Joel Mokyr Departments of Economics and History Northwestern University Berglas School of Economics Tel Aviv University OECD, Paris Sept.2014 1
Innovation pessimism Has the ideas machine broken down?
OECD, Paris Sept.2014 2
The new wave of techno-pessimism: The new technopessimist interpretation says that the low- hanging fruits of invention have been picked.
Future inventions will not have nearly as radical an effect as before.
For that reason, innovation will not be powerful enough to counter other economic “headwinds” and annual GDP growth will slow down to a trickle.
Many feel disappointed. Peter Thiel (of Paypal fame) has famously remarked “we wanted flying cars, instead we got 140 characters.” OECD, Paris Sept.2014 3
Is the world running out of ideas?
Perhaps the low-hanging fruits that have changed our lives have been picked: running water, chlorination, electricity, air- conditioning, antibiotics etc?
But science and technology’s main function in history is to make taller and taller ladders to get to the higher-hanging fruits (and to plant new and maybe improved trees).
Moreover, the old trees will keep sprouting new fruits, if only we give them proper care. OECD, Paris Sept.2014 4
What can an economic historian bring to this discussion?
From a purely technological point of view, if the patterns of the past hold (a big if), there is good reason to expect the rate of technological change to accelerate over the next decades, although it would be foolhardy to be more specific than that.
If so, the “tailwind” from technology is likely to be so powerful that it will overcome all “headwinds” from other factors.
OECD, Paris Sept.2014 5
Can I be sure? No. Far too many variables.
What will it do to GDP growth, TFP, measured L-productivity?
Can we learn anything from history here? Not so obvious.
The best I can do is point to three independent variables that I think made a difference in the past, and then plug in what I think their values might become in the future to estimate what technological progress might become in the future. But there are lots of omitted variables and the coefficients may be time-variant. OECD, Paris Sept.2014 6
Three factors that mattered in the past: •Artificial Revelation
•Good institutional set-up for intellectual innovation.
OECD, Paris Sept.2014 7
Artificial Revelation • The idea is basically that technological progress (“prescriptive knowledge”) makes scientific advances possible no less than the reverse.
• What drives scientific advances at any time? They are driven by many factors, but one of the most important is the tools and instruments available to scientists.
• Our senses and brains are too limited to observe and measure much of nature, which operates at scales, frequencies and bandwidths that we cannot observe. Many natural phenomena are also “too complex to compute” by hand.
OECD, Paris Sept.2014 8
This was certainly true for the scientific revolution in the 17th Century, which was in large part driven by new instruments and tools.
OECD, Paris Sept.2014 9
The best-known examples are of course “the great trio” of the telescope, the microscope, and the barometer, all developed during the early seventeenth-century, and that played a big role in the Scientific Revolution. But there are many others. Let me give you a few lesser-known examples from the era before and during the Industrial Revolution to drive the point home.
OECD, Paris Sept.2014 10
Boyle’s famous air pump Robert Boyle’s famous air pump,
built in the late 1650’s, which
showed once and for all that contra Aristotle, nature did not
abhor a vacuum, and thus paved
the road for atmospheric (steam)
OECD, Paris Sept.2014 11
Volta’s “pile” (1800) Volta’s battery provided chemists with a new tool, electrolysis, pioneered by (among others) Humphry Davy. He and other chemists were able to isolate element after element, and fill in much of the detail in the maps whose rough contours had been sketched by Lavoisier and Dalton. 12
And in medicine:
Joseph J. Lister (father of the famous
surgeon), inventor of the achromatic microscope that minimized both
chromatic and spherical aberration.
This made it possible eventually for
Pasteur, Koch and others to demon-
strate that infectious diseases were
directly linked to identifiable microorganisms.
OECD, Paris Sept.2014 13
This is true a fortiori in our age.
Science’s toolkit has grown enormously in the past decades. This expansion cannot but lead to rapid applications, in fields that are at times obvious and immediate but often unexpected. Examples are easy to come by.
Start with telescopy, in honor of Galileo: OECD, Paris Sept.2014 14
OECD, Paris Sept.2014 15
Adaptive optics 1. These are two images of the planet Uranus, one using an ordinary telescope, the other one in which the blurring caused by atmospheric distortions are corrected through adaptive optics. 2. Adaptive optics technology sharpens images by changing the shape of telescope mirrors up to 1,000 times per second. 3. It is believed to have more potential than Hubble’s telescope (and is a lot less expensive). OECD, Paris Sept.2014 16
Another example how new technology helps science:
first developed in 1986 by Dr. Leroy Hood’s
laboratory at CalTech,
critical in sequencing
OECD, Paris Sept.2014 17
And perhaps the most revolutionary:
OECD, Paris Sept.2014 18
New techniques that may have been originally designed for manufacturing or agriculture are “spilling over” to scientific research.
