have been ef­fec­tively zero for seven years. As I see it the world has to deal with six great dis­rup­tions at once: Mon­e­tary pol­icy Zero in­ter­est rates and quan­ti­tat- ive eas­ing are one of great dis­rup­tive in­no­va­tions of our time: neg­a­tive real in­ter­est rates — and in some places even neg­a­tive nom­i­nal in­ter­est rates — and cen­tral banks sim­ply print­ing money and buy­ing as­sets from banks. It’s an ex­per­i­ment that is hav­ing a pro­found ef­fect on the way all mar­kets and economies op­er­ate. Com­put­ing power When Gor­don Moore ob­served in 1965 that the num­ber of tran­sis­tors on an in­te­grated cir­cuit could dou­ble ev­ery year (and then re­vised that to ev­ery two years in 1975) it was early days for Moore’s Law. Forty years of that ex­po­nen­tial growth in com­put­ing power is now af­fect­ing ev­ery part of life. Moore’s Law is re­spon­si­ble for smart­phones, ro­bot­ics, ar­ti­fi­cial in­tel­li­gence, pro­gram­matic trad­ing in shares and ad­ver­tis­ing … the list goes on, and in­cludes all the things that Martin Ford is so wor­ried about. Cloud com­put­ing In a way this is an ex­ten­sion of Moore’s Law, but turn­ing both data stor­age and soft­ware into a ser­vice — op­er­at­ing ex­pen­di­ture rather than cap­i­tal ex­pen­di­ture — is it­self hugely dis­rup­tive and de­serves its own men­tion. Hy­per-Con­nec­tiv­ity In an in­ter­view this week with Stephen Bartholomeusz and my­self for Busi­ness Spec­ta­tor, the CEO of Tel­stra Andy Penn re­vealed that the com­pany had achieved a data speed over its mo­bile net­work of one gi­ga­bit per sec­ond in a test en­vi­ron­ment. Sim­i­lar down­load speeds are al­ready be­ing achieved over fixed line. This has meant the in­ter­net can be used for broad­cast­ing as well as con­nect­ing mil­lions of ma­chines (the “in­ter­net of things”) trans­mit­ting colos­sal amounts of data and stor­ing it in the cloud. Blockchain This is the tech­nol­ogy at the heart of Bit­coin, but its uses go far wider than that. Blockchain is a way of or­gan­is­ing and ver­i­fy­ing al­most any trans­ac­tion. It is early days, but al­ready clear that this tech­nol­ogy will even­tu­ally be at heart of a new bank­ing sys­tem and a new set­tle­ment sys­tem. Trust In my view this is the most pow­er­ful dis­rup­tive force of all. The in­ter­net has be­come a tool for hu­mans to deal di­rectly with each other. And it turns out that in even a world that is also be­ing dis­rupted by ap­palling acts of ter­ror­ism, there is an over­whelm­ing urge for peo­ple to trust each other. We go and stay in each other’s houses (Airbnb), get in each other’s cars (Uber), buy stuff from each other and pay first (eBay, etc). We tell each other about our lives and ideas (Face­book, Twit­ter, In­sta­gram, etc) and we’re now lend­ing each other money (So­ci­ety One, Lend­ing Club, etc). The will­ing­ness of hu­mans to trust each other and, separately, to ex­press them­selves to each other, is not a new thing, but the in­ter­net has un­leashed it and al­lowed it to be or­gan­ised. The so­cial urge in hu­man so­ci­ety has al­ways ex­isted, but has been con­fined to small com­mu­ni­ties sim­ply be­cause of the lack of any means to com­mu­ni­cate ef­fi­ciently across large dis­tances. The com­bi­na­tion of con­nec­tiv­ity and com­put­ing power has lit­er­ally put that abil­ity into our pock­ets — it’s with us every­where we go. It has meant that our in­her­ent urge to trust and com­mu­ni­cate can be both fully ex­pressed and or­gan­ised.

