End of the Rainbow (eBook)

1

Our Standard of Living

When it comes to determining how well off we are economically, one of the best measures is Gross Domestic Product (GDP) divided by the number of hours worked. It’s not a perfect measure, due to difficulties of data collection. For example, it counts depreciation as final production rather than an intermediate cost of production and it doesn’t include the vast amount of unpaid work people, particularly women, perform at home, but it’s the closet approximation we have.

As of 2019, the most recent year for which data is available, we in the United States produced $74 worth of goods and services for every hour we worked. In 1948, we only produced $21 an hour (inflation adjusted), about the same as modern Mexico (Bureau of Economic Analysis). This increase manifests itself in both an increased quality and quantity of goods and services. This is why we have nicer homes, more automobiles, and exciting gadgets like televisions, air conditioners, cell phones, and laptops that the people of 1948 could only dream of. It also means that we have more left over after satisfying our basic needs. In 1959, 19% of all consumer spending in the United States was for food, 9% was for clothing. In 2019, those figures were 7% and 3% respectively. (Bureau of Economic Analysis)

Progress of Civilization

How did we produce this great abundance? Fundamentally, it’s because we increased our knowledge of the forces of nature and harnessed them for our use. The most important breakthrough came in 1947, when three researchers at Bell Telephone Laboratories in Murray Hill, New Jersey discovered a device that could act as an electric switch without any moving parts. This was the first transistor and it made possible the electronic age. Today, much work done by humans in 1948 by is now handled by robots and computer software.

The arrival of low-cost computing power has also enabled an increase in the quality of products. Automobiles today use electronics to handle fuel injection and the timing of valves, allowing them to get much more power from each gallon of gasoline. All of this adds up to a more than three fold increase in output for a given number of hours worked.

Furthermore, we will continue to reap the gains of increasing scientific knowledge. Currently, studies are being conducted on mice where electrodes are implanted into their brains. The mice are then observed performing various exercises. (Spectrum IEEE) The purpose of this research is to understand how mice are able to learn. Such knowledge can be applied to create new artificial intelligence (AI) systems that are able to learn new things with small amounts of data. Right now, if you want to get a neural network to recognize a dog, you’d need to show it thousands of different images of dogs. That makes AI useless in unstructured environments where very unusual situations will come up. But if the AI only needed one image of a dog to figure out what a dog is, then that changes things dramatically. It becomes possible to automate many jobs.

In the following table, I have produced my rough estimate for how many jobs in each field could be automated based on my assessment of what technological progress will occur by the year 2060.

Automation Potential by Job Field