The new science then “feeds back” into technology, at times with enormous power. OECD, Paris Sept.2014 19
Finally, of course, the computer It is hard to think of a single field of research that has not been transformed by computers.
The real question often seems to be: what did we ever do before it?
My interest here is not in what the digital revolution does for productivity directly, but rather indirectly through its effect on science.
OECD, Paris Sept.2014 20
Computers allow research hitherto impossible Multiscale Models of Complex Chemical Systems, which allows the solution of the complex equations that govern the properties of quantum chemistry. Turbulence: the English applied mathematician Horace Lamb sighed in 1932 that “I am an old man now, and when I die and go to heaven there are two matters on which I hope for enlightenment. One is quantum electrodynamics, and the other is the turbulent motion of fluids. And about the former I am rather optimistic.” Today we can start to approximate the Navier-Stokes equations with supercomputers.
OECD, Paris Sept.2014 21
Perhaps just as promising: material science • Materials have historically been at the heart of economic
civilization. Hence terms such as “iron age.” • Historically, progress here has been always the result of
tedious and inefficient “trial and error” or highly uncertain
serendipity (classic example: William Perkins discovery of
aniline purple in 1856 and Bessemer steel the same year). • We now can simulate in silico the quantum mechanics equations
that define the properties of materials using high-throughput
super-computers to experiment with materials having pre-
specified properties. • This does not mean we do not need to test new materials, but it vastly shortens the testing time (by comparison, lithium ion batteries invented by Sony in 1991 took 20 years to develop) and increases the options by orders of magnitude.
OECD, Paris Sept.2014 22 22
2. Access costs Here is a simple argument: What is the total social useful knowledge that an economy has access to? Answer: it is the union of all individual sets of useful knowledge. Corollary: some very important pieces of knowledge that “are known” to this society are only possessed by very few extraordinarily smart individuals. And hence, access by others who do not have this knowledge but need it is important. Such access is costly. OECD, Paris Sept.2014 23
Why is access so crucial to sustained technological dynamism? • Part of it of course is that more and more production depends directly on access to the best science available in material science, biochemistry, combustion, etc. • Even if science is not directly very useful in guiding an inventor, it is still true that “le hasard ne favorise que les esprits préparés.“ ( Fortune favors the prepared mind.) • But it is also true that technology advances by “ideas having sex” as one writer famously described it. So if you have one idea, you need access to “a partner”. • Finally, inventors have to know what is already known, so OECD, Paris Sept.2014 24 that they don’t reinvent more wheels than is unavoidable.
This does not matter as long as users who need this knowledge have access to it. • But access can be costly. What determines access costs?
• Among many factors, clearly the cost of storing codified information and searching through it figure highly.
• In the past, the most important advances in information- storage and search-engine technology were the invention of writing, paper and the printing press.
OECD, Paris Sept.2014 25
But codified knowledge needs to be organized if access is to be fast and cheap with high searchability.
The Age of Enlightenment that preceded the Industrial Revolution blazed new trails in access capability in technology and in science, both for codified and tacit knowledge.
OECD, Paris Sept.2014 26 26
The eighteenth century version of the search engine was the encyclopedia. Alphabetized encyclopedias and indices to technical books were the Googles of their age.
And indeed, the paradigmatic enlightenment document is Diderot’s Grande Encyclopédie.
OECD, Paris Sept.2014 27 27
Pinmaking essay in Diderot’s encyclopedia
OECD, Paris Sept.2014 28 28
There were many other new organizations that helped reduce access costs to best-practice knowledge in the Age of Enlightenment. For instance, clubs and scientific societies provided meeting places for people who searched for experts.
(See my Gifts of Athena: Historical Origins of the Knowledge Economy, 2002).
This helps explain many things about the Industrial Revolution (which was mostly about textiles and steam, not better medicine).
Many of the top inventors in the Industrial Revolution had access to the “best-practice” science of the day or (at times) were scientists themselves. OECD, Paris Sept.2014 29
What about today?
If ICT has done anything, it has reduced access costs.
We no longer deal with “data” ̶ we have “meta-data,” amazing quantities of information that can only be accessed with sophisticated searchware.
We can search for nanoscopic needles in haystacks the size of Montana.
This is certainly not without its drawbacks, as both spies and advertisers know more and more about us.
