Myths about technological change

Seven Myths About

Technology Change

A/Prof Jeffrey Funk

Division of Engineering and Technology Management

National University of Singapore

More details can be found in

1) What Drives Exponential Improvements? California Management Review, May 2013

2) Technology Change and the Rise of New Industries, book from Stanford University Press, January 2013

3) Presentations on slideshare: http://www.slideshare.net/Funk98/presentations

Seven Myths About Technology Change

Performance improvements follow S-curve Slow down in old technology causes

search for new technology Costs fall as cumulative production rises Demand drives improvements Media attention means a technology will

diffuse We can’t analyze the timing of new

technologies The market doesn’t work

Time

Performance

The Myth: Performance Improvements Follow

S-Curves (Jumps and Diminishing Returns)

Emergence of New Technology

Where’s the S-Curve (e.g., Jumps) in Moore’s Law?

You have seen Moore’s Law, Haven’t You?

Moore’s Law for

Microprocessors

Source: http://www.fgarciasanchez.es/thesisfelipe/node5.html

Where’s the S-Curve in Magnetic Recording Density of Platters for Hard Disk Drives?

Luminosity per watt (lm/W) of lights and

displays

Organic

Transistors

Where are the Jumps?

Co

erc

ivit

y(O

ers

ted

or

Am

ps/

Me

ter)

Figure 2.5 Improvements in Coercivity of Magnetic Materials

0.1

1

10

100

1900 1910 1920 1930 1940 1950 1960 1970 1980

SmCo MaBl

PtCo Ferrites

Alnico Alloys Steel

General Trend Line

Trend LineFor SamariumCobalt Magnets

Computer

Processing Speed

Where are the

S-Curves?






The Myth: Slow Down in Old Causes Search for New Technology

Technologies exhibit diminishing returns and perhaps limits

As they experience diminishing returns, new technologies are searched for, found and developed

This causes the new technology to experience improvements, perhaps even jumps in performance

The Reality (1)

For many technologies, limits, diminishing returns, and jumps are not clearly evident in time series (see previous figures and next slide for examples)

This suggests that new technologies are searched for, targeted, and developed long before old technologies exhibit limits or even diminishing returns

One reason is that while physical limits do exist, apparently few have been reached

Organic

Transistors

Where is evidence of limits leading to search for new technology?

The Reality (2)

Even if diminishing returns exist, development

of new technology begins much earlier

Why? There are many technologies and many

potential applications for them

These technologies are targeted and pursued

in decentralized manner

◦ Millions of research scientists and engineers

◦ They independently pursue many technologies long

before commercialization occurs

Look for new types of materials

Consider many applications

The Reality (3)

The result is that improvement curves are relatively independent of each other

Each technology is pursued as opportunity, both individual and organizational opportunity

Rates of improvement reflect ◦ ability to find improvements

◦ perceptions about the potential for improvements

Scientists and engineers pursue these technologies in order to obtain publications, patents, and perhaps fame

The Reality (3)

In any case, if we try to change distribution of research effort ◦ We should understand different technologies, their rates of improvements, and their potential for further improvements

Independent of application, we should fund those technologies with ◦ greatest rates of improvement and ◦ greater potential for improvements than do others

But in doing so we should fund many technologies and many forms of them






Costs fall as cumulative production grows in learning or experience curve ◦ One suggested mechanism is that automated

manufacturing equipment is introduced, modified, and organized into flow lines

But learning curve ◦ Can’t be used until production has begun

◦ Assumes all components are unique to new product

◦ Doesn’t help us understand why some technologies experience more improvements than do other technologies

◦ Ignores work done in laboratories

Myth: Cumulative Production Drives Cost Reductions

Creating materials (and associated processes) that better exploit physical phenomena

Geometrical scaling ◦ Reductions in scale: e.g., integrated circuits (ICs),

magnetic storage, MEMS, bio-electronic ICs

◦ Increases in scale: e.g., larger production equipment, engines, oil tankers

Some technologies directly experience improvements while others indirectly experience them through improvements in “components” ◦ Computers and other electronic systems

◦ Telecommunication systems

Reality: What Drives Improvements?

