Knowledge-economy talk is currently, and rightly, focusing attention on issues of intellectual capital, lifelong learning, and people as the prime movers behind innovation. Companies, in turn, are starting to invest more in their organizations as learning systems, not just as information systems. History suggests, however, that it’s not easy to accelerate human learning and that progress is likely to be slow.
But something unexpected is about to happen. When inanimate objects learned to remember (that is, we had books), the world changed. When they learned to calculate (that is, we had computers), it changed just as radically. Now they are about to gain the ability to learn much faster than we human beings do.
How? In a BMW ad, a driver is stranded, his wheels spinning on an icy road. He uses the onboard customer-service system to ask for help. A BMW technician, examining the telemetry from the car, finds a solution in her computer, and downloads a software patch to the customer’s traction-control system. Problem solved — one car educated.
What’s next? Why not automate this process? Today’s car, of course, knows when its wheels are spinning and gathers lots of local data on temperature, on air pressure, on moisture, etc. Why not trigger a data upload every time there is a malfunction? Up in the sky, some Queen BMW uses genetic algorithms and simulations to breed a new, improved software-traction control release 404.3. (Such techniques are already in use, for example, to schedule tractors at Deere & Co.) Then the Queen downloads the new software — not to just one car, but to the entire fleet. This means that the car you start up this morning isn’t the same one you put in the garage last night! While you were asleep, it was at night school.
In principle, this is not new. When one 737 displays frayed wiring, the FAA insists that the whole fleet “learn” the solution. This is occasional and slow. In the software world, though, it’s already continuous. My Macintosh updates its operating system and virus protection weekly, without any help from me. Soon everything will be made of software — not just cars but toasters, toys, and tins of tuna. Services too. The role of the customer will be to drive the system into corners it hasn’t encountered before; the role of the product or service will be to gather data on how the system is performing, so that the whole “hive” can learn and benefit. Your toaster will read the bar code of your bread and learn not to burn it, so that no toaster will ever burn that high-sugar-content loaf again.
Creepy? Sound a bit reminiscent of The Truman Show? Could be, but I don’t think so. We like our updated 737s and antivirus programs, and we’ll come to like, rather than fear, an environment that’s always learning on our behalf.
When science-fiction writers alerted us that machines were becoming smarter than we were, they worried about robots, still thinking that anything that learned had to look like a person. As the computer gives way to special-purpose “information appliances,” appliances — and all the other products in the economy — are becoming computers, linked to enormous information resources and their colleges of peers.
The fast learning in the next decade is going to be done by the stuff around us. We’ll be lucky to keep up.
Christopher Meyer (firstname.lastname@example.org) is vice president and director of the Cap Gemini Ernst & Young Center for Business Innovation. He coauthored Blur: The Speed of Change in the Connected Economy with Stan Davis.