Robots can kill but they don’t un­der­stand us

“I’ve seen things you peo­ple wouldn’t be­lieve,” the vil­lain played by Rut­ger Hauer rem­i­nisces at the end of the film Blade Run­ner af­ter haul­ing Har­ri­son Ford’s char­ac­ter onto a rooftop and spar­ing his life. “Peo­ple” is the op­er­a­tive word as Roy Batty is not a per­son but an an­droid who es­capes to earth from a space colony and takes re­venge on the Tyrell Cor­po­ra­tion, his cre­ator. That is what I call a killer ro­bot – a be­ing that can hold an in­tel­li­gent con­ver­sa­tion with you be­fore wip­ing you out. It was science fic­tion in 1982, when Blade Run­ner, based on Philip K. Dick’s dystopian fan­tasy novel Do An­droids Dream Of Elec­tric Sheep?, came out. It is now faintly plau­si­ble – suf­fi­ciently for ar­ti­fi­cial in­tel­li­gence re­searchers to warn this week of the dan­gers of an au­ton­o­mous arms race. The killer ma­chines feared by those such as Mr Elon Musk, the founder of Tesla Mo­tors, and the­o­ret­i­cal physi­cist Stephen Hawk­ing are crude ter­mi­na­tors by com­par­i­son with the Nexus repli­cants in Blade Run­ner. No one would fall in love with an armed quad­copter that blows up en­emy sol­diers, as the hero of Blade Run­ner does with Rachael, the fe­male an­droid who does not re­alise that she is a repli­cant. Robots can mur­der us but they can­not un­der­stand us. Au­ton­o­mous killing ma­chines are be­com­ing re­al­ity – Is­rael al­ready has its Harpy anti-radar drone, which loi­ters in the sky be­fore choos­ing and de­stroy­ing tar­gets by it­self. A sen­tient, so­phis­ti­cated ma­chine with com­mon sense and the ca­pac­ity to grasp peo­ple’s moods and pre­dict be­hav­iour is still a dis­tant prospect. In the­ory, it will be cre­ated. Ar­ti­fi­cial in­tel­li­gence re­searchers do not see the bar­rier in prin­ci­ple to robots de­vel­op­ing higher rea­son­ing pow­ers, or the kind of phys­i­cal dex­ter­ity that hu­mans pos­sess. The last re­main­ing work­ers on car assem­bly lines are peo­ple who can at­tach screws nim­bly and reach in­side the body shells for elec­tri­cal wiring in a way that has de­feated robots to date. Ma­chines also pos­sess some ad­van­tages. They do not have to con­strict their pro­cess­ing units to fit into skulls, and they do not need to sup­ply them with oxy­gen, an energy-hog­ging tech­nol­ogy. Nor are they lim­ited by an evo­lu­tion­ary edict to re­pro­duce, rather than purely to get clev­erer. But de­spite rapid ad­vances in ma­chine learn­ing, vis­ual and voice recog­ni­tion, and neu­ral net­work pro­cess­ing – all the el­e­ments that are now trans­form­ing the po­ten­tial of ar­ti­fi­cial in­tel­li­gence – an­droids are not with us. Com­put­ers can beat hu­mans easily at chess, but poker at the high­est level is be­yond them – they would need to see through the other play­ers’ bluffs. “Com­put­ers are be­com­ing bet­ter and bet­ter at per­cep­tion tasks,” says Dr Fei-Fei Li, di­rec­tor of Stan­ford Univer­sity’s ar­ti­fi­cial in­tel­li­gence lab­o­ra­tory. “Al­go­rithms can iden­tify thou­sands of types of cars while I can tell only three of them. But at the cog­ni­tive, em­pa­thetic, and emo­tional level, ma­chines are not even close to hu­mans.” I have also ex­pe­ri­enced some­thing you peo­ple would not be­lieve – Google’s self-driv­ing car. The thing that struck me as it toured Moun­tain View in Cal­i­for­nia re­cently was that it felt hu­man. It ac­cel­er­ated from junc­tions con­fi­dently, even as­sertively, clos­ing the gaps with ve­hi­cles in front so oth­ers could not rush in. We would be safer if all driv­ers were equally calm and ra­tio­nal. In­side the car, you can see what it per­ceives with its sen­sors and rooftop radar. The out­lines of ob­jects around, in­clud­ing pedes­tri­ans, buses and other cars, are dis­played like hol­low, mov­ing shapes on the screen of a lap­top held by a Google engi­neer. The ob­jects are cat­e­gorised by dif­fer­ent colours, so the ve­hi­cle knows it should re­act to them and how far to steer clear. A self-driv­ing ve­hi­cle would, in other words, be a per­fectly ca­pa­ble killer ro­bot if you at­tach a mis­sile launcher to its roof, and ma­chine guns to its sides (not that Google would do such a thing, of course). It could cruise through cities, scan­ning for warm, slow-mov­ing, pink-coloured ob­jects to de­stroy. So it is not scare­mon­ger­ing for sci­en­tists to warn of ar­ti­fi­cial in­tel­li­gence re­search be­ing tainted by as­so­ci­a­tion with au­ton­o­mous weapons. The In­ter­net it­self emerged from re­search funded by the United States Depart­ment of De­fence in the 1960s, and mil­i­tary and space pro­grammes have the deep­est pock­ets and the keen­est in­ter­est in de­vel­op­ing cut­ting-edge tech­nol­ogy. What would be foolish would be to think the ad­vent of killer robots means that ma­chines are ready to take over the world. De­stroy­ing things is eas­ier than un­der­stand­ing or cre­at­ing them. Ar­ti­fi­cial in­tel­li­gence – the abil­ity to scan, process and an­a­lyse large data sets – is not the same as the ca­pac­ity to per­form most hu­man tasks (known as ar­ti­fi­cial gen­eral in­tel­li­gence). Even those who warn of ma­chines tak­ing jobs that are now per­formed by hu­mans ac­cept that man­age­rial, pro­fes­sional and artis­tic jobs that de­mand high-level rea­son­ing, em­pa­thy and cre­ativ­ity are still safe. A ro­bot that scans a set of fea­tures to iden­tify a woman, but can­not grasp her mood, or use com­mon sense to solve an un­ex­pected puz­zle, re­mains very lim­ited. “Quite an ex­pe­ri­ence to live in fear, isn’t it? That’s what it’s like to be a slave,” Roy Batty says to the hu­man bounty-hunter he has de­feated in com­bat be­fore reach­ing out and res­cu­ing him from fall­ing to his death. Let us not en­slave our­selves yet.

Is online pric­ing cus­tom-made?

