Surely the development of the electronic computer will be viewed by future historians as one of the great milestones in human history. The introduction of the computer into the manufacturing process carries the potential for changes as profound as those resulting from the development of the steam engine or the discovery of electricity.
The computer is qualitatively different from all other machines in several important respects. First, its fundamental mechanisms are electronic rather than mechanical. As a result, a computer can operate many orders of magnitude faster than other devices. Typically, it can perform about a million separate and discrete operations every second.
Second, a computer does not wear out in any normal sense of the word. Although semiconductor components do deteriorate as a function of age and temperature, this deterioration is not appreciably affected by use. A computer does not age any faster when being operated at top speed than it does when sitting completely idle.
Third and most important, a computer can store and manipulate large quantities of information. A computer can store instructions, make decisions, calculate formulae, and execute procedures. As a result computers are able to manage businesses, schedule factories, maintain inventories, and control many types of mechanical systems in the performance of complicated operations. In theory, if not yet in fact, computers are capable of performing almost all of the decision and control functions currently done by humans in the basic manufacturing industries.1
Computer-Aided Manufacturing
Computer-controlled machines and automatic factories are no longer science-fiction fantasy. Many hundreds of computer-controlled machine tools are already in operation throughout the United States and the world, and several computer-controlled factories have recently begun production.2-7 A technology forecast conducted by the University of Michigan predicts that by 1980 computer systems for full automation will be used in the manufacture of at least one-fourth of all parts, and by 1985 approximately 60 percent of all machine loading and scheduling will be done by computers.8
Recent advances in computer-aided design reveal just the merest hint of what will soon be not only possible but routine. Engineers, architects, and scientists are now able to use the computer as a design tool to sketch objects in three-dimensions, to visualize how parts and structures fit together, and to analyze the performance of electronic circuits.9 A designer can communicate with the computer through a keyboard, a drafting machine, or a computer-controlled display device similar to a home television screen.10 Present computer-aided design techniques often reduce engineering and drafting costs by a factor of three or more.11
In the future it will be possible to connect together computer -aided design equipment with computer-controlled machine tools. This will make it possible for an engineer to design a part at a computer terminal in his office. When he is satisfied with the design, he will be able to push a button and cause computer-generated control signals to be transmitted to an automatic machine tool, perhaps many hundreds of miles away, where the part will actually be produced without further human intervention.
Computer applications in manufacturing promise even more dramatic results. Over the past two decades enormous cost reductions in metal cutting, particularly in the aerospace industry, have been achieved by the rather simple expedient of operating machine tools under control of a magnetic or paper tape rather than human operators.12 Numerically-controlled machine tools routinely result in productivity increases of 150 to 400 percent in present job-shop environments.13 Recently, direct computer control has been added. Since 1970, low-cost mini-computers have become available that make it economical to dedicate an entire computer to controlling a single machine tool.
Preliminary results show that direct computer control can produce increases of 3 to 10 times in machine productivity, and even more dramatic improvements seem likely in the future.14
Computers also have potential applications in controlling automatic assembly machines, automatic parts -handling robots, automatic warehousing and inventory systems,15 and finally, computer-based management and factory scheduling systems. The Rand Corporation has estimated that, if computers were employed in controlling all these steps, o v e r all reductions in total manufacturing costs of two to four times are achievable.16
There are two important factors concerning cost reductions obtainable through the use of computers in manufacturing that tend to make the benefits extend far beyond the industrial sector of the economy. The first is that the basic manufacturing industries are the foundation stone upon which the entire socio-economic system rests. The cost of manufactured goods affects agriculture and the service industries as much as it affects the manufacturing industries themselves. The cost of tractors, combines, milking machines, irrigation equipment, fertilizer, fencing, and farm buildings all depend on the cost of manufactured goods. The same is true in the services. The cost of telephones, trucks, planes, railroads, typewriters, schools, books, hospitals, medical equipment, furniture, houses, even the cost of churches, vacation resorts, and sporting goods depends on the efficiency of the basic manufacturing industries. Of course, the cost of all raw materials such as steel, aluminum, concrete, oil, and nuclear power, as well as future costs of solar and geothermal energy, depends on the cost of manufactured materials. Thus, any significant reduction in cost in the manufacturing industries has a multiplier effect that ripples through the entire economy.
A second, even more important feature, of cost reductions in manufacturing is that they are regenerative. For example, if the cost of a machine tool is reduced, then the cost of another machine tool produced by the less expensive tool will be less expensive still.
