Liberate powerful technologies from the few.  Let everyone gain access to technology and expand their personal opportunities. Applied technology enriches daily life, saves labor and reduces costs and risks. As technologists we know this from personal experience. To reach a wider sphere of customers we can design our business ecosystems to produce a range of products, prices, benefits and tradeoffs.  Differentiate our offerings, be flexible in our options and diversify whom we serve.

DIY Drones, Freescale Semiconductors, Harvard University 

In 2007 Chris Anderson, the editor of Wired Magazine, founded a Do It Yourself (DIY) community to make personal drones. There had long been a worldwide hobbyist community making and flying model airplanes — but these were not autonomous. They were tethered — either by actual physical lines or by radio control units.  What they lacked was the ability to fly by themselves free from tether. Anderson’s insight was that the control technology to enable reliable long-range drones was becoming available at a price many could afford — in the range of a few hundred dollars.  As he put it in an article last summer,

“Just as the 1970s saw the birth and rise of the personal computer, this decade will see the ascendance of the personal drone. We’re entering the Drone Age.”[i]

A business ecosystem is initiated by bringing together the fruits of science with the needs of society.  In less formal terms, entrepreneurs liberate technologies and make them widely available.  William Gibson’s famous statement that “the future is already here, it’s just not evenly distributed yet”[ii] is not just an observation.  It also poses implicit questions: “Why isn’t the future distributed yet?  What would happen if it was?”  And most exhilarating: “Do you think we might be able to do it — to let the future loose?”  Wearable computers, ingestible sensors, wireless communications from inside the body, personal drones — the real lesson of the past decade or so is that the public itself, including hobbyists, DIY “makers,” kids in school and out, will spread the word and distribute the future.

You don’t achieve this just by extending Moore’s law — by simply going smaller, denser and faster.[iii]  That would be like attributing all the blessings of modern transportation to the advancement of metallurgy or internal combustion.  Moore’s law is necessary but not sufficient.  The new business ecosystems have an explicit intent to let people communicate knowledge among themselves and to liberate useful technologies.  They have gained this in part from the social fruits of the open-source movement and from the social networking experience in general. This social experience is joined with something I consider an essential and often unnoticed fruit of the first-generation ecosystems: modular and flexible components that can be put together to perform sophisticated functions by (relatively) unsophisticated makers. It was the modularity of electronic components that let the two Steves — Jobs and Wozniak — build a “phone phreaking” device to trick Bell System computers into allowing free long distance calls. They sold their “blue box” in the dorms of UC Berkeley, and subsequently, building on this experience, created and sold the personal computer.

The new generation ecosystems are explicitly designed for freedom, choice and flexibility — for sensing and satisfying customer pull.  I often think we need a new “law of flexibility” to complement and build on Moore’s law.  The more that Moore’s law takes us deeper into the atomic level, the more theoretical flexibility we have to redesign and reconstruct our world.  The great physicist Richard Feynman pointed this out years ago in a lecture that is widely viewed as founding the nanosciences. In an after-dinner speech to the American Physical Society in 1957, Feynman said,

“At the atomic level, we have new kinds of forces and new kinds of possibilities, new kinds of effects. The problems of manufacture and reproduction of materials will be quite different. I am, as I said, inspired by the biological phenomena in which chemical forces are used in a repetitious fashion to produce all kinds of weird effects (one of which is the author).”[iv]

Great combinatorial diversity is born of having more pieces to combine. As the size of pieces approaches the atomic level, the possible combinations approach infinity.  The phenomenon holds at other scales.  For example, a useful discussion of this phenomenon at the ecosystem level is “Population diversity and ecosystem services,” a widely referenced paper by Gary Luck, Gretchen Daily and Paul Ehrlich of Stanford.[v] They attempt to measure “ecosystem services” — that is, the services provided by ecosystems to those within and beyond their borders. We can think of this as the analogue of services a business ecosystem provides in response to a variety of customer pulls.  The variety of services depends on the variety of organisms making up the ecosystem, and the ability of these organisms to join in unique and useful combinations.