The smart­phone era has in­tro­duced a new list of rules for shop­ping. First, find a prod­uct and check the In­ter­net to com­pare the store’s price against that of a nearby com­peti­tor or online re­tailer. If the cur­rent lo­ca­tion wins, make the pur­chase. If not, ei­ther take the trip across town or press the check­out but­ton on your vir­tual shop­ping cart. There’s lit­tle doubt that the in­stant ac­ces­si­bil­ity of the In­ter­net has helped con­sumers find the goods they need for the prices they’re will­ing to pay. How­ever, with In­ter­net searches adding to dig­i­tal pro­files de­signed to glean as much in­for­ma­tion about con­sumer lives and habits as pos­si­ble, is the best price one that has been pre­de­ter­mined by cor­po­ra­tions based on a shop­per’s history? Dif­fer­en­tial pric­ing — the eco­nomic term for set­ting dif­fer­ent prices for the same prod­uct for dif­fer­ent cus­tomers — has long been a fea­ture of com­merce in one form or another. Se­nior cit­i­zens’ and vet­er­ans’ dis­counts, air­line tick­ets and ne­go­ti­ated rates on car lots all fall un­der the um­brella. Dy­namic pric­ing, in which busi­nesses change prices based on al­go­rithms that take into ac­count com­peti­tor pric­ing, sup­ply and de­mand, and other ex­ter­nal fac­tors, has grown more so­phis­ti­cated over time. The catch is that in the past the prac­tice was geared to­ward spe­cific de­mo­graph­ics or in re­sponse to broad mar­ket con­di­tions. But dig­i­tal data that tracks nearly ev­ery online ac­tion could give com­pa­nies the op­por­tu­nity some­day to track the ex­act price an in­di­vid­ual con­sumer is will­ing to pay for a good or ser­vice — a form of per­son­al­ized pric­ing on steroids. “The more a mer­chant knows about you, the more they can pre­dict your max­i­mum will­ing­ness to pay for a good,” said Alessan­dro Ac­quisti, Carnegie Mel­lon Univer­sity pro­fes­sor of in­for­ma­tion tech­nol­ogy and co-di­rec­tor of the univer­sity’s Cen­ter for Be­hav­ioral Re­search. “All of the trails of data that you leave as you browse around the In­ter­net are be­ing stud­ied to cre­ate a pic­ture of you which can be used not only to show you a cer­tain, par­tic­u­lar ad­ver­tise­ment, but also to show you a cer­tain, par­tic­u­lar price.” Mr. Ac­quisti, a renowned pri­vacy re­searcher who will ex­plore the is­sue as part of a two-year fel­low­ship with the Carnegie Cor­po­ra­tion of New York this year, said he was first drawn to the idea by a pa­per he wrote with for­mer col­league Hal Var­ian in 2001 pre­dict­ing the phe­nom­e­non. Since it was pub­lished more than a decade ago, there has been ex­po­nen­tial growth in dy­namic pric­ing ser­vices such as Ama­zon Prime or loy­alty re­wards pro­grams that of­fer dis­counts to cer­tain cus­tomers. What hasn’t been seen as of­ten is tar­geted pric­ing based on pur­chas­ing history, even though there’s some ev­i­dence it has oc­curred. A 2012 Wall Street Jour­nal in­ves­ti­ga­tion re­vealed that Sta­ples, Home De­pot and sev­eral other re­tail­ers gave dif­fer­ent con­sumers dif­fer­ent prices for the same prod­ucts based on lo­ca­tion data from con­sumers’ cell phones that in­di­cated how close the con­sumers were to a com­peti­tor’s store. Dif­fer­en­tial pric­ing is not illegal un­less the rea­son for the dif­fer­ence is based on re­liance on a cat­e­gory such as race, re­li­gion, na­tional ori­gin or gen­der. The prac­tice could also be illegal if it vi­o­lates an­titrust or price-fix­ing laws. But con­sumers who are aware that hag­gling is the norm in bricks-and-mor­tar places such as car deal­er­ships may not be aware that the same thing goes on online. A car sales­per­son might fig­ure that a cus­tomer in a $3,000 suit driv­ing a 1-yearold Mercedes is a good bet for selling at a higher price, but an In­ter­net re­tailer can tell by that same cus­tomer’s online buy­ing habits with greater ac­cu­racy that he’s likely to pay more for an espresso ma­chine than the cus­tomer with a history of shop­ping around for the best price. And while the well­dressed cus­tomer can hag­gle as read­ily as his grungier coun­ter­part on the sales room floor, it’s harder to bar­gain with an Ama­zon shop­ping cart. Since con­sumers don’t regularly com­pare prices with con­sumers from across town or across the coun­try buy­ing the same goods, it’s dif­fi­cult to tell how of­ten the prac­tice ac­tu­ally oc­curs, said Ali Lange, pol­icy an­a­lyst for the Washington, D.C.-based non­profit Cen­ter for Democ­racy and Tech­nol­ogy. “It’s def­i­nitely a con­cern for con­sumers. The way com­pa­nies can as­sume things about peo­ple based on their data is some­thing I don’t think peo­ple fully grasp,” she said. In an at­tempt to get ahead of the is­sue, the White House in Fe­bru­ary is­sued the re­port “Big Data and Dif­fer­en­tial Pric­ing.” Its ul­ti­mate con­clu­sion was that per­son­al­ized pric­ing was prob­a­bly scarce be­cause com­pa­nies aren’t sure they can tar­get cus­tomers’ needs ac­cu­rately and be­cause of a fear of back­lash against the prac­tice. The re­port agreed with Mr. Ac­quisti in one re­gard: rais­ing aware­ness of the po­ten­tial for dan­ger in or­der to to pro­tect con­sumer rights. “Given the speed at which both the tech­nol­ogy and busi­ness prac­tices are evolv­ing, com­mer­cial ap­pli­ca­tions of big data de­serve on­go­ing scru­tiny, par­tic­u­larly where com­pa­nies may be us­ing sen­si­tive in­for­ma­tion in ways that are not trans­par­ent to users and fall out­side the bound­aries of ex­ist­ing reg­u­la­tory frame­works.”

Real­botix pro­ject to trans­form the world of sex dolls

THE WORLD of sex dolls is about to get even stranger. A pro­ject, dubbed Real­botix, is cre­at­ing in­tel­li­gent dolls that can com­mu­ni­cate re­al­is­ti­cally with their own­ers. The pro­ject’s cre­ator hopes that the sex dolls will even­tu­ally be able to think for them­selves, while sat­is­fy­ing the cus­tomer’s phys­i­cal needs. Real­botix is the in­ven­tion of Matt McMullen who is best known for cre­at­ing “Real Dolls”, which are life-like sil­i­cone dolls that have be­come pop­u­lar in the in­dus­try. McMullen boasts that he has sold more than 5 000 life-size dolls since 1996, with prices from $5 000 (R62 026) to $10 000. As part of his new pro­ject, McMullen is fo­cus­ing on de­vel­op­ing con­vinc­ing ar­ti­fi­cial in­tel­li­gence on a ro­botic head that can blink and open and close its mouth. Ac­cord­ing to a re­port in the New York Times, he hopes to in­te­grate the dolls with mo­bile apps so that they can act as vir­tual as­sis­tants. Vir­tual re­al­ity head­sets could also be used sep­a­rately or along- side the phys­i­cal dolls. The dolls will also build on the Real Doll cre­ation by adding what McMullen calls a “cus­tomis­able pro­gram­ming of per­son­al­ity”. “The hope is to cre­ate some­thing that will ac­tu­ally arouse some­one on an emo­tional, in­tel­lec­tual level, be­yond the phys­i­cal,” he said. In the video by the New York Times, one of the dolls de­scribes her­self as “a pro­to­type of a very ex­cit­ing new form of adult com­pan­ion­ship”. – Daily Mail

