The Lights in the Tunnel Automation Accelerating Technology and the Economy of the Future_4 pptx

Automation, Offshoring and Small Business We tend to think of automation and offshoring as primari-ly impacting jobs in large corporations.. Instead of making a large ment in sophisticat

chess moves, then why can it not also iterate throughevery known legal argument since the days when Ciceroheld forth in the Roman Forum? Would this be a “lesser”form of legal creativity? Perhaps it would But would thatmatter to our lawyer’s employer?

Although the practical applications of artificial gence have so far emphasized brute force solutions, it is by

intelli-no means true that this is the only approach being taken inthe field A very important area of study revolves around

the idea of neural nets, which are a special type of computer

that is built upon a model of the human brain Neural netsare currently being used in areas such as visual patternrecognition In the future, we can probably expect someimportant advances in this area, especially as the engineerswho design neural nets work more closely with scientistswho are uncovering the secrets of howour brains work.One thing that probably jumps out at you as we speak

of lawyers and radiologists is that these people make a lot

of money The average radiologist in the United Statesmakes over $300,000 In fact, we can reasonably say thatsoftware jobs (or knowledge worker jobs) are typicallyhigh paying jobs This creates a very strong incentive forbusinesses to offshore and, when possible, automate thesejobs Another point we can make is that there is really norelationship between how much training is required for ahuman being, and how difficult it is to automate the job

To become a lawyer or a radiologist requires both collegeand graduate degrees, but this will not hold off automa-tion It is a relatively simple matter to program accumu-

lated knowledge into an algorithm or enter it into a base.

data-For knowledge workers, there is really a double dose

of bad news Not only are their jobs potentially easier toautomate than other job types because no investment inmechanical equipment is required; but also, the financialincentive for getting rid of the job is significantly higher

As a result, we can expect that, in the future, automationwill fall heavily on knowledge workers and in particular onhighly paid workers In cases where technology is not yetsufficient to automate the job, offshoring is likely to bepursued as a interim solution

Given this reality, it may be that the simulation weperformed in Chapter 1 was actually somewhat conserva-tive Look back at the table listing traditional jobs on page

59 Very few of these people are knowledge workers Inour simulation, we assumed that automation would fallevenly on some significant percentage of the average lights

in the tunnel We now see, however, that automation may,

in fact, arrive in a relatively “top heavy” pattern It maywell be that a great many of the brighter lights in our tun-nel will be among the first impacted

What does this mean for a business that offers ucts and services in the mass market? Clearly, it impliesthat automation may be poised to someday eliminate notjust untold millions of your potential customers—it is like-

prod-ly to hit hard at your best customers.

Automation, Offshoring and Small Business

We tend to think of automation and offshoring as

primari-ly impacting jobs in large corporations After all, it takes asubstantial investment to set up a relationship with anoverseas outsourcing firm or bring in specialized automa-tion equipment or software In the near future, however,both of these practices are likely to become increasing ac-cessible and inexpensive for even the smallest businesses.There is a significant trend toward breaking jobs intosmaller pieces or specific tasks—which can then be eitherautomated or offshored This capability is increasingly be-ing offered to small businesses either as pre-packagedsoftware or through easy to use online interfaces over theInternet Tax preparation is one area where this approach

is already widespread Instead of making a large ment in sophisticated automation software, a small busi-ness owner or manager will be able to visit a website andthen rent access to the software on either a per-hour orper-task basis I think it is very possible that the same willhappen with task-specific offshoring Competition be-tween service providers will quickly produce lower prices,easier to use online interfaces, and a wider variety of ser-vices

invest-The result will be rapid penetration of these practicesinto businesses of all sizes As we saw with the radiologistand the lawyer, once significant portions of jobs can beautomated, the number of workers employed will imme-diately begin to fall The U.S Small Business Administra-tion estimates that businesses with fewer than 500 em-ployees have generated from 60-80 percent of all job

growth over the past decade.25 As it becomes easier andcheaper for business owners to employ automation andoffshoring, we may well find that these practices will be-come a significant drag on America’s primary job creationengine.

