THE PROBLEM

In two companion papers (AP1.3A and AP1.3B), the trend toward automation of the work of civilization by computers and robots and the resultant loss of jobs was described. It was noted that almost all jobs that do not require the ability to recognize trends, sort information and use common sense to make decisions based on these things would be taken over by computers and robots. In addition, workers will need to find ways to enter new industries that are uniquely suited to small, fast changing businesses. A new potential source of work to counter this trend and provide employment for the coming generations was also described. It was claimed in AP1.3A that this new source of jobs (the SEMAN, a vessel that operates on the oceans to harvest energy from wind, waves and sun) is free from the threat of automation, but little proof of this claim was given. It was further noted that the SEMAN was capable of giving a middle class income. It was finally claimed that the SEMAN was capable of delivering work that is resistant to the impact of an economic downturn. But the SEMAN delivers sea-based jobs, and many people would rather work on land. In the second companion paper AP1.3B, the land jobs that are generated by the SEMAN are described, and it was noted that these jobs would be resistant to automation and recession as well. In this paper, the limits of automation will be explored, and the resultant characteristics of these potential new land and sea based, good paying, automation resistant jobs described. It will be found that in general, a job that works toward independence from outside economic forces has the greatest potential to provide work that is good paying and automation resistant, and such jobs are almost ready to come on line. Further, they are both land and sea based, and are also connected to the SEMAN. Finally, it appears possible that these jobs may provide a path for mankind to start working toward the basic allocation of food, clothing and shelter needed to supply a recession-proof living for each person upon birth.

Definition of Terms

Certain terms used in this paper have different meanings to different people, so in order reduce confusion; the following definitions will be used here.

Goal-the end result of a sequence of operations that improves the happiness and productivity of mankind.
Operation-the combination of a sequence of tasks that aims at the accomplishment of a goal.
Task-the fundamental unit of work that aims at completion of an operation.
Computer-a machine capable of performing logical operations and defining tasks for an operation that exists in cyber space.
Robot-a machine capable of carrying out tasks defined by a computer in the real world.
Automation-the process of replacing a human worker by a machine such as an assembly line, the Internet, a computer or a robot or a combination of these machines.

Desirable Job Types

The job types being sought in the world have the following characteristics.

The job should not be capable of being automated.
The job should yield a middle class income.
The job should be resistant to the impact of an economic downturn or recession.
At least some of these jobs should be land based.

These characteristics will be detailed below.

Limits of Automation

There are limits to the ability of computers and robots to take over the jobs usually done by people. The intrinsic character of the computers and robots determines these limits. The most important of these characteristic limitations are as follows.

Computers and robots are normally constructed to provide products for, do tasks for and increase the productivity of mankind, and so they are not able to determine how to fill these needs without input from mankind to define the goals of the operations performed. These inputs constitute the more creative aspects of work Computers are also not suited to recognizing trends (pattern recognition), sorting information and using common sense to make decisions.
Computers and robots are best suited to the tasks that are simple and repetitive that must be done for civilization to function, while allowing people to determine the goals and operations that accomplish the goals. Thus humans program the computers. Computers don’t normally program computers.
Computers and robots are not able to accomplish tasks and complete operations for incompletely defined goals, and cannot complete tasks for operations when the data on which the task are based are not a complete set. By using general knowledge of the goal of the work and how it can be done (often called common sense), a human can complete such tasks.
The cost of computers and robots is high and they are usually designed for a specific set of tasks in an operation, and so this equipment must be used enough for that operation to justify its initial expense. A human, on the other hand, is versatile, and can cover many different kinds of jobs, and his/her initial expense is lower, so it is easier to justify hiring him/her. Note, however, that computers and robots are usually cheaper to operate than humans, and their overhead is lower, so over the long run, if the utilization is high enough, computers and robots are cheaper and more efficient.