Ro­bots that flirt are on the way – but don’t panic, says ex­pert

Com­put­ers will have de­vel­oped “com­mon sense” within a decade and we could be count­ing them among our friends not long af­ter­wards, one of the world’s lead­ing ar­ti­fi­cial in­tel­li­gence sci­en­tists has pre­dicted. Prof Ge­off Hin­ton, hired by Google two years ago to help de­velop in­tel­li­gent op­er­at­ing sys­tems, said that the com­pany is on the brink of de­vel­op­ing al­go­rithms with the ca­pac­ity for logic, nat­u­ral con­ver­sa­tion and even flir­ta­tion. The re­searcher told the Guardian that Google is work­ing on a new type of al­go­rithm de­signed to en­code thoughts as se­quences of num­bers – some­thing he de­scribed as “thought vec­tors”. Although the work is at an early stage, he said there is a plau­si­ble path from the cur­rent soft­ware to a more so­phis­ti­cated ver­sion that would have some­thing ap­proach­ing hu­man-like ca­pac­ity for rea­son­ing and logic. “Ba­si­cally, they’ll have com­mon sense.” The idea that thoughts can be cap­tured and dis­tilled down to cold se­quences of dig­its is con­tro­ver­sial, Hin­ton said. “There’ll be a lot of peo­ple who ar­gue against it, who say you can’t cap­ture a thought like that,” he added. “But there’s no rea­son why not. I think you can cap­ture a thought by a vec­tor.” Hin­ton, who is due to give a talk at the Royal So­ci­ety in Lon­don to­day, be­lieves that the “thought vec­tor” ap­proach will help to crack two of the cen­tral chal­lenges in ar­ti­fi­cial in­tel­li­gence (AI): mas­ter­ing nat­u­ral, con­ver­sa­tional lan­guage, and the abil­ity to make leaps of logic. He painted a pic­ture of the near-fu­ture in which peo­ple will chat with their com­put­ers for fun – rem­i­nis­cent of the film, Her, in which Joaquin Phoenix falls in love with his in­tel­li­gent op­er­at­ing sys­tem. “It’s not that far-fetched,” Hin­ton said. “I don’t see why it shouldn’t be like a friend. I don’t see why you shouldn’t grow quite at­tached to them.” Richard Socher, an AI sci­en­tist at Stan­ford Uni­ver­sity, re­cently de­vel­oped a pro­gram called NaSent that he taught to recog­nise hu­man sen­ti­ment by train­ing it on 12,000 sen­tences taken from the film re­view web­site Rot­ten Toma­toes. Part of the ini­tial mo­ti­va­tion for de­vel­op­ing “thought vec­tors” was to im­prove trans­la­tion soft­ware, such as Google Trans­late, which cur­rently uses dic­tio­nar­ies to trans­late in­di­vid­ual words and searches through pre­vi­ously trans­lated doc­u­ments to find typ­i­cal trans­la­tions for phrases. Although th­ese meth­ods of­ten pro­vide the rough mean­ing, they are also prone to de­liv­er­ing non­sense and du­bi­ous gram­mar. Thought vec­tors, Hin­ton ex­plained, work at a higher level by ex­tract­ing some­thing closer to ac­tual mean­ing. The tech­nique works by as­crib­ing each word a set of num­bers (or vec­tor) that de­fine its po­si­tion in a the­o­ret­i­cal “mean­ing space” or cloud. A sen­tence can be looked at as a path be­tween th­ese words, which can in turn be dis­tilled down to its own set of num­bers, or thought vec­tor. The “thought” serves as the bridge be­tween the two lan­guages be­cause it can be trans­ferred into, say, the French ver­sion of mean­ing space and de­coded back into a new path be­tween words. The key is work­ing out which num­bers to as­sign each word in a lan­guage – this is where deep learn­ing comes in. Ini­tially the po­si­tions of words are or­dered at ran­dom and the trans­la­tion al­go­rithm be­gins train­ing on a dataset of trans­lated sen­tences. At first the trans­la­tions it pro­duces are non­sense, but a feed­back loop pro­vides an er­ror sig­nal that al­lows the po­si­tion of each word to be re­fined un­til even­tu­ally the po­si­tion of words in the cloud cap­tures the way hu­mans use them – in ef­fect a map of their mean­ings. Hin­ton said the idea lan­guage can be de­con­structed with al­most math­e­mat­i­cal pre­ci­sion is sur­pris­ing, but true. “If you take the vec­tor for Paris and sub­tract the vec­tor for France and add Italy, you get Rome. It’s quite re­mark­able.” Dr Her­mann Hauser, a Cam­bridge com­puter sci­en­tist and en­tre­pre­neur, said that Hin­ton and oth­ers could be on the way to solv­ing what pro­gram­mers call the “ge­nie prob­lem”. “With ma­chines at the mo­ment, you get ex­actly what you wished for,” Hauser said. “The prob­lem is we’re not very good at wish­ing for the right thing. “Hin­ton is our num­ber one guru in the world on this at the mo­ment,” he added. Some as­pects of com­mu­ni­ca­tion are likely to prove more chal­leng­ing, Hin­ton pre­dicted. “Irony is go­ing to be hard to get,” he said. “You have to be mas­ter of the lit­eral first. But then, Amer­i­cans don’t get irony ei­ther. Com­put­ers are go­ing to reach the level of Amer­i­cans be­fore Brits.” A flir­ta­tious pro­gram would prob­a­bly be quite sim­ple to cre­ate, how­ever. Many of the re­cent ad­vances in AI have sprung from the field of deep learn­ing, which Hin­ton has been work­ing on since the 1980s. At its core is the idea that com­puter pro­grams learn how to carry out tasks by train­ing on huge datasets, rather than be­ing taught a set of in­flex­i­ble rules. With the ad­vent of huge datasets and pow­er­ful pro­ces­sors, the ap­proach pi­o­neered by Hin­ton decades ago has come into the as­cen­dency and un­der­pins the work of Google’s ar­ti­fi­cial in­tel­li­gence arm, Deep­Mind, and sim­i­lar pro­grams of re­search at Face­book and Mi­crosoft. Hin­ton played down con­cerns about the dan­gers of AI raised by those such as the Amer­i­can en­tre­pre­neur Elon Musk, who has de­scribed the tech­nolo­gies un­der devel­op­ment as hu­man­ity’s great­est ex­is­ten­tial threat. “I’m more scared about the things that have al­ready hap­pened,” said Hin­ton. “The NSA is al­ready bug­ging ev­ery­thing that every­body does. Each time there’s a new rev­e­la­tion from Snow­den, you re­alise the ex­tent of it.” “I am scared that if you make the tech­nol­ogy work bet­ter, you help the NSA mis­use it more,” he added. “I’d be more wor­ried about that than about au­ton­o­mous killer ro­bots.”

Can an­droids pray for dig­i­tal sal­va­tion?