“Hardware” Jobs and Robotics

A “hardware” job is a job that requires some investment inmechanical or robotic technologies in order for the job to

be automated The automation of hardware jobs startedlong before the computer revolution Machines used onassembly lines, farm equipment, and heavy earth movingequipment are all technologies that have displaced millions

of workers in the past As history has shown, repetitivemotion manufacturing jobs are among the easiest to au-tomate In fact, as I mentioned, this is how the Ludditemovement got started back in 1811 However, the merger

of mechanics and computer technology into the field ofrobotics will almost certainly impact an unprecedentednumber and types of jobs Whether a specific hardwarejob is difficult or easy to automate really depends on thecombination of skills and manual dexterity required.For an example of a job that is very difficult to auto-mate, let’s consider an auto mechanic A mechanic ob-viously requires a great deal of hand-eye coordination He

or she has to work on thousands of different parts in avariety of different engines, often in highly varied states ofrepair In other words, a robot mechanic would face manyvisual recognition and manipulation problems similar tothe ones we discussed earlier with the robot housekeeper

In addition, the robot mechanic would require a muchhigher degree of problem solving skill than the housekee-per In fact, this diagnostic skill it not something thatcould be solved with software alone because it extends tonearly all the human senses A mechanic may listen to thesound an engine makes or even diagnose a problem based

on a specific smell

As things stand, we can say that becoming an automechanic is probably a pretty safe choice for the time be-ing But, as we said with the housekeeper, that does notimply the job will be safe forever Advances in robotictechnology will continue relentlessly until many of theseproblems are solved However, an even more importantfactor is likely to be changes made to the cars the mechan-

ic is working on Advancing technology has already pacted the way mechanics work; computerized diagnostictools are now used to read fault codes provided by micro-processors embedded in the engine We can expect thatthis trend will continue, and that, at some point in the fu-ture, cars may well be designed specifically to be worked

im-on by robotic mechanics

A truck driver is another example of a job that is

like-ly to be protected for the time being, but, in the long run,the reason will probably not be so much technology associal acceptance The military is already making substan-tial investments in automated trucks that could be used onthe battlefield These could be completely autonomous, orthey might be programmed to simply follow a lead truck.Similarly, many car manufacturers will soon be deployingcollision avoidance technology in cars These systems will

help drivers avoid mistakes that might lead to accidents;however over time they could evolve into technology ca-pable of driving the car autonomously—just as jet airlinersnowroutinely fly and land without assistance.

While the technology for automated cars and trucksmay arrive, it is somewhat difficult to imagine that mostpeople would be eager to share the road with 50-ton dri-verless trucks A second important issue would likely bethe power of the Teamsters union Once again, however, Ihave to give my standard disclaimer: this does not meantruck driving jobs will always be protected

The job types that are likely to be threatened fairlysoon by advances in robotics are the jobs that fit some-where between the auto mechanic and the repetitive mo-tion assembly line worker As an example, consider theshelf stocker in a supermarket or chain retail store Thisjob requires more flexibility than working on an assemblyline, but still falls far short of what the auto mechanic fac-es

The layout of a supermarket is standardized and couldeasily be programmed into a computer The aisles are wideand the floors are smooth; ideal territory for an industrialrobot Every item has a specific place on the shelves Barcodes make it a simple matter to identify items, and speciallocation markers could be placed on the shelves: a shelfstocking robot faces few of the visual recognition issuesthat challenged our housekeeping or auto mechanic ro-bots Designing a robot that could move inventory fromthe stock room and place it on shelves is certainly wellwithin the realm of possibility in the not too distant future

Needless to say, if a robot can be designed to stockshelves, then it can also be made to unload trucks andmove material of all types.