It should be noted that there are computer and robot designers hard at work trying to expand the scope of computer/robot capabilities to include the more creative aspects of human capability, and so allow computers and robots to work on the problems requiring creativity. Also, designers are working to expand the capability of computers to work indeterminate problems. Thus far, however, they have had very limited success, so the limits of automation described above are still valid. The reason for this lack of success lies in the difficulty for computers to understand the goal of and the meaning behind an operation or a task. Also, designers have difficulty defining and imitating “common sense” or general knowledge of the situation. This difficulty can be illustrated with the following example.

Normally, a circumstance (defined by a set of sensor data) indicates a task required to complete an operation necessary to achieve a goal. A rule set for that circumstance indicates what task is required to complete the operation toward that goal. If a computer is given a clear set of rules, it can apply them satisfactorily by selecting from a set of simple tasks. Goals are often fuzzy, however. This makes the operations fuzzy, the rules unclear and the task not quite correct. Thus, for some fuzzy goals, and for some complex operations, it may not be possible to establish a clear cause-effect relationship for a task in achieving the goal, and so it is necessary to determine which one of several rules applies to a task, based on the (often incomplete) sensor data available. In this circumstance, a human is indispensable. The human can use common sense and his/her “feel” for the goal to choose and even modify the task but still achieve the goal. For example, a robot is told to plant a tree at a specific position in a garden. When it starts to dig, it finds a large rock in that position. It is not equipped to penetrate the rock, and even if it were, it would not realize that the tree would not grow well surrounded by rock, so it must stop. A human would see that a small change in the position of the tree would not violate the esthetic goal of the garden plan (a fuzzy goal), and common sense would tell him/her that the rock does not underlie the whole garden, so he/she would move the location a little, and plant the tree. Here, the problem is a lack of understanding of the fuzzy goal and an incomplete understanding of the environment. Thus computers stumble because they cannot understand the meaning behind a task (a fuzzy goal) or have the or the pattern recognition capability to note the limits of the rock, or have the common sense needed to bypass the problems encountered in achieving that goal. In fact, since humans designed the goal, they are uniquely suited to modify it when necessary.

Thus, the jobs that actually can be automated by computers and robots fall into three categories in increasing order of difficulty of design, construction and operation.

Simple, fixed position, single task, fixed instruction, repeatable jobs.
Simple, mobile, single task, fixed instruction, repeatable jobs.
Complex, mobile, multi task, changing instruction non-repeatable jobs.

None of these will require understanding the goal or having common sense. The use of computer and robot devices in automating all three major job types will be described in the SOLUTION section below, and the jobs resistant to automation indicated.

Income Level

Jobs are not necessarily compensated by money alone, but if it were, an annual income of $50,000 to $100,000 would be roughly correct to maintain a comfortable life style for a family of two adults and two children. This is usually considered to be the income range for middle class families. If rent and food were provided by the job, a proportionately smaller income would be acceptable. If maintenance for equipment used in the job were necessary, a proportionately larger income would be required. This $50,000 to $100,000 income range is what will be aimed for in this paper.

Resistance to Economic Downturn

All economies are subject to oscillations. If a job were only good for one cycle (~5 to 10 years), it would be considered unstable and thus undesirable. If a job were good for a working lifetime (~30 to 60 years), it would be considered stable and desirable. Job stability through several recessions and extending toward a working lifetime is what will be aimed for in this paper.

Location of Jobs

For some, the location of the job and its “cultural environment” are very important. Jobs on the ocean would be less desirable. At least some of the new jobs should be land based.

THE SOLUTION-NEW EMPLOYMENT IN NEW JOBS

Existing Job Types

Let us start by going through the entire list of jobs that exist in the US according to the latest census, and describing how and why a large number of these jobs are fast disappearing from the modern economy as witnessed by current trends shown by a sequence of these censuses. In a world without automation, we would expect that the proportion of jobs in each category would remain roughly the same even as the population rises because the need for the output from each category is proportional to population, and so the need for workers in each category is also proportional to the population. Two things can change this conclusion, automation, which reduces the number of workers in a category for the same output, and new industries, which increases the number of categories, and so the number of jobs. In order to maintain employment for new generations of workers, new industries must balance automation. In this list of job categories, the change in the number of jobs in the existing job categories due to automation will be addressed.