The success of the smartphone is just one dimension of a much broader Cambrian-explosion of diverse products.  I consider this expansion of diversity every bit as fundamental as Moore’s law, even as it builds on that earlier law.  Maybe we should name it after whoever is best able to quantify it.  Gordon Moore’s original paper is a serious piece of scientific observation and insight.  Moore himself anticipated the issue of flexibility in his 1965 paper, and thought about it as a way to share the cost of achieving continuing advances in leading edge manufacturing.

“Clearly, we will be able to build such component-crammed equipment. Next, we ask under what circumstances we should do it. The total cost of making a particular system function must be minimized. To do so, we could amortize the engineering over several identical items, or evolve flexible techniques for the engineering of large functions so that no disproportionate expense need be borne by a particular array. Perhaps newly devised design automation procedures could translate from logic diagram to technological realization without any special engineering.

“It may prove to be more economical to build large systems out of smaller functions, which are separately packaged and interconnected. The availability of large functions, combined with functional design and construction, should allow the manufacturer of large systems to design and construct a considerable variety of equipment both rapidly and economically.”[vi]

Fast-forward five decades to today, and Moore’s vision is being realized. Consider Freescale Semiconductor, a global semiconductor company that sells inexpensive, power-efficient flexibility with more-than-good-enough speed.  Freescale manufactures and sells a dizzying range of chips packaged and ready to be interconnected, most of them available off-the-shelf through online distributors such as Mouser Electronics.  Mouser lists hundreds of Freescale products, most selling for far less than $100.[vii] One of these products soon to hit the market is a tiny microcontroller 1.9 x 2 mm that Mouser says is [viii]“with great potential for … portable consumer devices, remote sensing nodes, wearable devices and ingestible healthcare sensing.”

I spoke with Geoff Lees and Mario Centeno of Freescale.  Geoff leads the microcontrollers business and Mario is a senior executive in strategy.  They are based in Austin, Texas, home of the South by Southwest Music Festival as well as fabulous BBQ ribs.

I asked them how Freescale is able to offer such a wildly diverse range of chips, almost all at markedly low prices.  The answer:  Gordon Moore’s 1965 hope that “design automation procedures could translate from logic diagram to technological realization without any special engineering” has largely been accomplish by an ecosystem of independent electronic design automation companies and their partners.  Freescale is a sponsor, customer and beneficiary of this achievement, and manufactures chips at fabs around the world.

As Geoff explained,

“In the mid-2000s we began to see a trend toward using automated design software to pull together software models representing, say, one or more microprocessors plus a number of other functions, and then use the design software to ‘synthesize‘ the models into instructions to drive the manufacturing line in a fab, down to the molecular level.

“So EDA companies such as Cadence, Synopsys take a high-level language and synthesize it into actual real world physical logic.

“Automation in hardware design made small-run, specialized chips economically viable.  A newly expanded range of customers could be offered specialty chips at affordable prices.

“It’s clear that the hardware side of our industry has reached a level of efficiency that [allows us to] deliver flexible hardware solutions at the right price and the right power and the right profile for customers.  And that differentiation is now.”

Variety can now be achieved without exotic processes and unaffordable fabs.  Geoff is an admirer of what Intel accomplished in pursuit of a linear Moore’s law, and what their work taught the world.  On the other hand, he sees that most of what customers desire today can be achieved without exotic processes — including the tiny Internet-of-things devices presaged by his 2mm processor.

“Everything we are doing now is focused on lower power.  Extracting the best performance in power out of conventional silicon solutions, spread over many, many foundries, over many partners.  It’s really going to fuel these devices.”[ix]

In two years we will celebrate the 50th anniversary of Moore’s paper.  It is time we created a rigorous, well-publicized way to measure the advance of flexibility and design freedom so that we can celebrate our successes and motivate creativity.

