The Green Light / 207
CopyrightedMaterial–Paperback/Kindle available @ Amazon
prosperity.Thisiswhycapitalismhassucceeded,andoth-
ereconomicsystemshavefailed.
In time, I thinkit is likely that our perceptions will
shift so that we begin to truly recognize the economic
contribution that our individual marketplace decisions
make. Someday,the majority of peoplewill bevaluedin
economic terms not for what they directly produce, but
fortheirparticipationinconsumptionmarkets.Ifwecan
succeed in gradually extending that participation to the
billions ofpeoplewhoarenow trappedinpoverty—and
do so in a way that creates incentives to conserve re-
sources and minimize environmental impact—we may
findthattheresultingconsumerdemandiscapableoffu-
eling an engine that can drive us to unprecedented eco-
nomicheights.
The Green Light
The natural cycle in the tunnel is stable and reinforcing.
The vast majority of the consumers in the tunnel now
glowwithapredominantlygreenlight.Astimepasses,the
collectiveintensityofthelightscontinuestograduallyin-
crease.
Suddenly, we see that vast numbers of dim green
lights have begun to stream into the tunnel. These new
lights have barely enough intensity to make it past the
threshold,butonceinside,theyjointheriveroflightsasit
courses overthe panels onthe tunnelwalls. At first, we
sensethatthebusinessesinthetunnelarestrainingsome-
what to meet this new demand, but as time passes, the
cycleagainstrengthens.Thecollectiveintensityofthelight
THE LIGHTS IN THE TUNNEL / 208
CopyrightedMaterial–Paperback/Kindle available @ Amazon
in the tunnel begins to quickly increase. We also notice
that,overtime,eachofthenewdimlightsisverygradually
becomingbrighter.
As the newlightsstream into thetunneland arein-
corporatedintoitsnaturalcycle,weseethatsomebright
white lightsbegintopulse withnewenergy.New panels
arenowappearinginmanyplacesonthetunnelwallsthat
wereoncedark.Theentrepreneursandbusinessownersin
the tunnel are responding to the rapidly increasing de-
mand.
As the number of lights continues to increase, the
speed with which panelsareupdatedand thenumber of
newpanelsappearingonthetunnelwallsseemstoaccele-
rate. Although we had perceived the tunnel as being al-
mostinfinitelyvast,itnowappearsthatthewallsarecom-
pletelycoveredwithpanels.
Evenaswesensethis,however,thetunnelitselfbe-
ginstoexpand.Asnewpanelsrapidlyfillthespacesonthe
expandingtunnelwalls,wenoticethatsomeofthebrigh-
testwhitelightsarenowradiatingwithanunprecedented
intensity.Still,astheever-increasingcycleoflightcontin-
ues to parade through the expanding tunnel, we sense
stronglythatitisthe seeminglyinfinite number ofgreen
lights thattrulyencapsulates the collective energy, enter-
priseandhopeofallhumanbeings.
CopyrightedMaterial–
Paperback/Kindle available @ Amazon
APPENDIX / FINAL
THOUGHTS
AretheideaspresentedinthisbookWRONG?
(Opposingargumentsandmyresponses)….
Page
210
TwoQuestionsWorthThinkingAbout……
Page
223
Wherearewenow?F
ourPossibleCases……
Page
224
TheNext10-20years:SomeIndicatorsto
WatchFor.……………………………………
Page
227
OutsmartingMarx……………………………
Page
237
TheTechnologyPara
dox……………………
Page
239
MachineIntelligenceandtheTuringTest…….
Page
241
THE LIGHTS IN THE TUNNEL / 210
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Are the ideas presented in this book WRONG?
(Opposing arguments with responses)
In this section I have listed some of the arguments that
maybemadeagainsttheideasinthisbook,togetherwith
myresponses.Theseareeitherconventionalargumentsor
thingsIhavethoughtoforseenelsewhere.