Today, machines are used to make machines. Any basic improvement in technology that increases productivity at this point in the economy has a regenerative effect. It tends to make the cost of wealth-producing capital equipment spiral downward exponentially. Studies predict that even first generation computer-controlled factories will be no more expensive than conventiona1 plants despite their greater complexity. They may, in fact, be cheaper by as much as two-to-one. Second and third generation automatic factories may be five or ten times cheaper.
The fact that this regenerative effect does indeed occur in basic manufacturing industries is dramatically illustrated by the history of the computer industry. Computer-aided design and computer-aided manufacturing techniques have been used most extensively in the design and construction of computers themselves. Computers are built from semiconductor devices of astounding complexity etched on incredibly tiny chips of silicon. The processes involved in designing, manufacturing, and testing these devices would be entirely impractical, if not impossible, without computer assistance. Yet by using computers, these processes are quite routine and, more importantly, inexpensive. The technique of using computers to make components for other computers has been a major factor in producing spectacular cost reductions in the semiconductor industry. Basic electronic circuits such as flip-flops, that two decades ago cost $10 to $100 apiece, today are available by the thousands for only a fraction of a cent each.
Computers are also used in the assembly, wiring, and checkout of other computers. It is quite common to see a computer being used to control the very machinery by which it itself was assembled.
The effect of this regenerative, almost reproductive, interaction between product and process can be seen in the performance versus cost characteristics of computers over the past two decades. The cost of computing power has dropped dramatically ever since the invention of the first computer. In the 1950's, the cost of large-scale computing machines was well over $1 million. In 1970, the same computing capacity was obtainable for less than $50,000, a decrease in cost of 2000 percent in 15 years. Figure V-1 shows the downward trend in the cost of mini-computers over the past decade. The price of new mini-computers has dropped at a rate of over 20 percent per year, and there seems to be no indication that the trend is nearing an end. In fact, it may be accelerating.
Figure V-1. The price of mini-computers has dropped at a rate of 20 percent annually for over a decade, and there is no indication that the trend is nearing an end.
Between 1965 and 1975, the cost of computing power in terms of bits per second per dollar dropped by approximately 5000 percent.18
During the past few years, it has become possible to put entire computer subsystems on a single chip of silicon. This technological feat has given rise to a whole new breed of devices known as micro-computers. Recently micro-computers have become commercially available for less than one hundred dollars,19 and predictions are that by 1978 these devices will cost less than ten dollars.
The impact of this spectacular reduction in the cost of computers has been enormous. First, it has revolutionized entire industries such as banking and insurance. Second, it has made possible technological achievements in space exploration, nuclear research, and electronics development, including home stereo and television equipment and pocket calculators that would have been inconceivable otherwise. But by far, the most important long-range effects have yet to be realized. These will occur when the decision-making and control capabilities of computers are fully applied to the basic manufacturing processes such as steel making, tool and die design, metal cutting, assembly, and inspection.
Today the cost benefits of mass production are achievable principally for items that are produced in quantities of many thousands per year. This is because expensive and complex machinery is profitable only when it is kept busy. Mass-production machines are highly specialized and capable of manufacturing only one product or, at best, a few different products of nearly the same type.
Computer-based automation, however, is flexible. The information concerning how parts should be shaped, where holes should be drilled, and how pieces should fit together are stored not in the mechanical structure of the production line, but in the easily changed memory of the computer. Computer-based automation can be switched from the manufacture of one item to another almost instantaneously. Machinery can be kept busy by manufacturing small amounts of many different items as well as large amounts of only one item. The prospect for the near future is that computer-controlled automation will make small-lot manufacturing as economical as mass production is today. Nathan Cook of MIT predicts that computers and robots may reduce overall costs in small-lot manufacturing by 80 to 90 percent.20
The implications of this for society are far reaching, not so much because of the direct impact on consumer products, since most of these are mass-produced, but because of the regenerative effects on the manufacturing process. The ma-chines used for mass production are expensive primarily because they themselves are produced in small-lot quantities. If computer-based automation were to reduce the price of small-lot manufacturing to the level of present day mass production, then the price of the machines used for mass production might drop by a factor of ten or more. Since the capital cost of machinery, together with material costs, almost entirely determines the price of mass-produced items, reduced costs in small-lot production would eventually be reflected in mass-produced items as well.