Luminosity per watt (lm/W) of lights and

displays

Organic

Transistors

Note the names of materials

Co

erc

ivit

y(O

ers

ted

or

Am

ps/

Me

ter)

Figure 2.5 Improvements in Coercivity of Magnetic Materials

0.1

1

10

100

1900 1910 1920 1930 1940 1950 1960 1970 1980

SmCo MaBl

PtCo Ferrites

Alnico Alloys Steel

General Trend Line

Trend LineFor SamariumCobalt Magnets


Figure 2.6 Improvements in Energy Product of Magnetic Materials

Ener

gy P

rod

uct

(Meg

a-G

auss

Oer

sted

s)

0.1

1

10

100

1900 1920 1940 1960 1980 2000

steel

alnico alloys

fine particles

rare earths

General Trend Line

0.01

0.1

1

10

100

1000

1960 1965 1970 1975 1980 1985

Op

tica

l Lo

ss (

db

/km

)

Figure 2.9 Reductions in Optical Loss of Optical Fiber

Critical Temperature

for Superconductors


Technology

Domain

Sub-

Technology

Dimensions of

measure

Different Classes of Materials

Energy

Trans-

formation

Lighting Light intensity per unit

cost

Candle wax, gas, carbon and tungsten filaments,

semiconductor and organic materials for LEDs

LEDs Luminosity per Watt Group III-V, IV-IV, and II-VI semiconductors

Organic LEDs Small molecules, polymers, phosphorescent materials

Solar Cells Power output per unit

cost

Silicon, Gallium Arsenide, Cadmium Telluride,

Cadmium Indium Gallium Selenide, Dye-Sensitized,

Organic

Energy storage Batteries Energy stored per unit

volume, mass or cost

Lead acid, Nickel Cadmium, Nickel Metal Hydride,

Lithium Polymer, Lithium-ion

Capacitors Carbons, polymers, metal oxides, ruthenium oxide, ionic

liquids

Flywheels Stone, steel, glass, carbon fibers

Information

Trans-

formation

Organic

Transistors

Mobility (cm2/ Volt-

seconds)

Polythiophenes, thiophene oligomers, polymers,

hthalocyanines, heteroacenes, tetrathiafulvalenes,

perylene diimides naphthalene diimides, acenes, C60

Living

Organisms

Biological

transfor-

mation

U.S. corn output per

area

Open pollinated, double cross, single cross, biotech

GMO

Materials Load Bearing Strength to weight ratio Iron, Steel, Composites, Carbon Fibers

Magnetic Strength Steel/Alnico Alloys, Fine particles, Rare earths

Coercivity Steel/Alnico Alloys, SmCo, PtCo, MaBi, Ferrites,

New Classes of Materials Enable Improvements over Long Time Scale








Figure 2. Declining Feature Size

0.001

0.01

0.1

1

10

100

1960 1965 1970 1975 1980 1985 1990 1995 2000

Year

Mic

rom

ete

rs (

Mic

rons)

Gate Oxide

Thickness

Junction Depth

Feature length

Source: (O'Neil, 2003)

Moore’s Law for

Microprocessors

Reductions in Scale Drive Improvements in Capacity

Magnetic Recording Density

Other Technologies Benefit from Reductions in Scale

MEMS (micro-electronic mechanical systems) for many applications ◦ Gyroscopes, resonators, micro-mirrors

◦ Photonics, ink jet nozzles for printers, micro-gas analyzers

Bio-electronic ICs for many applications ◦ Point-of-care diagnostics, drug delivery

◦ chips embedded in clothing, body, etc.

DNA sequencing

Nanotechnology








Scaling in Production Equipment

We all know about economies of scale ◦ But some products benefit from economies of scale more

than do others

◦ Why? Some products benefit from increases in scale of production equipment more than do others

Largest benefits for ◦ chemicals, other continuous flow equipment

◦ furnaces and smelters

Smaller benefits for discrete parts equipment

But also large benefits for ◦ Semiconductor wafers, liquid crystal display (LCD), and

solar cell manufacturing equipment

Production of Liquids or Gases in a Continuous Flow Factory

Liquids and gases are mixed, separated, heated, cooled, filtered, settled, extracted, distilled, and dried in pipes and reaction vessels

Pipes ◦ Cost is function of surface area (or radius) ◦ Output is function of volume (or radius squared)

Reaction vessels ◦ Cost is function of surface area (or radius squared)

◦ Output is function of volume (radius cubed)

Example of Benefits of Larger Scale: Engines

Diameter of cylinder (D)

Cost of cylinder

or piston is function

of cylinder’s surface

area (πDH)

Output of engine

is function of

Cylinder/piston’s

volume (πD2H/4)