The economy will always create new jobs; we will
never have structural unemployment as a result of ad-
vancing technology
Thisistheideabehindthe“Ludditefallacy”whichIdis-
cussedatsomelengthinChapters2and3(seepages95
and 131). At present, I suspect that most economists
wouldprobablybelikelytoagreewiththisstatementand,
therefore, disagree with what I have suggested in this
book.Here,inanutshell,ismyargumentforwhyIthink
wewillendupwithaseriousunemploymentproblem:
Astechnologyadvancesandindustriesautomate,this
improvestheefficiencyofproductionandtendstomake
the products and services produced by those industries
more affordable. That leaves more purchasing power in
thepocketsofconsumers.Thoseconsumersthengoout
andspendthatextramoneyonallkindsofproductsand
servicesproducedbyavarietyofindustries.Someofthose
industries are relatively labor intensive, so they have to
hire more workers to meet this demand—andso overall
employmentremainsstableorincreases.Thisisthereason
Appendix / Final Thoughts / 211
CopyrightedMaterial–Paperback/Kindle available @ Amazon
that, historically, technology has not led to sustained,
widespreadunemployment.
Myargumentisthatacceleratingautomationtechnol-
ogywillultimatelyinvademanyoftheindustriesthathave
traditionallybeenlaborintensive.Additionally,theprocess
of creative destruction will destroy old industries and
createnewones,andveryfewofthesenewindustriesare
likelytobelaborintensive.Asaresult,theoverallecono-
mywillbecomelesslaborintensiveandultimatelyreacha
“tipping point.” Beyondthis point, the economy will no
longerbeabletoabsorbtheworkerswholosejobsdueto
automation: businesses will instead invest primarily in
moremachines.Ihavealsoarguedthatthisprocesswillbe
relentless,andifitisnotaddressedbysometypeofgov-
ernmentpolicy,wemayultimatelyseeaprecipitousdrop
inconsumerspendingasasubstantialfractionofthepop-
ulation loses confidence in its future income continuity.
That,ofcourse,wouldresultinevenmoreunemployment
andadownwardspiralwouldensue.
THE LIGHTS IN THE TUNNEL / 212
CopyrightedMaterial–Paperback/Kindle available @ Amazon
If technology resulted in unemployment, everyone
would already be out of a job because technological
progress has been going on for hundreds of years
Thisreallyjustamountstosaying“ithasn’thappenedyet,
so it will never happen.” History has proven time and
againthat,wheretechnologyisconcerned,somethingcan
beimpossiblesincethedawnofcivilizationandthensud-
denly, in the blink of an eye, become possible.Revolutio-
narytechnologies,suchastheairplaneandnuclearpower,
wherealldismissedas being impossible evenbypreemi-
nentscientistswhowereinvolvedintheresearchthatled
totheirdevelopment.
Today,mostofusacceptthattechnologywillcontin-
uetoadvanceandproducethingsthatwemightcurrently
viewasimpossible.However,westillthinktoonarrowly.
Weacceptthattherewillbenewtechnologies,newprod-
uctsandnewindustries,butmostofusarenotprepared
toacceptthatallthiswillchangethebasiceconomicrules
that we take for granted. But why wouldn’t that be the
case?Isthereafundamentalreasonwhyacceleratingtech-
nologyshouldimpactnearlyeveryaspectofourlives—but
notimpactthewaytheeconomyworks?AsIpointedout
inChapter2,advancinginformationtechnology—because
itenabledthecreationanddistributionoffinancialderiva-
tives—hascertainlyplayedanimportantroleintheseveri-
tyofthecurrenteconomiccrisis.Isuspectthatthisisjust
a preview of the economic impacts that technology will
haveinthefuture.
Appendix / Final Thoughts / 213
CopyrightedMaterial–Paperback/Kindle available @ Amazon
The population is aging. Once the baby boomers re-
tire, we will have a worker shortage not unemploy-
ment
Innearlyallthedevelopednations(andalsoinChina),the
populationsarerapidlyagingandretirementplansarepro-
jected tocome under intensepressure,astoofewyoung
workers have to support too many older retired people.
