Transcript Technology today and tomorrowx

TECHNOLOGY: TODAY AND
TOMORROW AND HOW DO WE
COPE WITH IT
Some common threads
Increasing returns to scale – as more products are consumed, the more profitable they become
Products are free or very cheap, but providers are very lean
Recombinant innovation
Bounty and spread
More value, but less of GDP
Winner takes all
Increasing replacement of labor by technology
Increasing returns to scale
With physical goods, the more a producer sells, the cost of raw materials and other factors of
production at some point of time start going up – at some point of time, there is diminishing
returns to scale
With digital goods, there is
◦ (often) Negligible marginal costs
◦ Network effects
Benefits – with a price
The cloud today allows these products to be distributed with very low cost
A boon for consumers, since these products are often free (Skype calls) or have a very low cost
compared to existing alternatives (Skype calls to traditional landlines)
However, these companies employ very few people
13-4
Benefits – with a price
Traditional companies (often brick and mortar companies) have large physical presences, which
required manpower, and such jobs often paid decent salaries
If they need to compete with the new players, they have to ditch their erstwhile business
models, firing much their existing workforce
Many industries are getting hollowed out in this fashion
Measuring value
Many digitized products deliver lots of value but do not contribute to GDP measures
McKinsey estimated the value of Internet services to be around $50 per person per month in the
US
Just like Big Data, there is a need to develop new metrics of measuring value and consumer
surplus from these products
Innovation – the recombinant view
Innovations do not get used up, but appear in other forms
“To invent something is to find it in what previously exists” – The Nature of Technology by Brian
Arthur
Economic growth occurs whenever people take existing resources and rearrange them in a more
valuable manner
The recombinant view - examples
Google’s self-driven car is essentially recombining existing technologies to make them more
useful – the internal combustion engine (or maybe battery technology in future), sensors, a fast
computer and Google Maps data
The Web itself is recombinant innovation: a network protocol TCP/IP, a markup language
protocol HTML and a PC application the browser
Facebook, Waze, Instagram – all use existing technologies to make them more useful
Recombinant innovation with digitization
When you recombine existing technologies that are in digital form, you can essentially create
building blocks for further innovations
◦ Existing web technologies combined = Facebook
◦ Facebook + other programming = Facebook apps
◦ Existing infrastructure + cheap sensors = better ways to maintain infrastructure
Digitization makes available massive amounts of data available for further analysis, and used in
parallel for many applications: building blocks of innovation available for the future
Moore’s Law just accelerates this trend exponentially into the future
Is innovation therefore limitless?
By this argument, innovation possibilities are limitless – just recombine existing digital building
blocks in new ways and create value
Not that simple: human beings have to enter the picture to figure out the new ways of
recombining
◦ In other words, we need huge number of eyeballs looking into these recombinations
New ways to look at innovation
Innocentive
Yet2
Kaggle
Quirky (also look at Quirky+GE)
IBM Jam
99designs
Challenge.gov
The Good Judgment project
One Billion Minds
Affinnova: a novel model of innovation
Bounty versus spread
Many trillions of photos have been taken ever, most of them in the past few years
Tens of billions of photos shared on Instagram so far, with a team of about 15 members
Compare that to Kodak, which at one point of time employed 1/3 of all people in Rochester, NY
(at its peak, nearly 150,000 people)
◦ George Eastman’s net worth at time of death $95 million
◦ He also created a thriving economy and middle class
◦ Facebook created 7 billionaires after its IPO, whose net worth was 10 times more than Eastman’s
The new economics
If products are available for free or near-free, and if a person is producing that product, then his
wage has to be less than what the product sells for
But if a product is available digitally available to a billion people, the person producing it gets the
individual surplus (x1 billion), with little increase in underlying costs
There is no law that says that this wealth should be spread
Who reaps the benefits?
Crowdsourcing creates enormous values for free – the creators of Facebook and YouTube are
extremely rich by getting the world to create their content for free
The Internet can be seen as a leveling force, but is it?