Can ro­bots make good Chris­tians? As com­puter science races ahead, at least one for­ward-look­ing Florida pas­tor sees a fu­ture for the faith in what­ever passes for a soul in ro­bots, an­droids, cy­borgs and other forms of ar­ti­fi­cial in­tel­li­gence. No, I’m not mak­ing this up. When the Rev. Christo­pher Benek, an as­so­ciate pas­tor of the First Pres­by­te­rian Church of Fort Laud­erdale, talks about ar­ti­fi­cial in­tel­li­gence, or “AI,” he wrote in a re­cent on­line es­say, “I am not talk­ing about iPhone’s Siri, a Roomba vac­uum or one of those toast­ers that can make per­fectly timed toast with a like­ness of Je­sus on it. ... I am talk­ing about an au­ton­o­mous crea­ture that has self-aware­ness.” When some­thing not only can think, rea­son, plan, learn, com­mu­ni­cate and per­ceive things, but also “feel love, sad­ness, com­pas­sion, joy, af­fec­tion and a mul­ti­tude of emo­tions,” he wrote, then it is not a great leap to think that “an AI that is very much like us but ex­po­nen­tially more in­tel­li­gent (could) par­tic­i­pate in Christ’s re­demp­tive pur­poses in the world” and “help to make the world a bet­ter place.” I see his point, although they also could make the world a worse place. Imag­ine, for ex­am­ple, ro­bots of dif­fer­ent de­nom­i­na­tions get­ting into a dis­pute over who has the best lock on eter­nal life af­ter their lease on this life burns out — if it ever does. Yet, at a time when much of the re­li­gious and po­lit­i­cal world seems to be at war with science, Benek has gained in­ter­na­tional at­ten­tion with his vi­sion­ary ideas about how ethics and moral­ity can sur­vive in our rapidly chang­ing techno-fu­ture. Ever since IBM’s Wat­son com­puter beat two for­mer win­ners on “Jeop­ardy!” in 2011, in­ter­est in ar­ti­fi­cial in­tel­li­gence seems to have ac­cel­er­ated, along with anx­i­eties about what it means for the fu­ture of us mere hu­mans. Best-sell­ing au­thor Ray Kurzweil, a direc­tor of en­gi­neer­ing at Google, has be­come the most widely known prophet of “sin­gu­lar­ity,” the widely the­o­rized time, per­haps as soon as 20 or 30 years from now, when com­put­ers will be as smart as hu­mans — and pro­ceed im­me­di­ately to be­com­ing much smarter than hu­mans. The chill­ing pos­si­bil­ity that, like Ben­der, the rogu­ish robot on “Fu­tu­rama,” fu­ture AIs might want to do with­out us “meat­bag” hu­mans has caused wide­spread an­droid anx­i­ety. In Jan­uary, the fa­mous physi­cist Stephen Hawk­ing and ad­ven­tur­ous SpaceX CEO Elon Musk pledged to do all they can to make sure that ar­ti­fi­cial in­tel­li­gence will ben­e­fit hu­mankind and not de­stroy our species. Good luck, guys. Mean­while, trep­i­da­tion about our robot fu­ture seems to be pop­ping up with new vigor in popular cul­ture, where science fic­tion has long been an out­let for our industrial-age anx­i­eties. The new movie “Ex Machina” of­fers Ava, a strik­ingly at­trac­tive fe­male hu­manoid, and the haunt­ing ex­is­ten­tial ques­tion, “Does Ava ac­tu­ally like you? Or is she pre­tend­ing to like you?” Only a month ear­lier we had “Chap­pie,” the story of a po­lice droid who be­comes the first robot with the abil­ity to think and feel for him­self. Ad­ven­tures en­sue. Still to come: “Avengers: Age of Ul­tron,” in which the vil­lain­ous robot taunts in the pre­views that, like Walt Dis­ney’s Pinoc­chio, “There are no strings on me.” And at Christ­mas, we are sched­uled to see Ge­orge Lu­cas’ lat­est “Star Wars” se­quel. That means the re­turn of star droids R2-D2 and C3PO with the sort of AI that we hu­mans love: They don’t let their su­pe­rior in­tel­li­gence go to their heads — or wher­ever else their cen­tral pro­cess­ing units might be in­stalled. If the sci-fi world is our guide, public con­cern about the fu­ture power of AI looms in the back­ground of our lives whether we want to con­front it di­rectly or not. Some sort of reg­u­la­tory safe­guards might well be in or­der, but we can hardly ex­pect Wash­ing­ton law­mak­ers to help us get along with ro­bots when they can hardly get along with one an­other. Be­sides, AI is way be­yond the tech­ni­cal know-how of a Congress that seems barely able to fig­ure out net neu­tral­ity. They aren’t alone. Mean­while, re­search in ar­ti­fi­cial in­tel­li­gence and our un­cer­tain robot fu­ture forges ahead. Benek’s ideas about bring­ing sal­va­tion to ro­bots doesn’t sound so nutty af­ter all. He ac­tu­ally raises an im­por­tant ques­tion: If your su­per­com­puter loses its moral or eth­i­cal way, who’s go­ing to tell it?

Col­lege pushes farm drones

When Clark State Com­mu­nity Col­lege Pres­i­dent Jo Alice Blondin first came to Ohio nearly two years ago, she knew us­ing drones with its agri­cul­ture pro­gram “was a real op­por­tu­nity.” Clark State isn’t the first com­mu­nity col­lege to ven­ture into us­ing un­manned aerial ve­hi­cles, also known as un­manned aerial sys­tems and com­monly ref­er­enced as drones. Sinclair Com­mu­nity Col­lege in Day­ton first looked into us­ing drones dur­ing a 2008 trade mis­sion trip to Is­rael, and a hand­ful of other U.S. com­mu­nity col­leges have de­vel­oped some type of drone pro­gram. But ear­lier this month Clark State be­came the lat­est in­sti­tu­tion of higher ed­u­ca­tion to re­ceive a cer­tifi­cate of au­tho­riza­tion to fly drones over fields near the Spring­field-Beck­ley Mu­nic­i­pal Air­port. Th­ese drones will col­lect data for stu­dents in the school’s two-year pre­ci­sion ag pro­gram, which be­gan this past Au­gust. “It gives our stu­dents real-time data that they can an­a­lyze and they can help real farm­ers with,” said Blondin, who spoke with this news­pa­per in Wash­ing­ton, D.C., as a guest of Speaker of the House John Boehner for this past Tues­day’s State of the Union ad­dress. “And with one in seven jobs in Ohio be­ing re­lated to agri­cul­ture, and within our re­gion there are nearly 1,200 farms, this is just a great op­por­tu­nity to test their skills.” The hope is to ac­quire ad­di­tional cer­tifi­cates to fly over more lo­cal farms and col­lect more data for the farm­ing in­dus­try. “In­dus­try will de­velop around how the farmer uses (the data),” Blondin said. “For ex­am­ple, he sees if there are low lev­els, or var­i­ous lev­els of chem­i­cals, and he needs to pro­vide chem­i­cals in this field — and I’m talk­ing down to the mil­lime­ter now, so you’re not wast­ing. “You’re ac­tu­ally sav­ing quite a bit of money and time by us­ing the data by tar­get­ing the real prob­lems that are in your field.” The ear­li­est drone data will be used in classes would be this May or next Jan­uary, said Blondin, whose pro­gram is re­ceiv­ing support from Spring­field-based Selec­tTech GeoSpa­tial. But those who en­tered the as­so­ciate de­gree pro­gram this past Au­gust will be able to grad­u­ate at the ear­li­est in May 2016. U.S. Sen. Rob Port­man, R-Ohio, said he has “worked ex­ten­sively” to strengthen unique drone ag ca­pa­bil­i­ties and in­crease Ohio’s lead­er­ship in de­fense and aero­space. “I am com­mit­ted to ad­vanc­ing our state’s lead­er­ship role in un­manned sys­tems,” said Port­man. “Ohioans have a unique abil­ity to de­velop and build cut­ting-edge tech­nolo­gies and aero­space equip­ment with work done at our aca­demic in­sti­tu­tions like Clark State, the Air Force Re­search Lab at Wright-Pat­ter­son, NASA Glenn Re­search Cen­ter in Cleve­land and in­dus­try part­ners.” U.S. Sen. Sher­rod Brown, D-Ohio, said the Mi­ami Val­ley is a hub for drone re­search, train­ing and de­vel­op­ment and “Clark State is build­ing on that mo­men­tum by train­ing stu­dents for the jobs of to­mor­row. “By ap­ply­ing UAS tech­nolo­gies to the agri­cul­ture in­dus­try, we help bol­ster Ohio’s lead­ing in­dus­try by en­sur­ing pro­duc­ers have the re­sources nec­es­sary to com­pete,” Brown said.