Result: output rises

faster than costs as

diameter is increased

Height

of

cylinder

(H)

http://en.wikipedia.org/wiki/File:Cylinder_2_(PSF).png

1

10

100

1000

10000

0.1 1 10 100 1000 10000

Rela

tive

Pri

ce p

er

Ou

tpu

tRelative Price Per Output Falls as Scale Increases

Steam Engine (in

HP) Maximum scale:

1.3 M HP

Marine Engine

Largest is

90,000 HP

Chemical Plant:

1000s of tons of ethylene

per year; much smaller plants

built

Commercial aircraft

Smallest one had

12 passengers

Oil Tanker:

1000s of tons

Smallest was

1807 tons

Output (Scale)

LCD Mfg Equip:

Largest panel size is

16 square meters

Aluminum

(1000s of

amps)

Electric Power

Plants (in MW); much

smaller ones built








Computer Processing Speed: Driven by Improvements in ICs

Bandwidth/Speeds for Wireline Telecommunication: Driven by

improvements in ICs, optical fiber, lasers, and photosensors

Source: Koh H and Magee C, 20016, A function approach for studying technological progress: application to

Information technology, Technological Forecasting & Social Change 73: 1061-1983.

Computers (e.g., tablet computers) networks of RFID tags, smart dust, and

other sensors Cloud/utility computing Internet content (e.g., mashups, 3D

content, video conferencing) Human-computer interface (touch, gesture,

neural) Mobile phones and mobile phone systems

(e.g., 4G, 5G, cognitive radio) Autonomous vehicles Holographic display systems

ICs Drive Improvements in Many Systems






The Myth: Demand Drives Improvements

Demand drives cumulative production

Cumulative production drives improvements ◦ Automated manufacturing equipment is introduced, modified,

and organized into flow lines

◦ Better products and processes are introduced

Implications: stimulating demand will lead to cost reductions. This is one reason why many governments subsidize the introduction of clean energy more than they subsidize R&D spending

Clayton Christensen’s theory of disruptive innovation also implies that increases in demand will lead to reductions in cost and improvements in performance

But……

Many improvements occur without product demand

◦ Scientists and engineers create materials to exploit physical

phenomena long before technology is commercialized

Even with scaling, demand is indirect driver

◦ Demand does provide money for increasing scale of

production equipment or reducing scale of features on ICs

and magnetic storage

◦ But often complementary technologies such as new

equipment are the bottleneck

And for some increases in scale, they reduce rates

of increase in unit cumulative production (e.g.,

engines)

Time

Performance

Many also Argue that Increases in Demand

Lead to Accelerations in Performance During

First Half of S-Curve

Accelerations in Rates of

Improvement

But…..

Few performance curves display an acceleration, i.e., jumps, during the first half of an S-curve

What about the few that do?

Do the accelerations reflect increases in demand?

Let’s look at example of superconductors

Rate of improvement for maximum

critical temperature of

superconductors experienced

acceleration in mid-1980s

But what caused the acceleration?

Was it increases in demand for

superconducting materials?

No!

It was because scientists found

new and unexpected class

(ceramics) of superconducting

materials (later, red, black, green,

purple)

We don’t need more demand! We

need scientists and engineers to

look for and find new classes of

materials!

What Does “Open Innovation” Tell Us About

the Role of Demand?

In the old world of closed innovation, ◦ vertically integrated firms developed components for their systems

◦ thus demand for systems and components were somewhat linked

In the new world of open innovation, ◦ different firms develop systems and components, i.e., vertical disintegration

◦ Most firms develop components for multiple systems

◦ Thus weaker link between demand for specific systems and components

Source: Henry Chesbrough, Open Innovation:

The New Imperative for Creating and Profiting from Technology

What Does Open Innovation Tell Us About

the Role of Demand? (2)

Open innovation in R&D is extending the vertical disintegration backwards into research

R&D is now conducted in a very decentralized world, millions of research scientists and engineers now exist

Firms (and professors) do research even when final product demand doesn’t exist ◦ Government funding, wealthy entrepreneurs, venture capital, patent protection, and development prizes support this research

Source: Henry Chesbrough, Open Innovation:

The New Imperative for Creating and Profiting from Technology






The most dangerous “model” is hype

Visibility is not a signal of growth

Many technologies never experience

diffusion, even if they are visible

For Example, Lots of Hype about

Energy storage for vehicles

Smart grid

Wind turbines

Solar cells

Nanotechnology

And many other technologies

But which ones will succeed?

Which technologies will become economically feasible and thus diffuse?