Doesthisimplythatthereislikelytobeanoverallshort-
ageofworkersaslargenumbersofolderpeopleleavethe
workforce? I think it is certainly possible this may be a
counteractive force that might tend to delay the impact
fromautomationtoacertainextent.Herearesomethings
toconsider:
Theimpactofautomationonaspecificjobcategoryis
reallynotrelatedtothenumberofworkersavailableto
perform that type of job. Once technology advances
tothepointwhereatypeofjobcanbeautomated,the
machinestodothiscaneasilybereplicated.Machines
donotneedtobeeducatedortrained,andsotheyare
notsubjecttothebottlenecksthatcreateshortagesof
workersinfieldssuchasnursing.Therefore,inconsi-
deringtheoverallimpactofmachineautomation,the
important criterion is not the number of workers
availablebut thetypes of jobsthat canbe automated.
Totheextentthatthereareworkershortageswithina
specific job category, that would actually tend to in-
creasetheincentiveforautomationtechnologiestobe
developedinthatarea.WealreadyseethiseffectinJa-
panwheresignificantworkisbeingtodonetodevelop
THE LIGHTS IN THE TUNNEL / 214
CopyrightedMaterial–Paperback/Kindle available @ Amazon
robots capable of assisting with care of the elderly
population.
Thecurrentconsensusviewseemstobethat,asare-
sult of the 2008-9 financial crisis (and its impact on
401k plans), many workers will remain in the work-
forcelongerthanoriginallyplanned.Thiswillgiveau-
tomationtechnologylongertocomeintoplaybefore
anyworkershortagesmaterialize.
Eveniftheaging populationdoestendtoretardun-
employment,thiswould,ofcourse,simplybeadelay-
ingfactor—notalong-termsolutiontotheproblem.
Ithinkwecancertainlyexpecttoseeworkershortag-
esinsomeareas,butthismayverypossiblybecombined
withanoverallunemploymentproblem.Thedangeristhat
increasingstructuralunemploymentwillunfoldinparallel
withthedemographicproblem.Isuspectthatmostofthe
projectionsregardingtheimpactofagingpopulationsas-
sume reasonably full employment among younger work-
ers.Ifthisdoesnotturnouttobethecase,thesituation
willobviouslybemuchworse.AsIpointedoutinChapter
3, a payroll tax-based system for supporting retirement
programsmightbecomecompletelyunsustainable.
Appendix / Final Thoughts / 215
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Businesses will never fully automate because of the
high initial capital investment and the lack of flexibil-
ity this implies
There are somevalid pointshere, and I think thatthese
factorsmay,inmanycases,servetoretardtheprocessof
automation—butinthelongruntheywillnotpreventit.
Somebusinesseswillcertainlydelayautomationbecauseof
thehighcapitaloutlaysrequired.However,overtime,ma-
chines will become more affordable, more reliable, and
moreflexible.Atsomepoint,astechnologyadvances,ma-
chineswillbegintooutperformworkerstotheextentthat
a non-automated business will not be competitive. Con-
siderthecaseofonlinebanking:itgenerallyoffersarange
ofservices,suchasautomaticbillpaying,thatcouldnotbe
offeredbyahumanbankteller.
Keep in mind that automation offers cost benefits
beyond simply eliminating wages. There are also payroll
taxes, benefits, vacationtime, managementissues (and if
youeliminateworkers,youcaninmanycasesalsogetrid
ofthefirstlinemanagers),etc.Therearealsosafety and
liability issues; consider the safety advantages of a fully
automatedwarehouse.
Theneed for technicaland economic flexibilitymay
alsotendtoholdbackautomationforatime.Ifabusiness
investsheavilyinspecificmachinestoproduceaparticular
productandthenthatproductdoesnotsellwell,itmaybe
stuckwith equipment itdoes notwant.The obviousan-
swertothatis that,inthefuture,automationtechnology
willbemoreflexibleandeasytoadapttodifferentprod-
ucts.Ithinkthemanufacturersofautomationequipment
THE LIGHTS IN THE TUNNEL / 216
CopyrightedMaterial–Paperback/Kindle available @ Amazon
arefullyawareofthisissueandwillbuildincreasinglyflex-
ibleproducts.
Thereisalsotheissueofeconomicflexibility:abusi-
nessthatemploysworkerscanlaythemoffinslowtimes,
while a more automated business will be stuck with its
machines.Again,Ithinkthat,inthelongrun,astechnolo-
gy advances, businesses that don’t automate will simply
not be able to compete: and that reality will overwhelm
otherconsiderations.