But it has enormous implications for further scientific research and technological advances. OECD, Paris Sept.2014 30
Anyone engaged in research can access vast banks of knowledge and data. Cloud technology is just getting started. We measure storage now not in petabytes but Zettabytes (a million petabytes) and Yottabytes (1000 Zettabytes)
(WHO makes up those terms? --- there is also “Brontobytes”).
As Matt Ridley has remarked, “The cross-fertilization of ideas between, say, Asia and Europe, that once took years, decades, or centuries, can now happen in minutes.”
Access costs have declined sharply for both codifiable and “tacit” knowledge. OECD, Paris Sept.2014 31 31
Public databases are a huge step forward in codified knowledge I will give you a few examples of modern databases that are important to medical research rather than management or clinical treatment, since that is my focus today. •Pubmed, an NIH-sponsored database, contains more than 23 million citations for biomedical literature from MEDLINE, life science journals, and online books. Many commercial databases such as ScienceDirect put out by Elsevier • Online Mendelian Inheritance in Man (OMIM), an online database that catalogues all the known diseases that have a genetic component, and links them to the relevant genes in the human genome and provides references for further research and tools for genomic analysis of a catalogued gene OECD, Paris Sept.2014 32
• Genbank maintained by NCBI (National Center for Biotechnology Information), contains all publicly available nucleotide sequences and their protein translations. As of July 2013, GenBank release 196.0 contains 152,599,230,112 bases. It doubles in size every 18 months… • Protein Data Bank (PDB) is a repository for the three- dimensional structural data of large biological molecules, such as proteins and nucleic acids. Contains as of 2013 over 94,000 proteins, nucleic acids, and complexes. • Stanford University HIV Drug-Resistant database contains a wealth of data on over 100,000 HIV patients, treated and untreated by such drugs as Protease Inhibitors.
OECD, Paris Sept.2014 33
All these databases are accessible free of charge, with no physical effort, at the click of a mouse.
Louis Pasteur never had it so good.
But what is true for medical science is true across the board, in material science, astrophysics, molecular plant genetics, and economic history.
The decline in access costs, whatever its other implications, must mean that technological progress is likely to accelerate OECD, Paris Sept.2014 34
What about institutions? In my new book A Culture of Growth (Princeton UP, 2015) I argue that the best incentive structure for the generation of useful knowledge is the reputation mechanisms involved in “open science” which emerged in early-modern Europe.
In some modified version, we still have that system! It has worked well (if not perfectly) in the past, and will continue to evolve. OECD, Paris Sept.2014 35
All the same, I am still worried All the technical conditions are ripe for progress to go on as before or faster. But bad institutions and politics can interfere. This can take many forms: • Outright resistance by entrenched interests. • Resistance by well-meaning ideologies suspicious of innovation. • Leading to: excess regulation • Leading to: lack of venture capital or entrepreneurship • Bad institutional set up of research funding (who pays? who allocates? who gets funded? who “picks winners”?). • New property forms and hence new forms of crime and insecurity. None of those seem all that acute in the US, though there are many worrisome signs. OECD, Paris Sept.2014 36
To sum up: •We are not like the late Roman Empire or Qing China, about to languish into an age of decline to be followed possibly by chaos and barbarism. •Technological progress is still remote from reaching a ceiling or even diminishing returns (and may never do so). •Economic growth, in an economically meaningful way (if not necessarily in a traditional NI accounting way) will continue, unless bad politics screws it up. •The Digital Age will be to the Analog Age what the iron age was to the stone age. •And we can’t even imagine what the Post-digital Age will look like. No more than Archimedes could imagine CERN. OECD, Paris Sept.2014 37 37
Thank you OECD, Paris Sept.2014 38
So where will technology be heading? We have already invented electric lights, internal combustion engines, a/c, and water chlorination, so these “easy” fruits have been picked. But can we predict where technology is heading?
We know to whom the “gift of prophesy” was given…
OECD, Paris Sept.2014 39
All the same: General areas in which it is reasonable to expect technological quantum leaps 1. Material Sciences (incl. nanoscience) 2. Energy (nuclear fusion or renewables?) 3. Advanced technology in medicine (personalized medicine) and life sciences incl agriculture, (GMO’s). [one small example: transgenic chestnut trees] 4. Robotics and automation OECD, Paris Sept.2014 40
5. Mass customization in manufacturing (3-d printing): Strati example! 6. Further advances in leisure technology (virtual reality). 7. Sharing and communications technology that allows higher and more efficient utilization of resources (telecommuting, Uber, etc.) 8. “Dumbing down” technology that makes very complex technology user friendly by frontloading the ingenuity.
OECD, Paris Sept.2014 41
OECD, Paris Sept.2014 42
Do we “need” all this new technology? YES! (If time, explain why).