In order to kick off the celebration, here are some notable contributions to flexibility and choice:

  • A spirit that encourages differentiation.  Human values, personal relationships and sustainable lives form the basis for professional partnerships.  These partnerships in turn encourage the best and most distinctive contributions by each partner to the whole.  This spirit extends to customers and encourages variety within the customer community, and helps customers guide ecosystem creativity.
  • Tools, components and programmability for flexibility.  In order to advance the frontier of differentiation, make use of advances in design automation software combined with modularity and programmability in key components:  Field Programmable Gate Arrays, Application Specific Integrated Circuits, multi-capability chips (e.g. six types of radios available and baked on the chip), and custom-designed and manufactured chips.
  • Rules designed as platforms for variety.  Adopt architectures and standards that enable differentiation rather than enforce uniformity.  Encourage community problem-solving that is self-reflective on this matter, so that the judgment of community members grows with the platforms themselves.
  • Common resources open to a variety of uses.  Make substantial contributions to the common base of technology on which the ecosystem lives. Examples include the Common Platform Initiative enabling IBM, Samsung and GlobalFoundries to make leading-edge manufacturing available through a foundry model.  At the software end of the spectrum, collaborative initiatives like the open-source Linaro initiative are developing shared flexibility-enhancing “middleware” to help bring new applications onto a range of hardware.
  • Appropriate technology diversity matching desired results.  Take advantage of the variation in manufacturing capability across the ecosystem.  Each time the leading edge in fabrication moves ahead, factories that only recently were seen as state-of-the-art are made available to a wider user base.  The equipment and the fab are paid for.  The people are experienced and the processes have been refined.  Implementation is straightforward and safe compared to the leading edge, design automation tools are mature and the yield of useable chips is likely to be comparatively high.

Taken together these contributions — and no doubt more I don’t know about — lower the cost and expertise barriers, enhance flexibility, increase choice up and down the value chain including for the end customer and application designer, and enable the ecosystem to expand into an ever larger array of markets.

They also give the lie to the widely held view that because leading-edge fabs are becoming more and more expensive, chip technology is becoming less and less accessible to small companies and to members of open ecosystems.  The opposite is true.  Because of concerted industry investments in democratization, open access and low barriers to entry, flexibility — or the law named after whoever is able to quantify this advancing frontier — continues to progress at a rate equal to and perhaps greater than Moore’s law.  The world needs the moon shots, the U.S. government-funded Defense Advanced Research Projects Administration (DARPA) investments, and the relentless Intel and IBM pursuit of the frontier.  But as these investments pay off, the knowledge they yield and the technology they leave in their wake are free for the populace to take up and distribute.  I conclude with Chris Anderson again:

“Today, all the sensors required to make a functioning autopilot have become radically smaller and radically cheaper. Gyroscopes, which measure rates of rotation; magnetometers, which function as digital compasses; pressure sensors, which measure atmospheric pressure to calculate altitude; accelerometers, to measure the force of gravity — all the capabilities of these technologies are now embedded in tiny chips that you can buy at RadioShack.  Indeed, some of the newest sensors combine three-axis accelerometers, gyros, and magnetometers (nine sensors in all), plus a temperature gauge and a processor, into one little package that costs about $17.[x]

This is choice, flexibility and access. RadioShack is the exemplar of Feynman’s law.

Postscript:

The accessibility revolution is not limited to semiconductors, and appears wherever the manipulation of atoms and molecules is valuable.  J.D. Deng manages Harvard’s Laboratory of Integrated Science and Engineering, a facility open to students with several clean rooms and extensive but small-scale Nano fabrication and Nano imaging capabilities.  Mostly the work is focused on Nano-Bio. Last fall I took an evening extension course with J.D.  We made sensors with carbon nanotubes that measured about a nanometer across and were “instrumented” (impregnated) with gold Nano particles. You can buy these instrumented nanotubes at lab supply houses.  We attached the nanotubes to silicon dioxide wafers by weak intermolecular forces (van der Waals forces). We checked our work with atomic force microscopes capable of imaging a molecule.  The microscopes fit on a desktop and cost a few thousand dollars.

J.D.’s course is complemented by one on microfluidics, labs-on-a-chip and soft lithography.  Soft lithography is applied to make micro channels and small compartments through which fluids such as blood plasma can be drawn.  There are many lab-on-a-chip applications. For example, small “fingers” made sensitive to particular molecules can be exposed to fluids.  When triggered, some fingers give off small electric charges, others change color or fluoresce.

