Have you noticed there has been a notable acceleration in technological progress in the past few years?
In 1977, Japanese innovators began a project to create an autonomous vehicle. Others followed. Progress was minimal and slow, until recently.
The Defense Advanced Research Projects Agency (DARPA) issued several competitive challenges for applicants to create autonomous vehicles, the first in 2004. The winning entry traversed 7.3 miles of the 149-mile route. Not particularly inspiring.
In 2005, five vehicles completed a more rigorous course, with durations ranging from about seven hours to more than 12. It might take weeks to get to Grandma’s.
In 2007, DARPA issued the Urban Challenge, requiring autonomous vehicles to traverse 59 miles in less than six hours, through live urban traffic. Six teams finished the course, with four achieving the goal between four and six hours, averaging about 14 miles per hour. Yawn.
In 2013, Google’s autonomous vehicle demonstrated unaided negotiation of city streets and freeways, speeding along at posted limits, carrying several passengers who chatted, read books and napped.
This is just one example of how technology advances are accelerating. If you investigated rates of progress in other fields, you would find similar, logarithmic change. As Ray Kurzweil pointed out in his book, “The Singularity is Near,” the rate of innovation itself is accelerating. Why is that?
Most experts point to the digital information revolution as the underlying cause. Certain advancements achieve great impact because they are broadly applied to multiple avenues of invention, and are called “general purpose technologies” (GPT) by economists.
The invention of the typewriter advanced the cause of producing written materials, for example, but it didn’t affect other human endeavors, and therefore its impact was localized to its purpose. The invention of distributed electrical power, however, applied to many things, including typewriters, and therefore had a much larger impact on human productivity.
Digitization of information, and its broad accessibility, is considered by many these days as a GPT, equal in significance to the invention of steam engines, which was the cause of the first industrial revolution.
As power increased and cost declined, digital information processing capability achieved almost universal applicability. We all know how our lives are now influenced by such technologies. What may not be so clear is how fast and furious the perfusion will continue to be.
Kurzweil’s analysis of the rate of acceleration of innovative changes results in some fairly staggering predictions. For example, by charting the rate of innovation acceleration in nanotechnology, biotechnology and artificial intelligence, he estimates that it will be possible to create human-level intelligence out of non-biological materials around 2045.
Others are tracking the rate of innovation as well. In a new book by Erik Brynjolfsson and Andrew McAfee called “The Second Machine Age,” the authors identify trends in this innovation acceleration that challenge our historical view of how invention and problem-solving occur.
For example, the digitization of information allows almost anyone access to the full body of knowledge about any given subject, and to receive updated information in nearly real time.
Millions of books dating back centuries have been input to the digital network that now spans the world, connecting hundreds of millions of people. Data sets from countless scientific papers are accessible online. And the use of non-professional investigators to solve intractable problems is rising.
In fact, recent examples of “crowdsourcing” solutions to difficult problems has demonstrated that solutions often come from people who are not trained or experienced in the field attempting to address that problem. Brynjolfsson and McAfee point to NASA as an example.
The space agency, perennially budget-challenged, was trying to figure out how to better predict solar particle events, or solar flares. The eruptions from sun storms cause damaging radiation that can destroy space-based electronics and threaten the safety of astronauts caught in it.
After 35 years of effort, NASA had come up empty. The best astrophysicists and scientists of NASA couldn’t produce a model that gave enough notice with enough predictive reliability. So the agency invited all comers to offer solutions by posting the problem on Innocentive, a clearing house for posing and solving scientific problems. Innocentive allows non-credentialed participants. You don’t need a Ph.D. in the requisite field to opine.
A retired radio frequency engineer from a small town in New Hampshire, Bruce Cragin, invented a method that provided 85 percent accurate predictions with eight hours advanced notice, and 75 percent accuracy with 24 hours notice. More than enough for astronauts and engineers to prepare for protections.
The examples are themselves accelerating, of expanding the participants in problem-solving through universal access of the information needed to work on problems. And the solvers aren’t confined to humans.
Computer algorithms combined with many different types of sensors and input devices, are doing things that only humans could do in the past, including complex pattern recognition, contextual interpretation and empirical learning. For example, advanced integrated circuit design is being done by computers containing advanced integrated circuits.
We need to be ready for the acceleration of advances that are now at an inflection point of rate. Just like historical advances that qualified as GPTs, digitization of information and its light-speed transmission allowing universal accessibility is now at a point where it will rapidly change everything it touches. And there’s nothing it won’t.
Sewitch is an entrepreneur and business psychologist. He serves as the vice president of global organization development for WD-40 Company. Sewitch can be reached at firstname.lastname@example.org