◦ To a cynic, it might seem as it created tools for the masses so that they can contribute their labor for
free
◦ “Everyone is free to play, but only a few reap the rewards”
◦ The Internet creates an incredibly efficient mechanism for harvesting the crowdsourced efforts in the
hands of a few organizations
GDP per capita and median family
income
The new uncoupling
For the first time in nearly 200 years of recorded history, wages have uncoupled from
productivity
Median salaries peaked in 1999, and has since then fallen
Income distribution is also becoming increasingly skewed
This is even affecting lifespans – the average white American male and female had decreasing
life expectancies from 1990 to 2008
The winners
Those who have the right assets
◦ Human (training and skills)
◦ Capital ( from machinery to financial assets)
The superstars (unique talent and/or luck)
Digital technologies have made these winners win more than those who have lost, but that is no
consolation for the latter
Skills-based earnings
Skills-based earnings
The role of technology in earnings
Routine jobs have been substituted by technology
Abstract jobs and roles have been augmented by big data and analytics and other ICTs
Net effect – increased demand for skilled labor and decreasing demand for unskilled labor
◦ More people earn advanced degrees than ever before, but even then, the relative demand for skilled
labor has gone up and that of unskilled labor down
The role of technology in earnings
Demand for non-routine jobs (cognitive or manual) has been steady or increasing, but that of
routine jobs has been going down
Once the recession was over, businesses did not feel the need for hiring for routine jobs and
replaced them with technology
Thus, technology is dividing the pie of earnings that go to labor versus capital more towards the
latter
Corporate profits and private wages
The net effect = rise of superstars
Share of income is getting higher at the top
◦ The top 1%
◦ The top 1% of the top 1%
◦ Slicing the top 1% at every level shows a fractal-like pattern
Digitization exacerbates these effects: think of the ways the Harry Potter novels can be
monetized, versus what Shakespeare (or even Tolkien) could do
◦ And the best gets monetized many, many times more than the next levels
◦ You would rather go with the best, so the best gets compensated many times more than the secondbest
Relative advantage = absolute benefits
The top quality provider can capture the entire market
◦
◦
◦
◦
◦
The next best provider might be nearly as good but it won’t matter
With digitization, capacity constraints lose meaning
You can reach out to the world without (technological) constraints
Network effects
Long tail effects
Which is why Facebook became the de facto social network within a very short period of time
Can bounty trump spread?
Technology has created bounty – in terms of new products, choice
But it also exacerbates the rise of superstars over everyone else
A case can be made that the gains of bounty do not overcome the lack of spread
◦ Too much of this and it can lead to lack of bounty in future
Technology and unemployment
One argument – technology can create Luddites, but the forces of capitalism creates
employment elsewhere
◦ Technology can lower prices, which increases demand for other goods (maybe newer goods), which in
turn leads to employment
The other argument – technology can create permanent unemployment if more and more jobs
get automated
◦ What if we are satisfied with the new cheaper goods and choose to consume less?
◦ What if technology changes faster than we and our institutions can adjust to? Replaced workers have to
learn new skills, which might be too much after a certain age
◦ The stigma of the long-term unemployed
Technology and unemployment
Further, as a result of technological progress, the cost of labor might go to near-zero, which is
not sustainable
◦ In fact, employers might find such labor unsuitable at even zero price
In fact, this is what seems to be happening since the 1990s: productivity decoupled from
employment
If it is indeed that Moore’s Law is behind this, the trends can only accelerate in future
r and g
r>g – the rate of return of capital versus economic growth: when r>g, there is economic
inequality and if it is easy to replace humans with technology, r will rise relatively faster than g
◦ Eventually the two will be the same (r cannot exceed g in the long run), but there can be a considerable
period of time when r>g, and technology can then feed unequal distributions of wealth
◦ That is when we will need policy interventions
Globalization
Can globalization explain the replacement of labor?
Maybe temporarily, but not over a longer horizon
◦ Increasingly, machines are replacing humans from the shopfloor, whether it is in the US or China or
India
◦ Foxconn
◦ Call centers being replaced with voice recognition software applications
◦ Offshoring is just an intermediate solution of the long-term trend: you cannot compete with Moore’s
Law in the long run
Implications for us
You don’t want to compete with machines which can be close substitutes for your labor –
especially ones that have or will soon have (Moore’s Law) cost advantages
Build on areas where humans are strong and machines are weak, so that they are complements:
in fact, it makes sense to be complements of something that is increasingly plentiful
“You will be paid in future based on how well you work with robots”
The role of creativity
Chess software perform best when handled by humans
◦ Strong machine < Strong player + weak machine
◦ “Weak human + machine + better process was superior to a strong computer alone, and more
remarkably, superior to a strong human + machine + weak process” – Kasparov
◦ Humans for strategic guidance
◦ Ideation: New ideas require humans
◦ But what about tomorrow? Will a “new idea” or “creativity” of today be just another computational
problem for a quantum computer?
