Smart Dust: Specialized Computers Fabricated to Be Smaller Than a Single Grain of Rice


“Sabotage #4: Mixing Noodles with Rice”, Image by Stefan

Back in 1977, Steve Martin put out a live album of his stand-up comedy performances called Let’s Get Small. Included was one of his signature routines called Well, Excuse Me. It went on to sell more than a millions copies. Much of it was laugh-out-loud hilarious.

Now, 48 years later, when the Internet of Things (IoT) has becoming a burgeoning global phenomenon, some very imaginative people have taken this notion (in name only), and  in a way that could have never been foreseen in any manner back then.

Certainly no excuses needed here. Rather, let’s have a look at this exciting new development.

Researchers at the University of Michigan, led by Professor David Blaauw, have recently fabricated a functional and autonomous computer measuring only 1 millimeter on each side. This device, dubbed the Michigan Micro Mote (M^3), was the subject of a most interesting article by Rex Sakamoto on entitled This Working Computer is Smaller Than a Grain of Rice, posted on April 6, 2015. I will summarize it, add some links and annotations, and pose a few questions. (The CNET article also contains an embedded video of a very informative recent report about this project on CBS News.)

This team’s work has been ongoing for more than ten years. With regards to the IoT, they believe that all of devices connected to it will require more “intelligence” and networking capabilities integrated into them whereby the M^3 could be the means to accomplish this.

The M^3’s current capabilities are photography and as temperature and pressure sensors. The researchers are now exploring a range of potential applications “ranging from medical to industrial” including:

  • Medical: How it can be “injected into the body” to take such temperature and pressure readings, as well as an electrocardiogram (EKG).
  • Energy: Assessing whether an existing oil well still contains any extractable reserves.
  • Consumer Goods: Attaching M^3s to everyday items such as keys and wallets to insure they are never lost inside or outside of the home.
  • Other potential apps on the project’s website include a platform containing “low-resolution imager, signal processing and memory, temperature sensor, on-board CMOS timer, wireless communication, battery, and solar energy harvesting that are all packaged in a 1mm3 volume through low-cost die stacking and encapsulation.”

In order to program and power up the M^3, the researchers created a means to accomplish this by using “strobing light at high frequency”. In turn, the M^3’s output is transmitted to an external computer by “conventional radio frequencies”.

The team’s current efforts involve reducing the size of the M^3 even further to a point where it may become the basis for a form of “smart dust“.

Just a few days ago in the April 10, 2015 Subway Fold post entitled The Next Wave in High Tech Material Science about metamaterials that can bend sound, light, radar and seismic waves, I speculated about some other potential applications for this emerging technology. With the M^3 so similar insofar as its originality and  potential to generate a myriad of applications not even considered yet,  my questions include:

  • Are there apps where the M^3 and metamaterials can be combined? What about in optical networks where metamaterials are using in the production of fiber cables where the M^3s could be used as sensors?
  • Would the M^3 make a via sensor for transportation infrastructure (roads, bridges, rails and so on), as well as the bodies, engines and electronics in cars, planes and trains? How about embedding them into buildings for additional safety technologies?
  • What safety and privacy protocols and policies will need to be developed and by whom? How can they be enforced?

Artificial Intelligence Apps for Business are Approaching a Tipping Point


“Algorithmic Contaminations”, Image by Derek Gavey

There have been many points during the long decades of the development of business applications using artificial intelligence (AI) when it appeared that The Rubicon was about to be crossed. That is, this technology often seemed to be right on the verge of going mainstream in global commerce. Yet it has still to achieve a pervasive critical mass despite the vast resources and best intentions behind it.

Today, with the advent of big data and analytics and their many manifestations¹ spreading across a wide spectrum of industries, AI is now closer than ever to reaching such a tipping point. Consultant, researcher and writer Brad Power makes a timely and very persuasive case for this in a highly insightful and informative article entitled Artificial Intelligence Is Almost Ready for Business, posted on the Harvard Business Review site on March 19, 2015. I will summarize some of the key points, add some links and annotations, and pose a few questions.

Mr. Power sees AI being brought to this threshold by the convergence of rapidly increasing tech sophistication, “smarter analytics engines, and the surge in data”. Further adding to this mix is the incursion and growth of the Internet of Things (Iot), better means to analyze “unstructured” data, and the extensive categorization and tagging of data. Furthermore,  there is the dynamic development and application of smarter algorithms to  discern complex patterns in data and to generate increasingly accurate predictive models.

So, too, does machine learning² play a highly significant role in AI applications. It can be used to generate “thousands of models a week”. For example, a model premised upon machine learning can be used to select which ads should be placed on what websites within milliseconds in order to achieve the greatest effectiveness in reaching an intended audience. DataXu is one of the model-generating firms in this space.