Anotherpointisthatbothofthesefactors(highcapi-
talcostsandtheneedforflexibility)maytendtopushthe
next wave of automation toward software applications
geared toward eliminating knowledge worker jobs. Soft-
ware is typically more flexible and has a lower up front
costthanexpensivemechanicalautomation.AsInotedin
Chapter 2, automation of these jobs, together with off-
shoring, may mean diminishing prospectsfor knowledge
worksandcollegegraduatesingeneral.
Machines may take over most unskilled labor, but
they will never be able to do skilled or professional
jobs that require lots of training and education
I think this is a dangerous misconception that stems, in
part, from a certain amount of hubris on the part of
peoplewhoare welleducated.Theconventional wisdom
isthatafencehasbeenerectedwithinoursociety.Onthe
lush, garden side of the fence, are workers who have
strongeducationsandtraining.Thesepeoplearebeneficia-
riesoftheinformationage.Onthetoxicwastelandsideof
the fence, are relatively unskilled workers. These people
Appendix / Final Thoughts / 217
CopyrightedMaterial–Paperback/Kindle available @ Amazon
havebeenheavilyimpactedbybothtechnologyandgloba-
lization.Theyoftensurvivebystringingtwoorthreepart
timejobstogetherorworkinlowwagejobswithfewben-
efits.Theobvioussolutionisforustofindawaytooffer
thesepeopleadditionaltraining—sotheycanhopoverto
thegoodsideofthefence.
Ithinkthattheproblemwiththisscenarioisthatthe
fenceisgoingtomove,anditmaymoveveryrapidly.Thegood
sideofthefenceisgoingtocontract,andincreasingnum-
bersofwell-educatedworkersaregoingtofindthemselves
suddenlyonthetoxicside.AsIpointedoutinChapter2,
we can expect fields such as artificial intelligence to ad-
vance rapidly in the coming years and decades. While
many college-educatedknowledge workers perform tasks
thatarecurrentlybeyondthecapabilityofcomputers,they
nonethelesshavejobsthatcanbebrokendownintoarela-
tively narrowset of tasks and routines. Over time, these
jobs will be increasingly subject to automation, and may
wellcomeundersignificantpressurefromoffshoringeven
sooner.Asthistrenddevelops,Ithinkthatthepsycholog-
icalimpactonconsumerswillrepresentasignificantdan-
gertotheeconomy.
EvenifIamwrongandincreasingunemploymentis
confinedprimarilytolowerskill workers, the overall im-
pact on our society and economy would be dramatically
negative.IntheUnitedStates,twothirdsofworkers—and
thereforeconsumers—donothaveacollegedegree.While
effortstoimproveeducationandtrainingarelaudable,the
realityisthatthisisnotlikelytooffsettheimpactofgeo-
metricallyacceleratingtechnology.Infact,Ithinkthatthe
THE LIGHTS IN THE TUNNEL / 218
CopyrightedMaterial–Paperback/Kindle available @ Amazon
argumentsI have put forthinthis book remainrelevant
even in the (I think unlikely) event that more educated
workersaresparedtheimpactofautomation.
A great many jobs require people skills and, there-
fore, could not be performed by machines
Ithinktherearedefinitelysomejobsthatmeetthisstan-
dard,butprobablynotanywherenearenoughtoavoidthe
overallproblemthatwouldresultfromautomation.Ialso
thinkthatpeopleoftentendtounderestimatetheextentto
whichtheirjobmightbesusceptibletoautomation.Bank
tellershave“people”jobs,butthathasnotstoppedpeople
fromusingATMsoronlinebanking.Ingeneral,consum-
ers seem quite receptive to automation and self-service
technologiesiftheyprovideconvenience.
Many workers whose interactions with other people
areprimarilyinternaltotheirorganizationmightfeelthat
they have jobs in which people skills are critical. If you
think about it, however, that is true only if most of the
otherjobsarealsobeingdonebypeople.Onceautomation
takesholdwithinanorganization,suchpeopleskillsmight
eventuallybefarlessimportant.
Finally, it is important to note that automation will
havebothadirectandanindirectimpactonjobs.Evena
personwhoholdsa job thatis completelysafe fromau-
tomation might still be impacted by declining consumer
demandresultingfromjobselsewhereintheeconomybe-
ingautomated.