What does it mean for us
When technology changes very fast that education cannot keep up, we run into problems: longterm unemployment, inequality, etc.
So education has to be a priority
There is a strong relationship between economic growth and higher test scores
Technology can help
The importance of networks
Remember that abstract, non-routine jobs are the ones we want to do – those that are
augmented by technology
Remember the nature of recombinant innovation: use existing building blocks in novel ways for
greater value
For that, we need lots of people looking at a problem
Thus the importance of networks
Innovation and serendipity often go together
Having networks help that process – more minds working at a problem
Find skills in non-traditional ways
Talent mismatch is costing the world economy $150 billion, according to LinkedIn estimates, but
the number might be much higher
◦ These numbers are based on additional productivity of the right workers and avoiding additional
recruiting costs
◦ But what about the option value of those networks not made?
topcoder
Knack.it (a news item)
Hireart
odesk and eLance
LinkedIn is developing an architecture by which skills can be found more easily
Will that take care of all problems?
Possibly not
Most people are not innovators, and thanks to Moore’s Law, the bar for what exactly constitutes
innovation is getting higher and higher
Why do we work?
◦ “The goal of the future is full unemployment, so we can play” – Arthur C. Clarke
◦ “Work saves a man from three great evils: boredom, vice and need” – Voltaire
◦ People want dignity, autonomy; people who work lead more fulfilling lives and live longer than those
who don’t
◦ All these have implications for taxes, social safety nets
Can there be other avenues for work?
Even big algorithms and big data need human tweaking
If big data is generating these huge amounts of messy data, can human beings help?
◦ Correctly labeling photos
◦ Extracting information from snaps of business cards
Amazon Mechanical Turk is now a thriving marketplace for crowdsourced work – an online
distributed problem-solving model, albeit not high-paying
Finding non-traditional avenues of work:
Rise of the peer economy
Airbnb
ClosetDash
Uber
Lyft
Sidecar
Relayrides
Getaround
Spinlister
Skillshare
Mealku
Taskrabbit
Zaarly
Rise of the peer economy
Estimated to be around $350 billion in transactions in 2013
It shows that there is a lot of resources (human or otherwise) that is not being used
Technology helps us extract value from such unused resources
It also points to the fact that our consumption models are being reshaped by technology
◦ The models of consumption and ownership are perhaps anachronistic
Peer-to-peer lending
LendingClub
Prosper
Fundingcircle
Kickstarter
Again, these models are meeting a need (both on the demand and supply side) that was not
being met by traditional financial entities, and technology has enabled that
The takeaway from sharing
The rise of the sharing economy is not yet significant, but could well be in future in tackling
unemployment
Uber growth is showing that it can radically change many existing models of not only
transportation but also of logistics
It also shows us how technology can surprise us in ways that we had not anticipated even a few
years back
There will be many new technologies coming up in our lifetime, and it will be up to us to be
open about them
◦ Flexibility will be key
◦ The exponential nature of progress of technology means that the best is always yet to come
Caution ahead
Because of Moore’s Law, errors and faults can get quickly compounded
◦ With everything connected on the Internet, a glitch somewhere in the system can cascade
catastrophically
◦ Tempting targets for criminals, hackers, spies
Technology today is powerful enough to destroy us
◦ Just like many things else in an increasingly digitized world, the limits of destruction will be less
constrained in future
Caution ahead
Will technology make us shallow?
◦ Where will investigative journalism come from if stories are ranked by algorithms?
◦ “Mile wide and an inch deep”
Will the Internet make us balkanized?
◦ We will stick only to those who correspond to our worldview: “ideological amplification”
◦ We can filter our world to one that corresponds to our worldview
This time it’s different?
Technology so far has always been for the good
So, history will suggest that we have reasons to be hopeful
What technology will do tomorrow will depend on what we decide as policymakers
◦ How do we support those who are left behind by technology?
◦ We need to think deeply what we will value
End credits
It is worth remembering that the specific innovations that we discussed in this course have in a
way been ‘factored in’ into our imagination; what will happen in future is almost certainly
different from what we discussed
Arthur C. Clarke’s “three laws of prediction”
1.
2.
3.
When a distinguished but elderly scientist states that something is possible, he is almost certainly
right. When he states that something is impossible, he is very probably wrong.
The only way of discovering the limits of the possible is to venture a little way past them into the
impossible.
Any sufficiently advanced technology is indistinguishable from magic.