Skeptical that robots might someday be stepping into

these jobs? Consider that as far back as 2005, CNET News

Blog published an article entitled “Why so Nervous about

robots, Wal-Mart?”26 The article pointed out that reportshad surfaced about Wal-Mart testing inventory-taking ro-bots These would be robots programmed to navigate theaisles at night and automatically take a complete store in-ventory When the CNET reporter contacted Wal-Martmanagement, he received an unusually abrupt denial thatWal-Mart was considering using robots in any way

We can take Wal-Mart’s management at its word andassume that it, in fact, has no plans to use robots In thelong run, however, that won’t matter At some point, ifone of Wal-Mart’s competitors tries to gain an advantage

by employing robots, then Wal-Mart and every othercompeting business will really have no choice but to fol-low suit The point of this is not to vilify Wal-Mart or anyother business that might someday choose to employ au-tomation We have to acknowledge that, in a free marketeconomy, every business has to respond to its competitiveenvironment and employ the best available technologiesand processes If it does not do so, it will not survive.History has shown that job automation very often in-volves pushing a significant portion of the job onto thecustomer Automation in the customer service area is real-

ly self-service This has been the case with ATMs, automated

checkout aisles and even self-serve gas pumps In the

re-cently opened Future Store27 near Düsseldorf, Germany,in-store retail sales and customer assistance is being auto-mated via a cell-phone interface Shoppers are able to getreal time assistance, while shopping, through their mobilephones They can also scan bar codes as they shop and, inthe near future, will be able to pay for their purchases di-rectly through their phones—presumably avoiding thecheckout isle altogether.

The specter of near fully automated supermarkets andchain retail stores is cause for genuine concern These arenowthe jobs of last resort These are the jobs that workersdisplaced from other industries take because there is noth-ing better available Look back at the table on page 59 Wehave already mentioned that 3.5 million cashiers are poten-tially at risk The table shows another 4 million retail sales-persons and 2.3 million laborers and freight, stock and ma-terial movers, as well as 1.7 million stock clerks and orderfillers What new jobs could we possibly find for all thesepeople?

Read any article in the popular press about the field ofrobotics and its potential future implications, and you willalmost invariably find a sentence pointing out that “in thefuture, robots will be used to perform tasks which aredangerous for people, or jobs which people don’t reallywant.” That is surely true, but it implies the somewhat

wistful assumption that robots won’t be used in jobs that people do want That is obviously a silly assumption Ro-

bots, and other forms of automation, will be used instead

of people as soon it becomes cost effective and profitablefor businesses to do so.*

“Interface” Jobs

A third type of job is what I call an “interface” job Thepeople who hold these jobs, to a large extent, fill in thecracks which currently exist between various informationformats and technologies As an example, consider whathappens when you apply for a home mortgage loan If youwork with an independent mortgage agent, he or she willprobably give you a paper application to fill out Next, youwill need to retrieve and make copies of your supportingdocumentation: pay stubs, tax returns, bank statements,insurance documents, etc

All of this documentation will be on paper or it will

be faxed to you A property appraisal will be done, and thereport will be forwarded to the loan agent Once the loanagent collects everything together, he or she will probablyfax it all to the bank, where a loan officer will review it.Ultimately, numbers such as your salary, credit rating, andthe equity to loan ratio will be plugged into a computerprogram and the loan will either be approved or denied.Clearly, the bulk of the labor associated with thisprocess is in collecting, copying, collating and faxing in-formation The intellectual portion of the job—either ap-proving or denying the loan—is probably already essential-

ly handled by a computer Throughout the economy, there

* For more on robotics and its potential impact on employment and

on society, see Marshall Brain’s “Robotic Nation” blog at

http://roboticnation.blogspot.com.

are probably thousands of jobs for clerks and office ers that continue to exist because of this clunky interfacebetween what exists on paper and what needs to be in acomputer.

work-Clearly, we cannot expect that this state of affairs willcontinue forever Financial statements are already availableonline Standard data formats are making it increasinglyeasy for computers to talk directly to one another The

“XML” standard is a very popular format that is alreadywidely used to move data between different businessesover the Internet Using XML, the computers at a manu-facturing company can talk directly to the computers be-longing to the company’s suppliers The continuing drivetoward paperless documents and seamless communication

is likely to eliminate many of these human interface jobs inthe coming years

The Next “Killer App”

Since the beginnings of the personal computer industry,computer hardware sales have often been driven by a par-ticular software application so compelling that it has moti-vated customers to purchase the machine required to run

it When the Apple II was introduced in 1977, it was itially a success within a relatively small group of computerhobbyists It wasn’t until the first electronic spreadsheet,VisiCalc, was developed that the Apple II began to gener-ate wider interest VisiCalc was the catalyst that helpedtransform the Apple II from an interesting toy into a truebusiness machine Likewise, when the IBM PC was intro-duced, Lotus 1-2-3 fulfilled the “killer app” role Later, it

in-was graphic design and desktop publishing software thatdrove the Apple MacIntosh to success.