Farming, fishing and mining jobs are largely definable, simple, mobile, single task and repetitive and so are amenable to automation by category 2 devices. A lot of the simple jobs are gone now. The remaining definable, simple jobs (mostly harvesters and machine operators) are gradually going to category 2 devices, leaving the business owner/manager and maintenance workers. A reduction of the share of the work force by 1/3 is expected in this simple category by 2050.

Manufacturing jobs are mostly definable, simple, fixed, single task and repetitive, and so they have almost disappeared, having been replaced by the assembly line and category 1 devices. The remaining jobs are those requiring category 2 and 3 devices such as 3-D printing manufacturing. By 2050, use of more advanced robots for these more complex tasks is expected to reduce this share of the work force by 1/3 toward designers/engineers, managers and maintenance workers.

Construction jobs are largely definable, complex, mobile, and single and multi task and changing instruction and so are amenable to automation by category 2 and 3 devices. These basic jobs (carpenters, plumbers, electricians, etc.) are disappearing slowly because of the difficulty of the computer/robot design, but eventually only architects, engineers, construction managers and maintenance workers will remain. By 2050, new computers and robots now coming on line are expected to reduce this share of the work force by ½.
Wholesale and retail trade jobs have been difficult to automate because they are hard to define and many people would prefer to see their purchases and trade with a person. The Internet offered a way around this difficulty, and so it is making inroads into this group of jobs. By 2050, computers are expected to reduce this share of the work force by 2/3.
Transport, warehousing and utilities are mostly definable, complex, mobile, multi task and changing instruction and so are amenable to automation only by category 3 devices. These devices are difficult to design, so this shift is going slowly. By 2050, computers and robots are expected to reduce this share of the work force by 1/5.
Information jobs are mostly indefinable, and so are not amenable to automation.
Finance, insurance, real estate, rental and leasing jobs are hard, but not impossible to define but people would prefer to trade with a person, so automation is going very slowly in this category of jobs. Little change is expected here.
Professional, science, management and administration jobs are mostly indefinable, and so are not amenable to automation.

Educational services, health care and social services jobs are difficult to define and people prefer to deal with people here, so this category is mostly not amenable to automation, but some of these jobs can be automated by category 3 devices and new health care treatments. This industry is expected to increase its share of the work force for a short time especially in health care as the population ages, and then start to lose its share as more efficient means of accomplishing these goals become available. The reduction will take place especially in health care and connected social services, where anti viral medicines and gene repair techniques (for cancer, heart disease and Alzheimer’s and other brain damage problems) that are just coming on line are expected to eventually have a dramatic impact on the number of health care and connected social services workers required. By 2050, a reduction of this share of the work force by 1/2 is expected to result. .

Arts, entertainment, recreation, accommodations and food services jobs are partly indefinable and the rest difficult to define, so if they are automated, it will be done slowly by category 3 devices. In the mean time, by 2050, an increase of this share of the work force by 1/3 is expected to meet new demand because of an increase in leisure time due to automation.
Public administration and other services jobs are mostly indefinable and so are not amenable to automation. Little change in this share of the work force is expected.
Armed forces jobs are mostly definable, complex, mobile, multitask and changing instruction and so are amenable to automation by category 3 devices leaving officers, technicians and maintenance workers. The difficulty and expense of the computer/robot design for this purpose and the reluctance to put this dangerous task in the hands of robots is expected to make automation extremely slow, so by 2050, only a modest change in this share of the work force is expected except through changing political climate. By 2050, a reduction in this share of the workforce by 1/5 is expected.

There are three other occupations that are important.