Appendix / Final Thoughts / 219
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Perhaps machines will ultimately take over most of
the work done by people, but this wont happen until
far into the future (hundreds of years, etc.)
Ingeneral,if myargumenthereturnsouttonot becor-
rect,thentherearereallytwobasicwaysinwhichitcanbe
wrong:
The argument could be fundamentally wrong. This
implies that the economy is capable of growing and
advancing technologically basically forever, without
creating an unemployment problem. In other words,
even 500 years from now, when society presumably
has technologythat is incomprehensible to us today,
theeconomywillstillprovideemploymentforthevast
majorityofpeopleinthepopulation.
The argument could be premature. Maybe the ideas
here are basically correct, but they won’t come into
play until far into the future. However, once we ac-
knowledgethatatsomepointinthefuture,theecon-
omy may become nearly fully automated, then as a
matter of mathematics, we cannot get to that point
withoutfirstpassingthroughatippingpoint—beyond
which structural unemployment will begin to be a
problem. The tipping point could occur long before
weexpectit.
IwouldguessthatifthetrendsIhavepresentedhere
donotdevelop,thenitwillmostlikelybebecauseIhave
gotten the timing wrong. However, I do think there is
someprettycompellingevidencethatwearefairlycloseto
THE LIGHTS IN THE TUNNEL / 220
CopyrightedMaterial–Paperback/Kindle available @ Amazon
thepointwhereacceleratingmachinecapabilityisgoingto
haveadramaticeconomicimpact(ifwehavenotalready
passed thatpoint).Pleaseseethe section“Wherearewe
now?FourPossibleCases”laterintheAppendixformore
onthis.
In the future, wages/income may be very low be-
cause of job automation, but technology will also
make everything plentiful and cheap so low income
wont matter
Thisisanideathatisoftenexpressedinconjunctionwith
adiscussionofadvancednanotechnology.Nanotechnolo-
gymayonedayofferthepromiseofmaterialobjectscon-
structed molecule-by-molecule, perhaps using self-
replicatingtechnologies.Somepeoplearguethat,takento
the extreme, thismightmean thatphysicalobjects could
beconstructedinawaythatisanalogoustothecreationof
virtual objects displayed on a computer screen (think of
the“replicator”inStarTrek).Thereare,ofcourse,afew
problemswiththis:
Expendituresonmanufacturedgoodsrepresentafair-
lysmallpercentage oftheaverageconsumer’sspend-
ing. Expenses like housing
*
and health care are far
*
Itisperhapsconceivablethatjobautomationmaysomedayleadto
somewhatlowerhousingcostsbecauseitcouldresultinalotofempty
officetowers andcommercial buildings.Thosebuildings might then
beconvertedtootheruses—includingperhapshousing.Ifknowledge
workerandofficejobsmigrateintothecomputernetwork,thereally
hot commercial real estate in the future might be in the locations
where companies likeGoogleand Microsoft are now building huge
Appendix / Final Thoughts / 221
CopyrightedMaterial–Paperback/Kindle available @ Amazon
more important. So even if the price of most goods
fell dramatically, consumers would still need an ade-
quateincome.
Today, we already have digital products that have a
zeromarginalcostofproduction.Theseproductsare,
forthemostpart,notfreebecausetheyhaveintellec-
tual propertyrights attached tothem.Wecan expect
thesameifadvancednanotechnologyarrivessomeday.
If there isn’t a way to protect and profit from these
propertyrights,itis veryunlikelythatinvestorswould
provide the enormous sums necessary to realize the
technology.
Advancednanotechnologyalmostcertainlyliesfurther
in the future than the automation technology that is
likely to threaten routine jobs. So it won’t arrive in
timetosolvetheprobleminanycase.
The Heads in the Sand Objection
IfotherargumentsagainsttheideasIhavepresentedhere
proveinsufficient,thenIsuspectthatmanypeoplewillbe
temptedtoturntothisone:
Somepeoplewillrejecttheideathatmachinesmight
begintoexhibitsomedegreeof intelligence—and,there-
fore,achievethecapabilitytoperformagreatmanyjobs—
simply because the implicationsare very difficult to deal
with. This irrational, but perhaps understandable, objec-
“serverfarms.”Theseareusuallyrelativelyisolatedlocationsfarfrom
naturaldisastersandotherthreatsandclosetoclean,reliable energy
(whichtodaymostlymeanshydroelectricpower).