In recent years, the highest sales growth for the puter industry has not been in high-end desktop comput-ers but instead in laptops and, lately, the newer netbookmachines that provide a simple and inexpensive way tobrowse the web At least in part, this probably results fromthe fact that the acceleration of computer hardware capa-bility has largely outpaced what is required to run most ofthe software applications of interest to the average user Ifyou are primarily interested in word processing, spread-sheets and web browsing, it may be difficult to justify thecost of a high-end computer when a lower cost or moreportable machine offers more than enough power to runthe software Likewise, it seems to be increasingly difficultfor Microsoft and other software vendors to continuallyadd new features to desktop productivity applications andoperating systems that are compelling enough to justifyexpensive upgrades

com-Yet the business models of both Intel and Microsoftdepend on continuing to sell ever more powerful proces-sors and new or updated software applications to take ad-vantage of that power If customers were to permanentlyturn away from the idea of faster processors, the businesswould quickly become commoditized, and Intel wouldlose its competitive advantage For that reason, we can besure that Intel, Microsoft and hundreds of other softwarecompanies are actively seeking the next killer app—something that will fully leverage the vastly increased

computer power that will be available in the coming yearsand decades.

I think that there are good reasons to believe that thisnext killer app is going to turn out to be artificial intelli-gence (AI) AI applications are highly compute intensiveand will take full advantage of all the computational powerthat newprocessors can offer Newstandalone AI applica-tions will appear, but more importantly, artificial intelli-gence is likely to be built directly into existing productivityapplications and operating systems, as well as the enter-prise software and database systems used by large busi-nesses

The market for AI software is likely to extend farbeyond the computer industry Increasingly sophisticatedrobots will demand the most advanced hardware andsoftware available High-end microprocessors and AIsoftware will also surely be used to build intelligence intoappliances, consumer devices and industrial equipment ofall kinds Ultimately, robots and other non-computer ap-plications may well eclipse the personal computer market

as the primary growth engine for leading-edge hardwareand software

Products that give some insight into what the futuremay hold are already on display Microsoft recently dem-onstrated a “virtual personal assistant” which appears as acomputer generated person on the screen The assistant iscapable of tasks such as making airline reservations orscheduling meetings and requires the most advanced

hardware available According to the New York Times,

Mi-crosoft’s virtual assistant can “make sophisticated

deci-sions about the people in front of her, judging things liketheir attire, whether they seem impatient, their importanceand their preferred times for appointments.”28 The Times

article also quotes a Microsoft executive who speculatesthat future applications might include a “medical doctor in

a box” that could help with basic medical issues

An artificial intelligence application that could pense basic medical advice is certainly a compelling idea,especially in light of the continuing problem with accele-rating health care costs However, it raises an importantpoint What education and training would we require of a

dis-person who dispensed such information? Would this dis-person

need to be a doctor? Perhaps not, but clearly this wouldnot be one of the low skill, low wage jobs that we oftenassociate with vulnerability to automation The reality isthat there is simply little or no relationship between thelevel of education and training required for a person to do

a job and whether or not that job can be automated Whiledoctors are probably not in danger of losing their jobs inthe foreseeable future, the same cannot be said for manythousands of knowledge workers and middle managers inthe private sector

It’s important to note that, while humanoid interfaceslike Microsoft’s virtual assistant make for great demonstra-tions, the AI applications that will likely displace know-ledge workers will not need such elaborate interfaces Theywill simply be workhorse programs that make the routinedecisions and perform the tasks and analysis that are cur-rently the responsibility of highly paid workers sitting incubicles all over the world AI capability may start out by