People who are self-employed owner/operators and live on their business or owners that live on their investments will increase. This occupation is expected to increase its share because more people will use investments and business ownership to become independent of the work-a-day world. By its very nature, this employment cannot be automated. This is expected to take a rapidly increasing share of the work force-see below for more on this category.
People who are employed in new innovative industries that form a new job category. The proportion of the work force in this category is expected to be small, because most new jobs are the incidental result of an attempt to make a profitable new industry. As a result, the new industry is designed to minimize the number of jobs in order to maximize the profit. This category is expected to account for less than 5% of the work force.
People, who are unemployed, disabled or are on welfare. The acceptable percentage of unemployed in an industrialized nation is roughly 3%. The acceptable percentage of those disabled or on welfare is roughly 2%. For the US, this 5 % works out to be is 4.6 million people. World wide (assuming the world economy becomes similar to the US), this percentage works out to be 123 million people. The unemployed category is expected to increase its share of the work force until a way is determined to move them into the professional, science, management and self employed owner/operator categories that cannot be automated, as new jobs of this type are generated. Eventually, this category will reach an irreducible minimum and support for them will have to be provided as civilization’s overhead.

Note that most jobs are amenable to automation, but some are not, and the people filling the jobs that are not, have job security. Note that the ones providing the most job security are, or are the equivalent of, a small business owner/operator. However, such a small business must be resistant to bad economic times and still provide the necessities of life or it will not survive a recession. Further, the business must provide a middle class income to maintain an acceptable standard of living for the owner.

Required Characteristics of New Jobs

In order to have the job characteristics of resistance to automation, middle class income, and resistance to recession, the new jobs should have the following characteristics.

The worker should be able to address incompletely defined goals, and/or use incompletely defined data, so computers and robots cannot replace him/her. In addition, the work should be complex, mobile, multi task, and require changing instructions in order to maximize the difficulty and cost of replacing the worker with a computer and/or a robot.
The product must be producible at a cost that always allows for sale at a profit.
The product must have a large, long-lived market that will maintain itself through an economic downturn.
The fundamental needs of life (food, water, shelter and domestic energy) should be available from the work independent of the short-term profitability of the business in order to insure that the worker can ride out a recession without losing his business or getting into debt simply in order to supply personal needs. Thus it must be possible to go into a kind of economic hibernation until business conditions improve.
The capital equipment necessary to accomplish these tasks will normally require a loan or mortgage to pay for it. Thus a certain minimum income is required to make mortgage payments.

Few, if any jobs can boast all of these characteristics. Yet by combining several occupations, a close approximation can be achieved. Consider the following.

Owner/operator is the job type most likely to require addressing incompletely defined goals and data (characteristic 1) because it requires certain creative characteristics on the part of the worker. Professional, science, management, arts and entertainment jobs also are likely to fit this requirement, but there are fewer people that are expected to fit the requirements for these latter jobs.
In order to insure that the product for sale can be produced at a profit (characteristic 2), the effort required of the worker, the materials required for production, the cost of the place the product is produced and the equipment needed to produce it must be minimized. The minimum is to produce the product with carpentry and electrician skills with hand tools in the worker’s shelter or home.
The products that maintain a strong demand during a recession (characteristic 3) are the basic needs for survival, namely food, water, shelter and domestic energy.
If the worker’s shelter is properly constructed, the shelter is near the ocean, and there is a source of domestic energy, the worker’s fundamental needs (characteristic 4) can be supplied by use of hydroponics for food, reverse osmosis for water and wind, waves and sun for domestic energy.
In order to make mortgage payments (characteristic 5), a certain minimum sale of product is required. If the demand in a recession is strong enough, this can be accomplished. This can be done on three levels, namely

Full mortgage payments.
Interest only.
Suspension of payments until sales improve.

The last two require patience on the part of the mortgage holder.