A year in gad­gets

One chap­ter in the saga of Ed­ward Snow­den, the Amer­i­can com­puter pro­fes­sional who leaked clas­si­fied in­for­ma­tion, was about how the US’ Na­tional Se­cu­rity Agency (NSA) was snoop­ing on phone con­ver­sa­tions, in­clud­ing those of politi­cians. In In­dia, the gov­ern­ment sought as­sur­ances from the US after rev­e­la­tions that the Bharatiya Janata Party was tar­geted by the NSA in 2010. We don’t know who is lis­ten­ing to our phone con­ver­sa­tions or read­ing our text mes­sages, but some­one prob­a­bly is. Gov­ern­ments can also col­lect user data from other sources, in­clud­ing cloud stor­age. Ad­di­tion­ally, web­sites, ad­ver­tis­ers and spam­mers could be ac­cess­ing your texts and brows­ing his­tory. Want to stop them? There are a few phones and apps out there that en­crypt con­ver­sa­tions. Black­phone (www.black­phone.ch) runs a cus­tom­ized se­cure ver­sion of An­droid—Pri­vatOS. The con­trol cen­tre lets you man­age which hard­ware or ser­vice each app can ac­cess—per­haps you don’t want a mes­sag­ing app to have ac­cess to the phone’s di­aller. Silent Cir­cle (www.silent­cir cle.com), which also part­nered the Black­phone prod­uct, of­fers a se­ries of se­cure apps for An­droid and iOS—mes­sag­ing, con­tacts, and en­crypted call­ing pack­ages. Whis­per Sys­tems (www.whis­per ys­tems.org) has the free-todown­load RedPhone and Tex­tSe­cure apps for An­droid. It has been sug­gested that Black­Berry Mes­sen­ger was the back­bone for con­ver­sa­tions be­tween pro­test­ers and their lo­cal/for­eign sup­port­ers dur­ing the London ri­ots in 2011. Twit­ter as­sumed that role dur­ing the 2010 pro-democ­racy demon­stra­tions known as the Arab Spring. And when the Turk­ish gov­ern­ment shut down parts of the In­ter­net dur­ing the 2013 Tak­sim Square protests—op­pos­ing the re­con­struc­tion of Ot­toman-era mil­i­tary bar­racks as a mu­seum-cum-shop­ping cen­tre—peo­ple started us­ing vir­tual pri­vate net­works (VPN) to ac­cess blocked com­mu­ni­ca­tion apps. In Hong Kong this year, the app of choice for the pro-democ­racy pro­test­ers was FireChat. The “off-the-grid” fea­ture uses var­i­ous com­mu­ni­ca­tion sen­sors in the phone, and can com­mu­ni­cate with other de­vices in a 200ft range. The de­vel­op­ers, Open Gar­den, re­ported that over 100,000 users from Hong Kong signed up within 24 hours on a late Septem­ber day. With the Chi­nese gov­ern­ment known to turn off the In­ter­net at the slight­est pre­text, this off­line app worked well. As we ush­ered in 2014, smart­phone class di­vi­sions were clear—af­ford­able phones were in­tensely frus­trat­ing to use. Users would have to spend up­wards of ` 30,000 for top-ofthe-line per­for­mance. But Xiaomi’s Mi3 packed in pow­er­ful hard­ware at a price point of around ` 15,000, and Mo­torola’s Moto E re­de­fined the mean­ing of an An­droid phone un­der ` 10,000. They changed things. The more fan­cied smart­phone brands have learnt the hard way. Google’s An­droid One project aims to take the smart­phone to the next bil­lion users. “Ter­ror­ists want An­droid One smart­phones more than AK-47s,” Prime Min­is­ter Naren­dra Modi said dur­ing a speech ear­lier this month in Jammu and Kashmir. Rais­ing funds for a business no longer needs suits, golf cour­ses, or reclu­sive billionaires. Now it can be demo­cratic, and on­line. Crowd­fund­ing plat­forms surged into the business space. Start-ups used so­cial me­dia as a plat­form to get in­vestors on board. For ex­am­ple, crowd­fund­ing plat­form Kick­s­tarter re­leased quar­terly num­bers at the end of Q1 2014—$1,244,868 (around ` 79,049,118) was pledged on av­er­age each day; 4,497 projects suc­cess­fully reached fund­ing goals; and 887,848 back­ers joined the plat­form. This is just the be­gin­ning. It took them a while, but com­pa­nies un­der­stood that smart­phones and tablets were be­com­ing pri­mary sources for ac­cess­ing, con­sum­ing and cre­at­ing con­tent. The PC has been left far be­hind. Ap­ple and Mi­crosoft are gear­ing their of­fice suites for phones. Ever­note’s Work Chat makes for­mal com­mu­ni­ca­tion crisp and breezy on a mo­bile de­vice. Banks are push­ing for mo­bile apps and on­line ser­vices. Mo­bile op­er­a­tors and their phone apps let users do ev­ery­thing from recharge to bill pay­ment. Plat­forms such as Paytm have evolved to of­fer a va­ri­ety of ser­vices—recharge, bill pay­ment and shop­ping—through one in­ter­face. Soon, you will be able to ditch all de­vices but one—the phone. “Ir­re­spec­tive of what de­vice it is, as long as you are us­ing a Mi­crosoft ser­vice, we are happy,” a Mi­crosoft ex­ec­u­tive, who didn’t want to be named, told us re­cently. The sub­se­quent an­nounce­ments tie in well with this change in think­ing. Of­fice 365 suite users now get un­lim­ited OneDrive stor­age (ear­lier it was 1 TB). MS Of­fice apps (Word, Ex­cel and Pow­erPoint) are now free to down­load and use on An­droid phones and iOS. Then there is the tie-up with Drop­box, and the of­fi­cial ex­ten­sion for the Chrome Web browser that lets users edit files within the browser—you don’t need to have MS Of­fice in­stalled. On­line stor­age space prices crashed in the sec­ond half of the year. The Mi­crosoft Of­fice 365 sub­scrip­tion now bun­dles un­lim­ited OneDrive stor­age. Google is of­fer­ing 100 GB space for $1.99 (around ` 125) a month. Box now of­fers 10 GB space with the free pack­age. Buy a new smart­phone, and chances are Google, Box or Drop­box will of­fer some free space. Smart­watches and fit­ness bands were a hot gad­get cat­e­gory through the year. Peo­ple were drawn help­lessly to the con­cept of ac­cess­ing apps, get­ting no­ti­fi­ca­tions and mon­i­tor­ing how many steps they had walked in a day through a sleek de­vice on their wrist. While they are costly, Vishal Gondal’s GOQii band did things a bit dif­fer­ently—adding a hu­man trainer to help you un­der­stand the data that the tracker gen­er­ates, and cre­ate a fit­ness sched­ule cus­tom­ized to each in­di­vid­ual. Mo­torola’s 360 is, so far, the best-look­ing smart­watch. The year of the selfie craze The Ox­ford English Dic­tio­nary’s def­i­ni­tion of selfie is “a pho­to­graph that one has taken of one­self, typ­i­cally one taken with a smart­phone or we­b­cam and up­loaded to a so­cial me­dia web­site”. For some rea­son, this be­came the most ex­cit­ing fad of 2014. Ev­ery­one went about click­ing them­selves (alone or with friends, against any back­drop pos­si­ble). Some­one try­ing to take a selfie with a gun man­aged to shoot him­self with both the cam­era and the gun. Another per­son got kicked in the head by the con­duc­tor of a mov­ing train for stand­ing too close to the tracks. There was some good stuff too: Cana­dian ad­ven­turer George Kourou­nis de­scended into a boil­ing lava lake in Van­u­atu’s Am­brym vol­cano with a GoPro cam­era and clicked a selfie be­fore heat made the cam­era un­us­able. South Korea has banned the use of un­reg­is­tered selfie sticks, a popular ac­ces­sory used to at­tach the phone to one end and hold it fur­ther away from your­self: They can ap­par­ently dis­rupt ra­dio fre­quen­cies. It is of­fi­cial. The Nokia Lu­mia smart­phones will now be known as Mi­crosoft Lu­mia. Mi­crosoft closed the Nokia deal ear­lier this year, but the re­brand­ing hap­pened in early De­cem­ber with the launch of Mi­crosoft Lu­mia 535. The more af­ford­able phones will con­tinue to sport the Nokia brand­ing, be­cause the brand had a bet­ter con­nect with a wider po­ten­tial cus­tomer de­mo­graphic. Last month, Nokia an­nounced the N1 An­droid tablet too—it clearly has plans beyond the Mi­crosoft ecosys­tem. Same day shipping and de­liv­ery will soon be rel­e­gated to the cat­e­gory marked snail’s pace. Ama­zon is work­ing on get­ting its drones into the air. The ser­vice is called Prime Air, and is cur­rently await­ing US Fed­eral Avi­a­tion Ad­min­is­tra­tion (FAA) ap­proval. Ama­zon says drones can help de­liver or­ders to cus­tomers in 30 min­utes. Th­ese drones can fly at 80.5km an hour and lift weights up to 2.3kg. Once the rules are in place, some­time in early 2015, we will see the first com­mer­cial tests. Real-world de­ploy­ment is still some way away, but is in­evitable.