Tom Davenport, a professor at Babson College and an analytics expert³, was one of the experts interviewed by Power for this article. To paraphrase part of his quote, he believes that AI and machine learning would be useful adjuncts to the human analysts (often referred to as “quants”4). Such living experts can far better understand what goes into and comes out of a model than a machine learning app alone. In turn, these people can persuade business managers to apply such “analytical insights” to actual business processes.

AI can also now produce greater competitive efficiencies by closing the time gap between analyzing vast troves of data at high speeds and decision-making on how to apply the results.

IBM, one of the leading integrators of AI, has recently invested $1B in the creation of their Watson Group, dedicated to exploring and leveraging commercial applications for Watson technology. (X-ref to the December 1, 2014 Subway Fold post entitled Possible Futures for Artificial Intelligence in Law Practice for a previous mention and links concerning Watson.) This AI technology is currently finding significant applications in:

  • Health Care: Due to Watson’s ability to process large, complex and dynamic quantities of text-based data, in turn, it can “generate and evaluate hypotheses”. With specialized training, these systems can then make recommendation about treating particular patients. A number of elite medical teaching institutions in the US are currently engaging with IBM to deploy Watson to “better understand patients’ diseases” and recommend treatments.
  • Finance: IBM is presently working with 45 companies on app including “digital virtual agents” to work with their clients in a more “personalized way”; a “wealth advisor” for financial planning5; and on “risk and compliance management”. For example, USAA provides financial services to active members of the military services and to their veterans. Watson is being used to provide a range of financial support functions to soldiers as they move to civilian status.
  • Startups: The company has designated $100 million for introducing Watson into startups. An example is WayBlazer which, according to its home page, is “an intelligence search discovery system” to assist travelers throughout all aspects of their trips. This online service is designed to be an easy-to-use series of tools to provide personalized answers and support for all sort of journeys. At the very bottom of their home page on the left-hand side are the words “Powered by IBM Watson”.

To get a sense of the trends and future of AI in business, Power spoke with the following venture capitalists who are knowledgeable about commercial AI systems:

  • Mark Gorenberg, Managing Director at Zetta Venture Partners which invests in big data and analytics startups, believes that AI is an “embedded technology”. It is akin to adding “a brain”  – – in the form of cognitive computing – – to an application through the use of machine learning.
  • Promod Haque, senior managing partner at Norwest Venture Partners, believes that when systems can draw correlations and construct models on their own, and thus labor is reduced and better speed is achieved. As a result, a system such as Watson can be used to automate analytics.
  • Manoj Saxena, a venture capitalists (formerly with IBM), sees analytics migrating to the “cognitive cloud”, a virtual place where vast amounts of data from various sources will be processed in such a manner to “deliver real-time analytics and learning”. In effect, this will promote smoother integration of data with analytics, something that still remains challenging. He is an investor in a startup called Cognitive Scale working in this space.

My own questions (not derived through machine learning), are as follows:

  • Just as Watson has begun to take root in the medical profession as described above, will it likewise begin to propagate across the legal profession? For a fascinating analysis as a starting point, I highly recommend 10 Predictions About How IBM’s Watson Will Impact the Legal Profession, by Paul Lippe and Daniel Katz, posted on the ABA Journal website on October 4, 2014. I wonder whether the installation of Watson in law offices take on other manifestations that cannot even be foreseen until the systems are fully integrated and running? Might the Law of Unintended Consequences also come into play and produce some negative results?
  • What other professions, industries and services might also be receptive to the introduction of AI apps that have not even considered it yet?
  • Does the implementation of AI always produce reductions in jobs or is this just a misconception? Are there instances where it could increase the number of jobs in a business? What might be some of the new types of jobs that could result? How about AI Facilitator, AI Change Manager, AI Instructor, AI Project Manager, AI Fun Specialist, Chief AI Officer,  or perhaps AI Intrapreneur?


1.  There are 27 Subway Fold posts in the category of Big Data and Analytics.

2.  See the Subway Fold posts on December 12, 2014 entitled Three New Perspectives on Whether Artificial Intelligence Threatens or Benefits the World and then another on December 10, 2014 entitled  Is Big Data Calling and Calculating the Tune in Today’s Global Music Market? for specific examples of machine learning.

3.  I had the great privilege of reading one of Mr. Davenport’s very insightful and enlightening books entitled Competing on Analytics: The New Science of Winning (Harvard Business Review Press, 2007), when it was first published. I learned a great deal from it and this book was responsible for my initial interest in the applications of analytics in commerce. Although big data and analytics have grown exponentially since its publication, I still highly recommend this book for its clarity, usefulness and enthusiasm for this field.

4.  For a terrific and highly engaging study of the work and influence of these analysts, I also recommend reading The Quants: How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It (Crown Business, 2011), by Scott Patterson.

5.  There was a most interesting side-by-side comparison of human versus automated financial advisors entitled Robo-Advisors Vs. Financial Advisors: Which Is Better For Your Money? by Libby Kane, posted on on July 21, 2014.