New Jobs to Fulfill the Need

Most new jobs are the incidental result of an attempt to make a profitable new industry. As a result, the new industry is designed to minimize the number of jobs in order to maximize the profit. Thus the number of jobs created is small. Here, an industry under development will be described that is designed to maximize the number of interesting, good paying jobs that cannot be automated and are recession proof. Specifically, the company Aquater2050 LLC that operates and builds the SEMAN or Seagoing Energy Module with Automatic Navigation-stabilization will be described. This SEMAN is designed to harvest clean energy from the oceans to replace fossil fuels and provide 400 million jobs on the sea and 150 million jobs on the land. None of the primary jobs indicated can be automated, but some auxiliary jobs connected with the SEMAN will be amenable to automation. These jobs and their differences will be described.

The SEMAN is an ocean going vessel that harvests and sells energy. In order to maximize the efficiency of this operation and so make the SEMAN profitable, the SEMAN designer collects a group of the needs of the operators in one package and, in the SEMAN, offers shelter, food, fresh water, domestic energy and transportation, so the operator can hunt for and harvest energy without being concerned about his/her basic needs. Note that this set of capabilities works toward independence from outside economic forces such as recession or change of consumption habits. Whatever happens in the outside world, the operator has his/her basic needs covered, and the sellable product, energy, is required by all, and so the market will never disappear. Economic conditions can drive down the selling price, but they can never cause the need to disappear. Note also that the SEMAN provides a means to sequester carbon dioxide, which will be required for the foreseeable future. Note also that a computer would have no means of determining what the needs of mankind are and which can be grouped together and how to group them without human input, so a human designer is required. On the other hand, construction of a SEMAN consists of a series of easily defined (carpenter and electrical tasks), repeated tasks that can be automated with robots. Further, since a large number of SEMAN are required, the robot’s expense can be amortized over many SEMAN. Clearly, a human must fill the design jobs, and a computer cannot replace those humans completely, but robots can accomplish construction. Note also that people with relatively simple skills can also accomplish construction; so common workers can build SEMAN in developing countries in the initial stages of the development of this industry.

The operation of a SEMAN consists of choosing an operation zone in an energetic (wind speed >15KN) area of the ocean, choosing an operating direction to optimize energy production, and operating the vessel in such a fashion as to maximize the amount of energy harvested for sale on land, maximize food production for personal consumption, and allow for the pursuit of some profitable side businesses. The maintenance of the vessel, however, consists of a set of fixed, definable tasks (such as setting the angle of the sails and rudder for maximizing energy output) that can be automated, thus allowing the SEMAN to operate with few crewmembers. This allows the business to save on operating expense. The choice of the operation zone and operating direction are not completely or clearly defined, and so are not amenable to automation, but most of the crew positions can be automated. The designer and the captain (owner/operator) appear to be safe in their jobs. The captain can take his/her family along since there is food and shelter available. The SEMAN can and probably will operate in groups or “pods”, to provide company, overcome loneliness, and insure help in case of trouble. Note that one SEMAN captain, by use of robotics and computers, can operate and control a few nearby SEMAN to reduce the number to operators and captains required if needed to cover the need for energy produced. However, the need for captains and operators can never be completely eliminated.

Note finally, that the land-based jobs connected with the SEMAN will consist of small, probably franchised and often isolated businesses, widely spaced around the coasts of the world or on the ocean in order to conveniently receive and sell energy, and make, outfit and trade with SEMAN owner/operators. Because of the isolation of these businesses, it will usually be necessary for these owner/operators to provide their own food, water and domestic energy by use of the same techniques as SEMAN. These businesses will, of necessity, remain small and separated and are not amenable to automation. However, because these businesses cannot pursue optimum energy gathering conditions like SEMAN, they cannot always provide enough energy. The SEMAN must provide back up energy under these conditions.

