Our Economy’s Productivity Plateau

I’m of the age where we have bridged the digital divide. I saw how computers have made seemingly slow inroads in more and more of our lives. As a child, if I wanted information, I had to go to a book, maybe an encyclopedia or perhaps a knowledgeable adult. When we traveled, there was always a navigator who was looking at a map and keeping the driver informed of the next turn. Complex calculations required a piece of paper and a calculator, I even remember using an abacus. Phone calls were only made from fixed positions and text messages were notes written on a piece of paper and hand delivered.

While there is a sharp contrast in the convenience of the technology we have today, what is interesting is with that convenience, we don’t really get things done any faster. Having an easy to access GPS navigator may eliminate the need to occasionally stop and ask for directions, we still have to drive at roughly the same speeds. We can find information faster, but we don’t seem to do anything better with that knowledge.

As an engineer that saw both hand drawings and computer-aided design, I anticipated that designs would come dramatically faster with computers. I was really surprised to see that designing with CAD did not decrease the design time. We are able to get more detailed models, solve issues before they are manufactured and usually reduce the hard cost of development, but it still takes about the same time. Some would even argue that it takes longer now. Why is that?

This, of course, isn’t limited to just engineering, it goes across the entire economy. We do have more data, and we are learning a great deal from that data, but that data isn’t making us more productive. Productivity is the value that is created in a given timeframe and we still generate roughly the same value in the same amount of time as we did 50 years ago.

I would propose that the industrial revolution and our ability to innovate in that toolset hit the optimal point half-century ago. The machines that humans use to do things faster and better have stabilized and the improvements have not significantly increased outputs. Cars, tractors, factory automation, shipping, distribution, etc. each of these areas is only marginally better than they were a half-century ago. Add to that, the cost and effort to make those improvements may have outweighed the productivity gains or just evened them out.

How do we improve productivity, if we still think that is the goal? The answer seems to be rooted in how productivity is measured, the output per person over time. The limitation is in the person. A person seems to have a finite ability in augmenting the machines that they control. We need to eat, sleep, and engage in a range of activities to keep our minds sharp. Study after study has shown that increasing a human’s time on the job, doesn’t improve productivity, especially when thinking is the primary requirements. How do we get past the human? We need the machine to start thinking!

The idea that the machine can think, or should think has been around for a long time. This has been both anticipated and feared. To have our machines prepare our food, keep our environment tidy and take care of the menial parts of our life is very compelling. We could end up becoming slaves to our machines is what is feared. For us to be more productive though, we need the machine to discriminate at a level that we would call intelligence. For the car to drive, it needs to be able to react to unforeseen changes in the road. For a good haircut from a machine, it needs a sense of style. For a machine to design a building, it needs to make countless decisions that are subjective in nature.

With the many ideas about how this future looks, I see the same potential for good and ill that technology improvements have made throughout my life. I don’t see the machines primarily looking like humans as has permeated our vision of the future. I think most of our smart machines will continue to look like appliances. I think that we will see new machines that tend to look and act more like insects or small animals, machines that can move, monitor and react to the world around us. Machines that are cheap, efficient and specialize in specific tasks. Productivity will then happen with the machines directing themselves without the need for humans to direct every step of the process. The wheels are already in motion for this change to take place, how long it takes to move to this economy is now the question.

Tiny Learning Machines

Google has everyone talking about their new Google Assistant that can make a voice call and almost master simple tasks with all the nuance of a person.  We are seeing the tip of the iceberg as the dominant tech companies roll out various tools of Artificial Intelligence. The long awaited and maybe dreaded day of truly smart machines has arrived.  What does all this mean for our future?

While the walking and talking robots are the devices which are trying to bring intelligence to life, more likely is that this intelligence will come in a much smaller scale.  The real advantage of AI is in what it can do for sensors and control devices.  We already can make very small and cheap programmable devices, by adding machine learning to those devices we can dramatically increase their capability and use.  We can make them even more energy efficient while allowing them to provide better information.  Small devices that can decode human voices, better catch environmental changes, and control devices based on being taught by the user instead of being programmed set commands.  What might some of these tiny learning machines actually do?

In the home

Tiny sensors that have small batteries and recharge via indoor solar panels or even breezes through the house.  Sensors listen and see what is happening and communicate wireless to a central controller.  Being smart they know when something actually changes and don’t do anything until they are alarmed.  They learn the difference between the window being broke and a dish breaking.  They can recognize those that are regularly in the house and those that are guests or strangers.  They can listen and respond to voice commands, without sending requests out to the cloud.

In the world

If a tree falls in the woods, and no one is there to hear it, does it make a sound?  We can make tiny and cheap sensors that can sprinkle the earth and let us know what is happening in places that we can’t observe.  From finding new species and plants to better understanding our climate, we have so much more to learn about this world we live in.  By having smart devices we can just learn what is relevant, what is truly unique instead of getting raw data.

Eliminate redundancy

In each of our lives, we end up doing many redundant and tedious tasks.  Simple things to do but they just take time and energy away from other more important activities.  If we can easily teach our devices to do new tasks, we have the ability to have our devices help us.  This was where programming and software started, but now the age of AI is opening up the idea that we can all become teachers and trainers for our devices.

Transcend Innovations helps our clients accomplish their dream to change the world.  If you have a device that you would like to create, we have the people can turn it into a reality.  For more information visit our website at or email us at