The revolutionary feature of this process is that it feeds on itself. Less expensive machinery makes the production of new machinery less expensive. When automatic factories begin to manufacture automatic factories, cost reductions will propagate exponentially from generation to generation. The introduction of computers into the manufacturing process thus has the potential for increasing productivity on a scale never before conceivable. Eventually the cost of finished manufactured goods may fall to only slightly above the cost of unprocessed raw materials. U this ever occurs, the expense of production will become virtually independent of the complexity of the manufacturing processes.
The self -regenerative properties of automatic factories are unprecedented in non-biological systems. Their potential impact is so overwhelming that the entire concept has an air of unreality. To some it may even suggest the notion of perpetual motion. It might be useful, therefore, to digress for a moment to point out the distinction between sell-regeneration and perpetual motion.
Sell-regeneration is the phenomenon that results when a complex organism like a living cell (or an automatic factory) uses energy and information to assemble raw materials into other organisms like itself. The secret of regeneration is the information that directs and controls the various steps in the manufacturing process. The most important of these steps is the duplication of the information itself so that succeeding generations can carry out the same procedures. Of course, the act of self-regeneration can also produce useful by-products such as the food and oxygen that are produced by plants in their self-regenerative activities.
In biological organisms, it is the ability of the DNA molecule to store information and control biochemical processes that enable the living organism to function and reproduce. In a similar fashion, it is the ability of computers to store information and control manufacturing processes that raises the serious possibility of self-regulating self-reproducing factories and industries.
The critical distinction between sell-regeneration and perpetual motion is in the supply of energy. Self-regeneration requires a net input of energy from the sun or some other external source. Perpetual motion proposes that useful work can be produced with no net input energy. Perpetual motion has never been observed to occur in nature and is considered to be theoretically impossible. Self-regeneration, in sharp contrast, not only is possible, but is commonly observed in nature and is, in fact, the process by which we all got here in the first place.
The fundamental scientific knowledge necessary to create sell-regulating sell-reproducing factories is already known.21 That is not to say, however, that major engineering advances are not still needed in order to actually build and operate such plants.22 The situation is similar in many respects to that which existed in 1960 in regards to sending a man to the moon. Now, as then, the basic scientific knowledge is available. There are no theoretical reasons why such factories cannot be built. All that is needed is a large-scale commitment of resources and manpower to a clear and certain goal.23
As was mentioned earlier, there have been a number of studies and reports that conclude that computer-controlled manufacturing systems are not only technically feasible but economically practical. In some areas automatic factories have already begun production, and others are in the process of planning, or construction. None of the available studies suggest that the cost of developing this technology for widespread use would be anywhere near as expensive as the Apollo moon expedition.
Among the critical items that still need further development are the computer programs for overall planning and scheduling, and the materials handling systems, or industrial robots for loading and unloading parts, changing tools, performing assembly operations, and inspecting for defects. 24
Computers and Robots
At present, most industrial robots are not computer-controlled. Instead, they use plug-boards, potentiometers, or plated-wire memory systems that restrict their capabilities to preprogrammed operations of a very simple nature.25 Industrial robots today are typically used only for such tasks as unloading die presses, spot-welding auto bodies, and picking up parts from one predetermined point and placing them in another. However, research is now being conducted in several laboratories in the United States and abroad into techniques whereby industrial robots under computer control can sense conditions in the external environment and adjust their own programs to compensate for misalignments and variations in dimensions.26 Computer-control techniques are developed that are similar to the muscle control circuitry in the human brain. These techniques will allow robots to be taught behavior patterns much in the same way that a child learns to use its own hands.27 Computer programs have also been developed by which robots can sense the environment, make logical deductions, and plan their own course of action based on their sensory input.28
Within a very few years, this research could lead to industrial robots capable of performing many, if not most, of the manipulative and assembly tasks that presently require human workers. Once these types of robot operations become practical in a factory environment, the primary function of human workers will be to set up production runs, program the robots, and then allow the machines to run under computer control. It will then be possible for factories to operate four shifts per week with only one shift of human labor. This fact alone would produce an almost instantaneous productivity in-crease of 300 percent.
Once industrial robots are capable of performing sophisticated machining, assembly, and inspection operations automatically, it is perfectly feasible for them to be used to construct other industrial robots just like themselves. This step would initiate a regenerative, or reproductive, process similar to that which exists in the computer industry. The result would be an exponential decline in the cost of industrial robots.
