Published in the Huffington Post.
"It is technology, more than science, which ignites growth. Phases of strong growth come years after scientific discoveries.”
Brian Arthur  defines science as the understanding of physical phenomena technology as the use of physical phenomena to implement a precise function. For example, science teaches us that some materials can be both conductive and insulating. The resulting technology uses this particular property to create a transistor (which enables to control one current for another current) and then a microprocessor (which is able to carry out logical operations).
It is technology, not science, which ignites growth. Phases of high growth come years after scientific discoveries, and only occur when associated technologies become affordable enough to be used by a majority of people. The scientific foundations of the Internet date back to the 1970s, but the effect on growth was only apparent in the late 1990s. What stimulated this growth was HTML technology, which enabled the creation of websites easily and quickly.
Several decades passed between the first prototypes of steam engines (Papin’s in 1679) and the moment their use became widespread, resulting in both substantial productivity gains and - already - concerns about unemployment. The steam train was invented in 1804 in the United Kingdom. Yet, it took twenty more years to see in France the first line linking Saint-Etienne to Andrézieux, and it took France close to 50 years to go from a steam machine capacity of 14,000 horsepower in 1833 to 500,000 in 1880.
Combining existing technologies. Sometimes technology evolves through breakthroughs, when a new physical phenomenon is mastered - for example, to produce new and more efficient way to store electricity. But most of the times it evolves through the combination of existing technologies. For example, a hydroelectric power station is composed of a tank, turbines, electric generators and electrical equipment.
Two figures play a key role in technological progress. First, the researchers or the research engineers who discover or master how physical phenomena can be used to create new technological bricks. Then, the inventors who assemble existing technologies to design a new innovative solution - from the engineers who improve the efficiency of an electric motor to the "makers" who build in their garage a connected device by assembling components bought on the Internet and a battery collected from a broken toy.
Most of these inventors have limited resources and must use components available to the public that they assemble in their garage - that is how Apple or HP started decades ago. Therefore, growth is greatly accelerated whenever "technology bricks" are developed, allowing new technologies to be easily assembled to meet needs. Those bricks are the technologies, the infrastructures, the languages or the development environments that make complex technologies accessible to everyone- just as HTML did for the Internet, the microprocessor for microcomputers, or the C programming language for Computer development. With those bricks, using advanced technology becomes as simple as assembling the bricks of a Lego game.
A whole generation could master the fundamentals of computing with the first microcomputers in the 1980s, before the increasing complexity (and cost) of systems reduced considerably the number of people capable of developing computer programs. Cloud computing platforms have the potential to democratize once again the development of applications - whether for connected consumer devices (like Pebble’s smartwatch app platform, recently bought by Fitbits) or for industrial processes (like GE’s Predix digital industrial cloud platform).
The industrial Internet Boom. More generally, access to numerous technological bricks has been greatly simplified. It takes only a soldering iron and a few dollars to make an object “connected” and “smart”. Dedicated networks such as Sigfox enable communication for $1 per device and per year. Crowd funding allows anybody to find the resources to scale up a first prototype to a global success.
Experts agree on the potential of industrial digital in the future: if it is still currently behind on the mainstream Internet, it would eventually exceed $8 trillion, a size comparable to the one of the consumer Internet. Moreover, the impact of the industrial Internet will be seen in the growth statistics, unlike its consumer counterpart: when an application saves us free time, the GDP figures do not change, whereas when a factory saves working hours, the measured productivity increases. Large potential, direct impact on growth and maturity of “technological bricks": all the ingredients are there for an unprecedented wave of growth whose premises, although still modest, are already visible.
Governments often focus their efforts more on science than on technology. Their centralized structures are usually more suited to large-scale projects and public research than designing the regulations supporting technology dissemination or inspiring and incubating tens of thousands of individual innovators. Though, there are differences across counties – for example those who witnessed the development of personal computers in the 1980s in Europe noticed the advantage of Northern countries, where high school students had more time for experimentation.
Are they real differences across countries? To estimate how countries rank in terms of "innovation composition" compared to other countries, we analyzed the volume of Internet search by country of origin for components used by "makers". For "ESP8266"  (widely lauded for its low cost in the realization of intelligent connected objects), the Netherlands lead (18% of searches), followed by Germany (16% distributed in 4 regions - Bavaria, Bad Württemberg, Berlin, North Rhine-Westphalia), Russia (13%), Poland (13%), Spain (9%), Italy (8%) and India (7%). The top 10 for search on "Arduino"  includes countries like Sri Lanka, El Salvador or Estonia.
In early November 2015, the "Tensorflow" deep learning library was transferred to the public domain by Google. It's a very promising move: everyone can from now on use this technology that is behind many of Google's innovative services. But who has heard about this news? We will know that governments are ready to seize opportunities that arise, the day they will be as attentive to opportunities this type of decision offers, as we are to the defensive discussions on topics such as “Uberization” or the supposed risks of artificial intelligence...
Thanks to Joyce Bessis for her help on the english version of this article.
 Arthur, W. Brian. The Nature of Technology: What It Is and How It Evolves. Reprint edition. Free Press, 2009.
 The ESP8266 is a low-cost Wi-Fi system-on-chip with full TCP/IP stack used by hobbyist to build connected devices.
 Arduino is an open source project designing microcontroller kits used by students or hobbyist.