THE LIGHTS IN THE TUNNEL / 222
CopyrightedMaterial–Paperback/Kindle available @ Amazon
tion to the idea that machines might someday begin to
think and reason was first articulated by the founder of
computerscience,AlanTuring(Pleaseseethelastsection
ofthisAppendix).
Turinginitiatedthefieldofartificialintelligencewith
his1950paper“ComputingMachineryandIntelligence.”
Here’showTuringexpressedwhathecalledthe“Headsin
theSand”Objection(which,ofcourse,herejected):
“Theconsequencesofmachinesthinkingwouldbetoodreadful.Let
ushopeandbelievethattheycannotdoso.”
55
Appendix / Final Thoughts / 223
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Two Questions Worth Thinking About
One
Whileyoumightnotagreethatitwilleverbeareality,itis
easytoimagineaneconomywithnohumanworkers.Ob-
viously if autonomous machines could do everything,
peoplecouldspendtheirtimedoingwhatevertheyliked.
Canyouimagineamarketeconomywithnoconsumers?
Two
Most economists would probably agree that long-term
economicgrowthandprosperity(perhapsasmeasuredby
growth in per capitaGDP)istiedprettycloselyto tech-
nologicalprogress.Thisisthesameassayingthatsociety
becomes more wealthy largely because the machines we
use to produce goods and services get better over time.
Weallassumethateconomicgrowthcancontinueindefi-
nitely. This implies that machines will have to continue
gettingbetterbasicallyforever.
Canamachinekeepgettingbetterforeverwithoutsomeday
becomingautonomous?
THE LIGHTS IN THE TUNNEL / 224
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Where are we now? Four Possible Cases
On the next page, I’ve reproduced the scarygraphfrom
Chapter3,whichshowsthepotentialimpactofmachines
throughout the economy becoming more autonomous.
Remember that this graph shows average income for a
“median range” ofpeople in theeconomy; it’s basedon
whatstatisticianswouldcalla “truncatedmean”withthe
poorestandwealthiestpeopleremovedbeforecalculating
theaverage.Thereasonfordoingthisisthatweknowau-
tomation will tend to concentrate income and perhaps
dragupoverallaverageincome—atleastforawhile.Ulti-
matelyhowever,thecollectiveimpactonthebulkofcon-
sumerswillbecomeoverwhelming,andagraphofperca-
pitaGDPshouldhaveasimilarshape.
If the basic shape of this graph is correct, then the
realquestionis:whereareweonthegraphnow?Ithink
there are four possible cases, as shown on the diagram.
KeepinmindthatwhilethegraphIhavedrawnissmooth
toshowthegeneraltrend,inrealitythegraphwouldlikely
have many short-term ups and downs. Thiswill makeit
verydifficulttofigureoutwhereweare.
Appendix / Final Thoughts / 225
CopyrightedMaterial–Paperback/Kindle available @ Amazon
Value Added (Wage) of Average Worker Operating Average Machine
Also: Overall Wealth of Society (GDP per capita will look similar)
1. IftheideasI’vepresentedinthisbookarewrong,then
the conventional wisdom is correct, and the current
crisis is just an aberration. We should eventually get
backontrackandcontinueclimbingthegraph.
2. If the shapeofthegraph is basicallycorrect,but we
arestillfarawayfromthepointwhereautomationis
goingtobecomeimportant,thenweshould,likewise,
getbackontrackandcontinueclimbing.
3. Ifwearesomewhereclosetothepointwherethedot-
tedandsolidlinesdiverge,thenwearegoingtoseein-
creasingeconomicimpacts,andwewillhavedifficulty
in achieving sustained, long-term growth. If I had to
bet,Iwouldchoosethiscase.