Glimpse of the Fu­ture

CHANGES in tech­nol­ogy are hap­pen­ing at a scale which was unimag­in­able be­fore and will cause dis­rup­tion in in­dus­try af­ter in­dus­try. We are not ready for this change, and most of our lead­ing com­pa­nies won’t ex­ist 15­20 years from now. Here are five sec­tors to keep an eye on: 1. Man­u­fac­tur­ing. Ro­bot­ics and 3­D print­ing have made it cheaper to man­u­fac­ture in the United States and Europe than in China. Ro­bots such as Bax­ter, from Re­think Ro­bot­ics, and UR10, from Uni­ver­sal Ro­bots, have arms, screens which show you their emo­tions, and sen­sors that de­tect what is hap­pen­ing around them. The cost of oper­at­ing these is less than the cost of hu­man la­bor. We can now have ro­bots work­ing 24/7. These ro­bots will be­come ever more so­phis­ti­cated and do most hu­man jobs. The man­u­fac­tur­ing in­dus­try is surely go­ing to be dis­rupted in a very big way. This is good news for Amer­ica, Europe and parts of Asia, be­cause it will be­come a lo­cal in­dus­try. But this will be bad for the Chi­nese econ­omy — which is largely de­pen­dent on man­u­fac­tur­ing jobs. In the next decade, ro­bots will likely go on strike, be­cause we won’t need them any­more. They will be re­placed by 3D print­ers. Within 15 to 20 years, we will even be able to 3D print elec­tron­ics. Imag­ine be­ing able to de­sign your own iPhone and print it at home. 2. Fi­nance. We are al­ready wit­ness­ing a con­tro­versy over Bit­coin. Crowd­fund­ing is shak­ing up the ven­ture­cap­i­tal in­dus­try. We will soon be able to crowd­fund loans for houses, cars and other goods. With card­less trans­ac­tions for pur­chas­ing goods, we won’t need the types of phys­i­cal banks and fi­nan­cial in­sti­tu­tions we have. 3. Health care. Ap­ple re­cently an­nounced Healthkit, its plat­form for health in­for­ma­tion. It wants to store data from the wear­able sen­sors that will soon be mon­i­tor­ing our blood pres­sure, blood oxy­gena­tion, heart rhythms, tem­per­a­ture, ac­tiv­ity lev­els and other symptoms. Google, Mi­cro­soft and Sam­sung will surely not be left be­hind. With these data, they will be able to warn us when we are about to get sick. AI­based physi­cians will ad­vise us on what we need to do to get healthy. Med­i­cal­test data, es­pe­cially in fields such as on­col­ogy, is of­ten so com­plex that doc­tors can­not un­der­stand it. This will be­come even more dif­fi­cult when they have ge­nomics data to cor­re­late. When you com­bine these data with the med­i­cal­sen­sor data tech com­pa­nies are col­lect­ing on their cloud plat­forms, we will have a med­i­cal rev­o­lu­tion. We won’t need doc­tors for day­to­day med­i­cal advice. Ro­botic sur­geons will also do the most so­phis­ti­cated surg­eries. 4. En­ergy. Five years ago, we were wor­ried about Amer­ica run­ ning out of oil; to­day we’re talk­ing about Saudi Amer­ica, be­cause of frack­ing. And then there’s so­lar en­ergy. So­lar prices have dropped about 97 per­cent over the past 35 years, and by the end of this decade we will achieve grid par­ity. Grid par­ity means it’s cheaper to pro­duce en­ergy at home on your so­lar cells than to buy it from utilities. Move for­ward an­other 10 or 20 years, and it will cost a frac­tion as much to pro­duce your own en­ergy as to buy it from the grid. Util­ity com­pa­nies will be in se­ri­ous trou­ble. If so­lar keeps ad­vanc­ing in the way it is, it will eclipse the fos­sil­fuel in­dus­try. So­lar is only one of maybe a hun­dred tech­nolo­gies that could dis­rupt the en­ergy in­dus­try. When we have un­lim­ited en­ergy, we can have un­lim­ited clean wa­ter, be­cause we can sim­ply boil as much ocean wa­ter as we want. We can af­ford to grow food lo­cally in ver­ti­cal farms. 5. Com­mu­ni­ca­tions. Note how AT&T, Ver­i­zon and Sprint have seen their land­line busi­nesses dis­ap­pear. These were re­placed by mo­bile — which is now be­ing re­placed by data. When I travel abroad, I don’t make long­dis­tance calls any more, be­cause I just call over Skype. Soon we will have WiFi ev­ery­where. In prac­ti­cally ev­ery in­dus­try, I see a ma­jor dis­rup­tion. The vast ma­jor­ity of com­pa­nies that are presently the lead­ers will likely not even ex­ist. Ex­ec­u­tives ei­ther are not aware of the changes that are com­ing, are re­luc­tant to in­vest the money re­quired for them to rein­vent them­selves or are pro­tect­ing legacy busi­nesses. Most are fo­cused on short­term per­for­mance. New tril­lion­dol­lar in­dus­tries will come out of nowhere and wipe out ex­ist­ing tril­lion­dol­lar in­dus­tries. This is the fu­ture we’re headed into, for bet­ter or for worse.