The astounding thing is how easy these labs-on-a-chip are to make. Patterns are etched into silicon wafers by computer-controlled electron beams. PDMS — common silicone caulk — is poured over the etched wafer surface and fills the negative spaces.  The PDMS is allowed to harden and then carefully peeled off.  Intricate positive structures of soft, bio-stable PDMS remain on the underside.  These tiny channels, gates, mixing rooms and wells are excellent structures for micromanipulation of fluid.  Layers can be joined to make more complex fluid processing systems — labs-on-a-chip.  The structures are simple but can have features down to about six nanometers—about half the Moore’s law threshold in semiconductors today.[xi]

Student projects typically mix or match techniques to make small systems.  In our class, I planned a lab-on-a-chip using a mix of detection techniques to look in parallel for 20 micro molecules in blood. These particular molecules have been recently discovered to signal and perhaps cause — by cascade effects — positive real-time effects of aerobic exercise on the heart and pancreas. Speaking to J.D. one afternoon, I commented on the extraordinary flexibility of the bio Nano techniques, and their ability to be mixed and matched into applications.  “Yes,” said J.D., “it is like there are all of these components available on a table in front of us, and we just draw lines through a handful of them and out to what we want to do.”[xii]

That is a perfect homolog for the strands and fabrics linking functions on a system-on-a-chip, chips in a system, people and teams in a value chain, and companies among each other in an ecosystem.


NOTES

[i] “How I Accidentally Kickstarted the Domestic Drone Boom,” Chris Anderson, Wired Magazine, June 22, 2012 http://www.wired.com/dangerroom/2012/06/ff_drones/all/

[ii] “Moore’s law to roll on for another decade,” Gordon Moore speaking to the International Solid-States Circuits Conference, IEEE, February 10, 2003, CNET, February 10, 2003 http://news.cnet.com/2100-1001-984051.html

[iii] William Gibson, Wikiquote, Fresh Air, NPR (31 August 1993) {unverified}, he repeated it, prefacing it with “As I’ve said many times…” in “The Science in Science Fiction” on Talk of the Nation, NPR (30 November 1999, Timecode 11:55)  http://en.wikiquote.org/wiki/William_Gibson

[iv] “There’s lots of room at the bottom,” Richard Feynman, Caltech Engineering and Science, Volume 23:5, February 1960, pp 22-36. http://www.zyvex.com/nanotech/feynman.html

[v] “Population diversity and ecosystem services,” Gary Luck, Gretchen Daily, Paul Ehrlich, Trends in Ecology and Evolution, Vol. 18, No. 7, July 2003 http://max2.ese.u-psud.fr/epc/conservation/PDFs/luck.pdf

[vi] “Cramming more components onto integrated circuits: With unit cost falling as the number of components per circuit rises, by 1975 economics may dictate squeezing as many as 65,000 components on a single silicon chip,” Gordon E. Moore, The experts look ahead (editorial section), Electronics, Volume 38, Number 8, April 19, 1965 http://download.intel.com/museum/Moores_Law/Articles-Press_Releases/Gordon_Moore_1965_Article.pdf

[vii] Freescale Products, Mouser Electronics Online Catalog, April 1013 http://www.mouser.com/freescalesemiconductor/

[ix] Geoff Lees and Mario Centeno, Freescale Semiconductor, Austin, Texas, personal communication, March 2013

[x] “How I Accidentally Kickstarted the Domestic Drone Boom,” Chris Anderson, Wired Magazine, June 22, 2012 http://www.wired.com/dangerroom/2012/06/ff_drones/all/

[xi] “Recent progress in soft lithography.” J.A. Rogers, R.G. Nuzzo, February 2005, Materials Today 8 (2): 50–56. doi:10.1016/S1369-7021(05)00702-9.

[xii] J.D. Deng, Laboratory of Integrated Science and Engineering, Harvard University, Cambridge, Massachusetts, personal communication, February 2013