We see that this way of life can provide:

A basic allocation of food, clothing and shelter needed to supply a living for each person.
A salary that gives a middle class income and is relatively independent from outside economic forces because it comes from the sale of basic necessities of life such as energy, fertilizer, food and water.
A job that is automation resistant because it consists of tasks and goals that are incompletely defined, and are based on data that are not a complete set.
A minimum income requirement during recession that is small if the mortgage requirement can be temporarily suspended in case of recession.

Few other industries existing or under development are known that provide all of these characteristics, but many can be modified to provide them. It is believed that now is the time to start modifying jobs to provide these characteristics for all citizens that are capable of working.

It is important to note that by adjusting the amount of automation used for land construction versus the number of SEMAN a seagoing captain controls, it is possible to adjust the ratio of sea to land jobs within limits. Thus, people can choose the lifestyle (land or sea) they wish to pursue within limits.

SUMMARY AND CONCLUSIONS

The job should not be capable of being automated.
The job should yield a middle class income.
The job should be resistant to the impact of an economic downturn or recession.

In order to have these job requirements, the job type should have the following characteristics.

The worker should be able to address incompletely defined goals, and/or use incompletely defined data, so computers and robots cannot replace him/her. In addition, the work should be complex, mobile, multi task, and require changing instructions in order to maximize the difficulty and cost of replacing the worker with a computer and/or a robot.
The product must be producible at a cost that always allows for sale at a profit.
The product must have a large, long-lived market that will maintain itself through an economic downturn.
The fundamental needs of life (food, water, shelter and domestic energy) should be available from the work independent of the short-term profitability of the business in order to insure that the worker can ride out a recession without losing his business or getting into debt in order to supply personal needs.
The capital equipment necessary to accomplish these tasks will normally require a loan or mortgage to pay for it. Thus a certain minimum income is required to make mortgage payments.

Few, if any jobs can boast all of these characteristics. Yet by combining several occupations, a close approximation can be achieved. Consider the following job types.

Owner/operator is the job type most likely to require addressing incompletely defined goals and data (characteristic 1). Professional, science, management, arts and entertainment jobs also are likely to fit this requirement, but there are fewer people that are expected to fit the requirements for these latter jobs.
In order to insure that the product for sale can be produced at a profit (characteristic 2), the skills required of the worker, the materials required for production, the cost of the place the product is produced and the equipment needed to produce it must be minimized. The minimum is to produce the product with carpentry and electrician skills with hand tools in the worker’s shelter or home.
The products that maintain a strong demand during a recession (characteristic 3) are the basic needs for survival, namely food, water, shelter and domestic energy.
If the worker’s shelter is properly constructed, the shelter is on or near the ocean, and there is a source of domestic energy, the worker’s fundamental needs (characteristic 4) can be supplied by use of hydroponics for food, reverse osmosis for water and wind, waves and sun for domestic energy.
In order to make mortgage payments (characteristic 5), a certain minimum sale of product is required. If the demand in a recession is strong enough, this can be accomplished. This can be done on three levels, depending on the patience of the lender, namely
1. Mortgage only.
2. Interest only.
3. Suspension of payments until sales improve.

The company Aquater2050 LLC that will operate and build the SEMAN or Seagoing Energy Module with Automatic Navigation-stabilization was found to have these characteristics. This SEMAN is designed to harvest clean energy from the oceans to replace fossil fuels and provide 400 million jobs on the sea and 150 million jobs on the land. This way of life can provide:

A basic allocation of food, clothing and shelter needed to supply a living for each person.
A salary that gives a middle class income and is relatively independent from outside economic forces because it comes from the sale of basic necessities of life such as energy, fertilizer, food and water.
A job that is automation resistant because it consists of tasks and goals that are incompletely defined, and are based on data that are not a complete set.
A minimum income requirement during recession that is small if the mortgage requirement can be temporarily suspended in case of recession.

Notes

1. More details and references are given on Aquater2050.com.

2. For those interested in aiding the completion of the SEMAN prototype to speed the development of these new jobs, a donation button is provided on this site.