We’re all familiar with Moore’s Law: the observation that the number of transistors in an integrated circuit doubles every two years or so. We’re also familiar with the plethora of corollaries to the law, pertaining to everything from network speed to hard drive capacity to the number of pixels in our digital cameras. It seems that the natural behavior for technology advancement follows an exponential growth curve.
However, not all innovation follows such a curve. Organizational and process improvements, in particular, seem to proceed at a glacial pace. Management fads come and go, and they don’t even seem to be getting much better, let alone better at a faster rate.
In development shops, the Agile Manifesto is now fifteen years old or so, and yet organizations still struggle with it. Where’s Moore’s Law when you need it, eh?
The problem today’s organizations face is that digital transformation relies both upon technology advancement as well as organizational innovation – and yet, it seems our inability to accelerate the latter may be putting the whole kit and caboodle in jeopardy. If only we could bottle up the secret to Moore’s Law and pass the bottle around.
It’s true that Moore’s law is about fifty years old, more than a lifetime for most of the people who have been riding the exponential express for their entire careers. Yet, while such exponential behavior is now all around us, one question strangely appears to be as yet unanswered: why.
Kurzweil’s Key
Why? What is it about technology innovation that naturally proceeds along an exponential path? Note first of all that calling Moore’s Law a law is a misnomer; in reality it is simply an observation. An observation so ingrained now in the planning of technology innovators that perhaps it has become self-fulfilling – but such a sui generis explanation is certainly no explanation at all.
I had the great fortune of asking Ray Kurzweil this very question at a recent conference. Kurzweil, of course, is a noted inventor, futurist, and deep thinker who has been writing about the exponential pattern of technology innovation for many years now, what he calls the law of accelerating returns.
He has extended his argument well into the future, leading him to theorize about computers that get so smart they completely swamp our human intelligence, or perhaps to technology so advanced we’re able to augment our brains with it.
Fascinating stuff, to be sure, especially to us lifelong science fiction fans. And while some of his prognostications sound outlandish, there’s no arguing with the exponential behavior of technology innovation – and once you get your head around the wheat on the chessboard lessons of such growth, one would be hard pressed not to allow Kurzweil a good measure of leeway in his predictions.
Our focus here at Intellyx, however, is more short term. Planning for the next year or two is difficult enough without worrying about whether we’ll be able to back up our consciousnesses to the cloud before we die of old age. Let us therefore return to the question of the day: why?
Kurzweil’s answer is essentially to point to the feedback loop intrinsic to all exponential innovation. Essentially, each generation leverages the one before. Intel 386 processors enabled the brilliant minds at Intel to create the 486, and then the Pentium, and so on. The same with memory capacity and Ethernet speeds and so on ad infinitum.
Perhaps. Yes, feedback loops do seem to be part of the answer, but they don’t give me the warm fuzzies I would require in order to believe they are the whole story.
What is it about human innovation that requires such steadfast adherence to each exponential curve? Furthermore, what happens when exponential curves give out? After all, transistors can only get so small. What happens when Moore’s Law drives right past the single-atom transistor and keeps on going?
Moore’s Law before Moore
Kurzweil, as you might expect, has thought long and hard about this question. In fact, he extends Moore’s Law into the past, well before the 1971 invention of the integrated circuit, as shown in the following diagram.
Moore’s Law Extended Back to 1900 (Source: Kurzweil AI)
As electromechanical calculating machines gave way to relay-based computers and then vacuum tubes and eventually transistors, Kurzweil’s favored metric of calculations per second per dollar stuck closely to an exponential curve (represented as a straight line on this logarithmic scale) – even curving upward a bit.
In other words, just as a particular computing approach (what Kurzweil calls a paradigm) runs out of steam, human innovation comes up with another just in the nick of time.
Again my question: why?
Feedback loops don’t help us answer this question now. The inventors of the transistor didn’t look at the fact that vacuum tube technology had reached its limit in order to schedule the timing of their invention. Human innovation doesn’t work that way.
In other words, disruptive innovation and incremental innovation are at opposite ends of a spectrum – and yet, still adhere to the same curve. Such adherence can’t be coincidence. There must be some underlying principle of human innovation – or broadly speaking, human behavior – that drives both ends of this spectrum of innovation.
I don’t think Kurzweil has an answer to this question. At least, several days of poking around his writing haven’t uncovered an answer. (Ray, drop me a line if I’ve missed something!)
The good news: I believe I’ve cracked this nut. In fact, I approached this question from an entirely different angle in an earlier Cortex where I discussed the nature of disruptive innovation. The insight necessary to answer the question of why depends upon treating any large group of people as a complex adaptive system, where innovativeness is one of the emergent properties of that system.
As I explained in that Cortex, shifting such a system from high connectivity to low connectivity increases evolutionary change. Furthermore, disruptions also lead to increasing rates of evolutionary change, in particular when the system exhibits low connectivity. Such change leads to adaptation to the disruptions in the environment, so combining low connectivity with disruptions leads to periods of rapid innovation and adaptation.
Emergent behavior essentially appears as patterns out of patternlessness. Take a bunch of independent actors (scientists, business people, engineers, etc.), give them a number of external motivations (the drive for profit, the human need to create something new, and so on), as well as various resources – without the high connectivity, rigid reporting structures of traditional hierarchical organizations – and stir.
The end results are specific, repeatable patterns we can take to the bank: in this case, the exponential growth of the law of accelerating returns.
The Intellyx Take: Extending Exponential Innovation to the Organization
The law of accelerating returns applies to many evolutionary processes entirely separate from the behavior of members of Homo sapiens. In fact, Kurzweil points to the evolution of our species itself as an example, from single-celled life to the creation of DNA to the Cambrian Explosion, right on up to the evolution of sentience – a pattern that also follows an exponential curve.
That being said, even for the most technical of innovations, the most important component systems of the complex system of systems we call an organization are people, not technology subsystems. Innovation, after all, is a human endeavor. Technology innovation is itself a set of organizational processes.
What all of these processes have in common – a property, in fact, of all complex systems – is self-organization. Systems as large as galaxies and as small as the cells in your body self-organize. Sentience isn’t required – and paradoxically, often gets in the way, as I’ve discussed in a recent article on self-organizing teams.
Sometimes, however, these well-dressed primates walking the planet get it right. People ask Google their secret to innovation. The answer: self-organization. The same for NetFlix. And if they can do it, so can you.
Whether such self-organization will lead to exponential improvements, however, is an open question. It’s essential to the continuous delivery model that characterizes DevOps – but progress at even the best-run DevOps organizations can only go so fast.
Clearly there’s more to applying the law of accelerating returns to organizational change – but that discussion will have to wait till a future Cortex newsletter. Stay tuned!
Jason Bloomberg is covering this week’s CES conference for Forbes. Vendors at CES with disruptive offerings for the Internet of Things are invited to contact him at agility@intellyx.com.
Intellyx advises companies on their digital transformation initiatives and helps vendors communicate their agility stories. As of the time of writing, none of the organizations mentioned in this article are Intellyx customers. Image credit: Kurzweil AI.