It largely depends on the rate of improvement ◦ Fast rates of improvement increase the chances

that a new technology will become economically feasible

Fast rates of improvement reflect ◦ Creation of new materials

◦ Technologies that benefit from changes in scale and the implementation of these changes in scale

◦ Technologies that benefit from reductions in scale have particularly rapid rates of improvement

For Example, Consider Mobile Phones? (1)

In early 1980s, one study concluded there would be about 1 million mobile phones in use by 2000

Some would say we under estimated the need for mobile phones

I say we under estimated the impact of Moore’s Law on the cost of mobile phones

Lesson: pay attention to rates of improvement and not to hype (or lack of hype)

Mobile Phones? (2)

In early 2000s, many believed that location services were a huge market

Until recently no one used these services

Until recently some would say we overestimated the need for such services

I say we ◦ over estimated the impact of Moore’s Law on the cost of such services for short term

◦ under estimated the impact for long term

Lesson: pay attention to rates of improvement and not to hype






Myth: We can’t analyze the timing of new technologies, because….

New technologies just appear like bolts of lighting

Someone finds a technology that no one noticed before

Someone comes up with new concept or new business model and “wallah”

Some other form of “unexpected event” occurs

The Reality (1)

Literally thousands of new technologies

are being developed right now

◦ the concepts have been understood for years

◦ champions exist for every one of them

◦ these champions believe these new

technologies will become economically feasible

Some of these technologies are

experiencing more rapid improvements

than others

The Reality (2)

Which ones are experiencing rapid

improvements?

Of course, unexpected events do

occur…but we want to be the ones who

cause these unexpected events

In the end, everything is about

probability, what are the most probable

scenarios?

Analyzing the Timing of New Technologies

If components drive the improvements in performance or cost of new technology, ◦ We must understand the components and their rates of improvement

If we had understood the importance of ICs, even in the 1970s and 1980s, ◦ We would not have been surprised by success of personal computers, mobile phones, and similar technologies

Analyzing the Timing of New Technologies (2)

For technologies that benefit from finding materials that better exploit physical phenomena, ◦ we can use the rates of improvement to better understand the expected changes in performance, cost, and thus economic feasibility over time

◦ if many new classes of materials have been found, this increases the chances that some form of this technology will become economically feasible


For technologies that benefit from changes in scale ◦ we can use actual rates of improvement and role of supporting components in these changes in scale to better understand expected changes in performance, cost, and thus economic feasibility over time

◦ For new technologies for which little data is available, we can use data from similar systems to analyze the expected benefits from changes in scale


Finally, demand does impact on costs and prices. Increases in demand can ◦ reduce the development costs per unit ◦ enable larger volumes and thus larger production equipment

But, manual assembly and some production “equipment” benefit little from increases in scale ◦ e.g., assemblers of iPhones!

And increases in demand are the last factor and often the least important






The Myth: The Market Doesn’t Work Markets are very short-sighted New technologies will not be developed

unless there is strong government intervention

Governments must target new technologies in response to specific problems ◦ State the problems ◦ List potential solutions ◦ Fund and develop the potential solutions

The Reality

Markets work reasonably well Markets have replaced hierarchies (i.e.,

vertical disintegration) to a large extent even in R&D (i.e., open innovation)

Firms develop systems, components, and materials for those components even when ◦ the system for the components and materials are not clear

◦ and thus potential applications and demand for them are not clear

The Weakness of Markets

They under invest in R&D, particularly research

They under invest because ◦ there are large uncertainties

◦ in addition to uncertainties, information slips out and thus firms can’t appropriate all the benefits from research

What should governments do?

Government policies should support research,

perhaps much stronger than they currently do

These policies should reflect how technology

change occurs and not fall victim to the many

“myths”

The worst policies involve targeting new

technologies in response to specific problems as

if one is the “emperor of the universe”

◦ State the problems

◦ List potential solutions

◦ Fund and develop them

What should governments do? (2)

Government policies should support

technologies that

◦ are experiencing rapid improvements or

◦ have the potential for rapid improvements

Our research helps identify these technologies

◦ Ones that benefit from creating materials to…..

if many new classes of materials are being found,

improvements will probably follow

◦ Ones that benefit from reductions in scale

these technologies experience rapid rates of improvement

Final Words

Myths about technology change reduce our ability to develop new technology

Overcoming these myths can help us ◦ implement better strategies and policies

◦ more effectively find technologies that are or will become economically feasible

Business

Myths about technological change