Ama­zon work­ers get help­ing hand from robots to fill or­ders

Tracy, Calif. — This hol­i­day sea­son, Ama­zon’s lit­tle helper is an orange, 320pound ro­bot called Kiva. The robots — more than 15,000 of them com­pa­ny­wide — are part of Ama­zon’s high-tech ef­fort to get or­ders to cus­tomers faster. By lifting shelves of Ama­zon prod­ucts off the ground and speed­ily de­liv­er­ing them to em­ployee sta­tions, the robots dra­mat­i­cally re­duce the time it takes for work­ers to find items and put them into boxes for ship­ment. On the eve of Cy­ber Mon­day, Ama­zon’s year-old ware­house in Tracy, Cal­i­for­nia, was buzzing with ac­tiv­ity as the re­tailer pre­pared for one of the big­gest shop­ping days of the year. Kiva robots zoomed around the floor with un­canny pre­ci­sion, hoist­ing shelves con­tain­ing video games, col­or­ing books and stuffed an­i­mals. Yel­low bins filled with mer­chan­dise zipped by on con­veyor belts. A group of em­ploy­ees, some sport­ing red Santa hats, spent their break do­ing a se­ries of arm stretches. Since ac­quir­ing ro­bot-maker Kiva, a Mas­sachusetts company, for $775 mil­lion in cash in 2012, the e-com­merce re­tailer has been in­creas­ingly im­ple­ment­ing au­to­ma­tion at its gar­gan­tuan ful­fill­ment cen­ters. Ki­vas, which re­sem­ble over­grown Room­bas, are ca­pa­ble of lifting as much as 750 pounds and glide across Ama­zon’s ware­house floors by fol­low­ing rows of sen­sors. Sun­day was the first time Ama­zon pub­licly un­veiled the Tracy ware­house, which boasts the company’s lat­est “eighth-gener- ation” ful­fill­ment cen­ter tech­nol­ogy, in­clud­ing 3,000 Ki­vas. Ten Ama­zon ware­houses, in­clud­ing two in Cal­i­for­nia, are clas­si­fied as eighth-gen­er­a­tion; Ama­zon now has109 ful­fill­ment cen­ters glob­ally. Dave Clark, Ama­zon’s se­nior vice pres­i­dent of world­wide op­er­a­tions and cus­tomer ser­vice, said that be­cause Kiva-equipped fa­cil­i­ties elim­i­nate the need for wide aisles for hu­mans to walk down, eighth-gen­er­a­tion cen­ters can hold 50 per­cent more inventory than older ware­houses. More stor­age ca­pac­ity means a wider se­lec­tion of mer­chan­dise, fewer chances of prod­ucts be­ing out of stock and more pos­si­bil­i­ties for same-day de­liv­ery, he said. “It’s sort of a vir­tu­ous cy­cle,” he said dur­ing a tour of the fa­cil­ity. The robots, he added, have also cut pro­cess­ing times for or­ders to as lit­tle as 13 min­utes from well over an hour. The Tracy cen­ter, which is more than 1 mil­lion square feet in size and has 1,500 full-time em­ploy­ees, is still not at full ca­pac­ity. It cur­rently houses 21.5 mil­lion items (3.5 mil­lion unique SKUs), with plans to in­crease that num­ber to as much as 27 mil­lion items next year. On a peak day — such as Cy­ber Mon­day — the ware­house ships 700,000 items. Seat­tle-based Ama­zon has been adding new tech­nol­ogy to many of its ful­fill­ment cen­ters. Some ware­houses uti­lize RoboS­tow, a 6-ton ro­bot that moves mer­chan­dise pal­lets as high as 24 feet di­rectly onto Kiva robots, and “vi­sion sys­tems” that can re­ceive an en­tire trailer of inventory in as lit­tle as 30 min­utes by cap­tur­ing an im­age of the trailer’s con­tents. Work­ers at the Tracy fa­cil­ity said they en­joyed the ease of Kiva robots bring­ing prod­ucts di­rectly to them, though some con­ceded they missed walk­ing the for­mer maze of aisles to find prod­ucts. Clark de­clined to dis­cuss the cost of each ro­bot, or how much Ama­zon spent build­ing the Tracy ware­house. But he noted, “we’re very happy with the eco­nomics.” He said in­creased au­to­ma­tion hasn’t led to re­duced staffing lev­els at newer ware­houses be­cause the company con­tin­ues to grow rapidly. “That growth is driv­ing in­creased hir­ing. We con­tinue to add em­ploy­ees, and no em­ployee has been neg­a­tively im­pacted by Kiva com­ing on board,” he said.

© PressReader. All rights reserved.