Machines Getting
Better
Machines Fully
Autonomous
Time
Machines Becoming
Autonomous
Conventional Wisdom
(Most economists
believe this)
Value
THE LIGHTS IN THE TUNNEL / 226
CopyrightedMaterial–Paperback/Kindle available @ Amazon
4. Ifthingshavegottenawayfromus,thenwecould,in
fact, be much further along than we imagine. This
could perhaps be explained by suggesting that con-
sumer borrowing masked the reality of the situation
overthelastfewyearsandthatthecurrentcrisisisthe
beginning of the reckoning.This is anugly scenario,
butIdon’tthinkitcanbedismissedcompletely.Ob-
viously,ifthisisthecase,weneedtoadoptnewpoli-
ciesrapidly.
Appendix / Final Thoughts / 227
CopyrightedMaterial–Paperback/Kindle available @ Amazon
The Next 10-20 years: Some Indicators to
Watch For
TheideasthatIhavepresentedinthisbookmightbesaid
toconstituteahypothesis,orperhapsatheorythatisstill
lacking in substantial supporting evidence. This book is
obviouslynotbasedonananalysisofhistoricaleconomic
data,butratheronarational,andIthinksomewhatcon-
servative, projection of existing trends in information
technology.Nonetheless,Ibelievethatthereisalreadyat
leastsomeempiricalevidencetosupportthishypothesis.
I cannot reasonably expect anyone to accept these
ideasonthebasisofwhatispresentedherealone.Myin-
tentinwritingthisbookhasbeentotrytoraisethegener-
allevelofawarenessregardingthisissue.Myhopeisthat
economistsandothersmaybegintolookforevidenceof
the trend toward an automated economy with an open
mind,sothatwewillhaveareasonablechanceofaddress-
ingtheriskswemayfaceinthefuture.Towardthatend,
herearesomegeneralindicatorsthatIthinkmayconsti-
tute reasonable evidence that we are moving along the
pathIhavesketchedouthere.
Weak consumer spending pushes investment toward
cost cutting and labor saving technologies
Nearlyeveryoneseemstobeprojectingthatwhenrecov-
eryfromthecurrentdownturncomes,itwillberelatively
weak. Economists are once again anticipating a “jobless
recovery.”(Whenisthelast timewehadarecoverythat
wasn’tjobless?)Inthefaceoflingeringunemploymentand
THE LIGHTS IN THE TUNNEL / 228
CopyrightedMaterial–Paperback/Kindle available @ Amazon
tightcredit,consumerdemandinthe United Statesisal-
mostcertaintobeunimpressiveforsometime.Thewind
has been knocked out of the world economy’s primary
consumptionengine,anditremainsunclearwheresustain-
able futuredemandwillcomefrom. I’veseen severalar-
ticlesinthefinancialpressthatpoint outthatourfuture
prosperitywilllikelydependonusconsuminglessandex-
porting more. Yes, but who is going to buy all those ex-
ports?
Ifprojectionsforconsumerspendingremainunopti-
mistic,manybusinessesarelikelytoholdbackongeneral
technologyinvestmentastheywaitforamoresustainable
recovery.Asaresult,wemaycontinuetoseerelativelylow
levels of venture capital flowing into start-up firms for
some time. In the midst of this, it may become evident
that one of the few bright spots is the market for new
technologyproductsthatresultinimmediatecostsavings.
Wemightseeventurecapitalincreasinglybegintoflowto
start-upcompaniesthatarefocusedonlaborsavingtech-
nologiessuchasroboticsandartificialintelligence.
*
Some
ofthesenewventuresmightfocusonembeddingintelli-
genceintotheenterprisesoftwareusedbylargecorpora-
tions,while others createtoolsthat canbeusedinsmall
businessesviaInternetinterfaces.Significanteffortislikely
tobeputintomachinelearningtechnology,sothatauto-
mationalgorithmscanbeeasilytaughttoperformavariety
*
Obviously, even if thecurrentconsensusiswrongand weendup
withasurprisinglystrongrebound,wearestilllikelytoseesignificant
investmentintheseareas,aswellasamuchhigherleveloftechnology
investmentingeneral.
Appendix / Final Thoughts / 229
CopyrightedMaterial–Paperback/Kindle available @ Amazon
ofjobs.Becauseautomatingthejobsofrelativelyunskilled
workersoftenrequireshighcapitalinvestmentinmechan-
icallycomplexmachines,itmaywellbeofficeandknow-
ledgeworkerswhoaretheprimaryinitialtargetsofthese
newtechnologies.
