How tech is changing business

kw: book reviews, nonfiction, business, technology, artificial intelligence, trends

My, my, what a long time it took me to work my way through this book! It goes to show that I still have a poor mind for business. During the latter half of my career in IT, the managers and even some supervisors would speak of the "business reasons" for doing one thing or another. One day I asked a manager named Carol, "What is a 'business reason'?" She replied, "It's something people are willing to pay for." The thought had never entered my head. I have always done things for reasons such as "it is interesting", "it will make this or that task easier", "it does things in a more excellent way" and so forth. Getting paid was nice, but it wasn't my focus. When I heard a new company president speak of having a "passion for profits", I sent him an e-mail explaining how I had always had a passion for excellence, and that profits seemed always to follow. His response was so disturbing, revealing such abysmal blindness to everything I find meaningful, that I immediately sought work in a different company among the Dupont family of companies, and luckily found one within a few months.

I am not sure what I expected once I saw the cover of Machine Platform Crowd: Harnessing Our Digital Future by Andrew McAfee and Erik Brynjolfsson. Something more techie than what it delivered, certainly. But the authors' application of technological trends to present and future business was sufficiently appealing that I read it all.

The three words that begin the title emphasize the subjects of the book, which is a follow-on to their book The Second Machine Age. These words outline three dichotomous trends that are driving businesses:

• Mind and Machine
• Product and Platform
• Core and Crowd

The trends are toward the right, and it is uncertain how far each will proceed. I debated with myself, whether to use "versus" rather than "and". But these pairs are not truly at odds; rather they are synergistic and supplementary to each other. For example, I built much of my career as a scientific programmer and systems analyst on discerning the appropriate tasks for the Machine to do, so as to free up people's Mind to do the things that we do better. From the beginning of the Computer Century (now about 70 years along), computational machinery has been called "mechanical brains", and the term "artificial intelligence" began to be applied even before ENIAC's tubes first lit up.

We now have pocket phones and nearly-affordable wristwatches that are millions of times as computationally powerful as ENIAC (this article includes notes on its speed of computation). But only within the past decade have "AI applications" begun to carry out tasks that are still – usually – done better by people and many animals. Many Sci-Fi stories bring us ideas of giant computers somehow becoming conscious more-or-less by accident (e.g., "Colossus" and "The Moon is a Harsh Mistress"). There is a reason for that. Nobody yet has the slightest idea how to define consciousness in any unambiguous way, and therefore, no idea how to write appropriate code to "do consciousness". To repeat myself, I define "genuine artificial intelligence" thus:

That a mechanism, electronic or electromechanical, carries out its own investigation, does its own research, and obtains a patent or at the very least has its patent application accepted by the U.S. Patent Office.

For the time being, the next generation or two at least, there will remain numerous "real world" tasks that minds will perform better than machines. The authors contend that nearly any repetitive task, including many now deemed "too creative" for a machine to carry out, will over time become the province of machine work, and that humans will be squeezed out. Will the day arrive when humans are no longer permitted to pilot an automobile? Cook their own meals?

The discussion of Product and Platform was harder for me to follow. Having a viable Product is the essence of a Business Reason for doing something. People pay for products, including those more squishy "products" we call "services." For example, technically, nursing care is a "service", but in the context of business, it is a product, delivered as a series of "service tasks" by a skilled person on behalf of another. Where does that fit into the notion of a "platform"? I think I understand that a platform packages products and services to make them easier for a producer to deliver and for a consumer to order and obtain. Will there one day be a platform like Uber for nursing care? I am almost afraid to look; it may already be out there. But there is still the need for the nurse-person (one day, a nurse-machine?) to physically do something to or for the person receiving nursing care.

Then, Core and Crowd. Hmm. I look on this as an expansion of Mind and Machine, where the "machine" has become a human-machine synergy we call the Crowd. I love the Citizen Science efforts out there, 73 of which (to date) are available under the Zooniverse umbrella. I have participated in about a dozen of them, and am most recently active in three that are of most current interest to me. A few years ago I classified more than 6,000 galaxies in one of the early Zooniverse projects. The machine part is the image delivery and questionnaire system. I and thousands of others (many minds) do the crowd part. The designers build in lots of redundancy, so as to spot errors and the occasional troll. The key to such projects is good planning and curation.

The authors focus on more business-oriented crowd projects. Their aim is to show that many untutored folks find innovative ways to solve problems that the "experts" would never think of. Very frequently the synergy of various "out of discipline" methods come together to do something ten or 100 times as well as the best that the "experts" had produced.

This principle comes home for me. Although I long aspired to be a scientist, because I was someone who nearly always wrote software for other scientists I had little occasion to publish; I wrote stuff to support work that other scientists published about. But the key paper of mine that made it into a peer-reviewed journal (Computers and the Geosciences) applied some sideways thinking to the numerical analysis of stiff differential equations used to simulate complex chemical reaction networks. I mixed principles used by astronomers in orbital mechanics with methods devised originally by civil engineers. In my dissertation, I used, and described, another numerical method that applied descending reciprocals to Runge-Kutta methods so that linear equations (linear in the "Diff Eq" sense) could be solved to any order desired. It was just a little part of my research, but crucial for certain computations that were otherwise too lengthy to carry out on the mainframes of the late 1970's.

So, I have rambled a lot into technical areas, mainly to cover up my difficulties "getting" the business focus of the book. It is written as a self-help text, with summaries and guiding questions following each chapter. It is written for business managers and executives. It is well enough written to hold my interest, even where I was in over my head.

Not to end on a downer, but I must quibble: on page 271 it is stated that the "amino acids" are strings of the genetic bases A, C, G and T. Those who know how wrong this is, just take comfort in "the old college try" that McAfee and Brynjolfsson gave it, when they were even more out of their depth than I am in their realm of expertise. (Hint to others: ACGT make genes, which are translated into proteins, composed of amino acids that do NOT include ACGT. That is why it is called translation.)