Today, the best industrial robots cost between $20,000 and $60,000 each.29 The addition of computer-control systems typically boosts this cost by another $10,000 to $30,000. This is far too expensive to justify investment for most present day applications, particularly if the robots are operated for only one shift per week. However, once these machines are installed in an environment where they can operate two, or three, or four shifts per week the economic picture changes dramatically. Present day industrial robots often show more than 50 percent per year return on investment for two-shift operation.30
If the regenerative effect of robots producing other robots would result in anything like the 20 percent per year reduction in cost experienced by the computer industry, the price of even sophisticated computer-controlled robots might fall to several hundred dollars within two decades. Such a cost, when prorated for a 168-hour week, would amount to an effective robot labor rate of only pennies per hour. Return on investment might exceed 100 percent or even 1000 percent per year.
These are facts that are either terribly frightening or tremendously exciting, depending upon whether computer-controlled robots and automatic factories are viewed as a threat to the economic value of human labor or as a potential source of wealth for everyone. Almost surely, if computers and robots are cast in the role of competitors to human labor, then human workers will lose just as surely as John Henry was eventually replaced by the steam drill. However, if the ownership of future automatic factories is shared by a large percentage of the population and if the wealth created by automated industries is distributed so as to increase the income of everyone, then the benefits of automatic manufacturing may completely eliminate poverty, not only in the United States, but throughout the entire world.
The range of possibilities is enormous. It stretches all the way from widespread hardship to unprecedented affluence. Unfortunately, the existing income-distribution system contains no mechanisms designed to prevent direct competition between robot and human labor. It is thus not surprising that there exists no public support for a major national effort to accelerate the pace of robot development. The productivity gains that could strengthen our nation economically and solve many of our problems of rising costs and dwindling resources simply cannot be vigorously pursued in a society where income is so overwhelmingly dependent on wages and salaries.
Lack of public support, of course, will not indefinitely delay the robot revolution. The technology will eventually develop due to simple market pressures.31 A second industrial revolution is certainly coming whether the average American wants it or not. The world economic system is structured such that automatic factories are inevitable. Other nations are making serious efforts to avail themselves of the unprecedented wealth-creating capabilities of superautomation regardless of what we do.32 Japan, for example, has already committed more than one-quarter billion dollars to research and development in computer-aided manufacturing and robot technology.33 Current Japanese plans call for the construction of a prototype automatic factory for the manufacture of machine tools to be completed by 1980. This plant is of a size that would ordinarily employ 700 to 800 workers, but will require only 10 persons to operate.34 If this prototype plant is successful, other automatic plants will be constructed immediately. It is entirely possible that within two decades Japanese machine tools will dominate world markets the way Japanese cameras and electronic products do today.35
Other countries are also aware of the potential economic benefits of robot development. Norway, Sweden, West Germany, and the Soviet Union all have vigorous research programs in robot technology and computer-based manufacturing.36 The United States is no longer the only technologically sophisticated country in the world. The Russians proved that with Sputnik. If we ignore this new technology or if we simply allow market forces in this country to provide all the incentives for its development, we do so at great peril.
Robot technology, like computer technology, has military as well as economic implications. Any country that develops the capacity to run its factories around the clock seven days per week with only a few human workers will have a tremendous advantage both economically and militarily. If nothing else, this capability would allow military weapons to be produced in virtually unlimited quantities at extremely low costs. But, even assuming that such plants were never used for military production, the country that possessed such a large surplus of efficient production facilities could easily dominate the world economically simply by selling manufacturing capacity at rates far below what countries using less efficient methods could hope to match.37
In many different ways, the development of machines that can create wealth unattended by human workers and, in a sense even reproduce themselves, has potential historical significance that is difficult to project. It is almost as if a new race of creatures were to visit the earth and offer to work at substantially zero wages and produce offspring in perpetuity. Such an event would be bound to have profound effects on human history at least as great as any scientific discovery or political revolution that has ever taken place.
Whether this event results in unprecedented benefits or economic chaos depends largely on whether we can devise satisfactory answers to the questions: "Who owns these machines? Who controls them, and who gets the wealth they create?"
These are questions that go to the very heart of the income distribution system. As long as we have a system in which only a tiny minority of the people own or control virtually all of the wealth creating capital stock, and the rest of the population must rely on selling their labor for income, we will have a situation where automatic machines and advanced technology will inevitably threaten the security and personal dignity of the average person. Only if we can devise a means by which everyone can share in the control of modern technology, as well as in the wealth that it creates, will the fantastic capacities of the coming generation of superautomation be released to assist mankind in solving the urgent problems of our society.