Offshoring and automation begin to penetrate small
business and possibly combine to capture higher val-
ue jobs
Isuspectthatmosteconomistsdiscountthepotentialfor
outsourcing to invade the small business arena because
theyassumethatthecostsandinconveniencesassociated
withsettingupoffshorerelationshipsconstituteabarrier.
Myconcernisthattheoffshoringindustry,especiallyifit
facesdiminisheddemandinotherareas,willmovetomi-
nimize this barrier by adopting Internet-based interfaces
thatmakeiteasyforsmallerbusinessestooffshorespecif-
icjobsandtasks—withoutanyneedtotravelorengagein
complex negotiations. I think it is likely that exactly the
samethingwillhappenwithincreasinglysophisticatedau-
tomation software. If thiswere to occur, it would even-
tually undercut the U.S. economy’s primary job creation
enginetoasignificantdegree.
Asecondpossibilityisthatasincreasedinvestmentin
artificial intelligence-driven productivity tools begins to
bearfruit,thesetoolsmaybe combinedwithhighlyedu-
cated, but young and inexpensive, offshore workers to
capturejobswhicharecurrentlyheldbyhighlypaidknow-
ledge workers with deep experience. As AI software ad-
vances, it will increasingly begin to encapsulate what we
THE LIGHTS IN THE TUNNEL / 230
CopyrightedMaterial–Paperback/Kindle available @ Amazon
think of as “experience” and “judgment” within specific
fields of knowledge. An intelligent and educated young
worker wielding such toolsmighteventuallybecompeti-
tive with professionals and workers who currently com-
mandveryhighsalaries.
Labor intensive areas of the economy begin to see
increased automation
AsIhavepointedoutseveraltimesinthisbook,thepri-
mary danger to the U.S. economy will be when labor-
intensive industries, especially in the service sector, be-
comesusceptibletoautomation.Inareassuchasretailand
fastfoodwherewagesarelow,automationmaybeheldat
bay for a time by the high capital costs of automation
equipment. However, as technology advances and costs
fall,Ithinkitisinevitablethatatsomepoint,thetradeoff
will begin to make sense and competitive pressures will
pushbusinessesandindustriestowardautomation.
New technology industries fail to create significant
numbers of jobs
Wecanexpectthattechnologicaladvancewillgiveriseto
entirelynewindustriesinthefuture.However,thereality
isthatfewifanyofthesearelikelytobelaborintensive.
Bytheirverynature,thesenewindustrieswilltendtorely
on information technology and will offer relatively few
opportunitiesforaverageworkers.Thereisalsoariskthat
these newindustries maydirectlycompete withand ulti-
matelydestroyexisting,morelaborintensiveindustries.
Appendix / Final Thoughts / 231
CopyrightedMaterial–Paperback/Kindle available @ Amazon
One exception to this may be the so called “green
collar” jobsthatinvolveinstallingsolar panels, windtur-
bines, and so forth. These, however, are really one-time
infrastructure jobs. Alone, they will not be sufficient to
providesustainableemploymentgrowth.
Diminishing prospects for college graduates
Unfortunately,as I suggested inChapter2, I thinkthere
are reasons to expect that the assumption that a college
degreeisatickettosuccessmayultimatelybechallenged.
Thereareacoupleoftrendsthatmaydevelop:
Unemployment,underemploymentandlowwagesfor
recentgraduatesmaydevelopasautomation,andpos-
sibly offshoring, have a disproportionate impact on
entry-levelpositions.Thiswillbeanespeciallysignifi-
cant problem inlight of theenormous debt burdens
carriedbymanygraduates.
Older, highly paid knowledge workers, professionals
andmiddlemanagersmayfindthattheirjobspresent
plump targets for both offshoring and automation.
Manyofthesepeoplearelikelytobemiddleagedwith
substantialfamilyobligations,andwillhavefewgood
prospects. Age discrimination lawsuits may well clog
thecourtsofthefuture.
As I pointed out in Chapters 2and3, the long-term
impact of a diminishing economic incentive for average
peopletopursuefurthereducationwouldbedisastrous.In
myopinion,thisrealityprobablyconstitutesthesinglebest