Will a machine ever take your job?

NORMAN DePOE October 1 1955

Will a machine ever take your job?

NORMAN DePOE October 1 1955

Will a machine ever take your job?


When push-button machines learn to push their own buttons, that’s called automation. "Thinking” machines are already running some factories and they’re not going to stop there. What’s it going to mean to Canada — new leisure or unemployment?


EVEN THE most casual of newspaper readers has recently become aware that something new —and big — is going on in North American industry. It’s already being described as the Second Industrial Revolution. The difference is that the first one used machines to do things—to cut, to spin, to weave, to shape, to lift — but always for a human operator. Now the engineers have taught machines to think for themselves, and eliminated the human hand on the levers.

These self-operating self-regulating machines, which can pour out goods better, faster and cheaper than ever before, are the tools of a new industrial way of life called automation, whose implications are frightening or roseate, depending which way you see it. This new technology, hailed as the gimmick that will liberate man at last from labor, and denounced as the final triumph of the soulless machine over human endeavor, will make radical changes in nearly every basic condition of our daily lives. The question is, what will those changes be? Will they spell paradise or chaos -unlimited leisure or nation-wide unemployment?

The process behind this suddenly fashionable catchword—automation is already making dramatic and sweeping changes in factory production. It has invaded scores of “non-mechanical” white-collar jobs. It has a hand in mailing out the Family Allowance cheques received by hundreds of thousands of Canadians every month. Thousands more of us, though we may not call it by name, already use automation to heat our homes or dry Monday’s wash. Automatic oven controls cook dinners for thousands of Canadian housewives. Many of the meal’s ingredients may be ready-to-use products prepared in the first place by other automatic machines.

To Norbert Wiener, the American mathematical wizard whose book Cybernetics and Society is the most popular text on automation, this continuing drive to perfect almost self-sufficient machines and invent new ones is an open invitation to an “abrupt and final cessation of the demand for the type of factory labor performing purely repetitive tasks.” And eventually, he predicts, we’ll be deep in unem-

ployment on a scale that will make the Depression of the 1930s seem like a pleasant joke.

This apocalyptic view is emphatically not shared by people like Benjamin Fairless, former chairman of the board of United States Steel. This giant company has moved into automation as fast as developments warranted it, sees the changeover as a kind of revolution, admittedly. But, Fairless told a business audience in Johnstown, Pa., recently that fears of mass unemployment were “just plain silly.”

What is the process that can produce such fiat contradictions? How does it work? Most of all, what is it going to do to us—and how soon?

The first thing that experts in the field point out is that automation is not new. And it is certainly not, as some people seem to think, ‘‘just a word coined recently for mechanization.” While it does include mechanization, and a lot of it, automation is a great deal more than merely using machines. And while it may not be new, it has recently picked up enough speed to qualify as a revolution both in appearance and effect.

Perhaps the simplest way of explaining the difference is to say that mechanization is a way of replacing muscle power while automation replaces not only muscle power but also (and the qualification is important) routine brain power. Mechanization can give you a factory where men only have to push buttons; with automation, most of the buttons push themselves.

The ideal automated factory of the futurestill a long way off for most industries—would be a place where raw materials (ordered by machines as needed) were delivered at one end, passed rapidly through a series of operations (all carried out by automatic selfregulating machines), and emerged at the other end as neatly packaged finished products. Inside the plant, there might be a dozen or so engineers, doing little but studying control panels and servicing machines which, by flashing a light or buzzing a buzzer, indicated they had broken down.

Science fiction? A pipe dream? More than ten years ago, during World War II, American scientists ran the giant Oak Ridge

Continued on page 62

Will a Machine Ever Take Your Job?


atomic installation from a single central control room. It was linked to ten miles or so of control panels requiring, on the average, only twenty human operators to the mile.

Or take a modern oil refinery, which comes within a hair’s breadth of the automation ideal. The raw material —crudeoil—arrives by automated pipeline. Its passage through the plant is regulated by pre-set controls, which can be adjusted to determine what the end products will be, and which maintain the right operating conditions at each stage in refining. Finally, the finished gasoline and other products are drawn off automatically to tank storage, to tank cars—or to another automated pipeline that will take them hundreds of miles without human help to a distribution point.

All this is accomplished with surprisingly few men in relation to the amount of complex machinery controlled. At the Winnipeg refinery of Imperial Oil, for instance, only one hundred and seventy employees can be counted in an area of four hundred acres. They process twelve thousand barrels of crude oil a day.

McKinnon Industries, of St. Catharines, Ont., moved this year into the forefront of automation with a new assembly line to turn out V-8 engines for General Motors of Canada in Oshawa. All of the basic work on the engine blocks—more than eight hundred separate operations—is accomplished with only twenty-seven men. In a few months, when further automatic controls are installed, the number will be cut to twenty-one.

These few workers are spread out along one thousand and twenty feet of massive and complex machinery which broaches, mills, grinds, hones, reams, taps, turns the blocks in any desired direction (including upside down) and positions them with pinpoint accuracy. Much of the human work consists merely of replacing tool bits when the machines signal they are wearing out. Most of the inspection, too, is purely electronic.

Even a breakdown doesn’t stir up much human activity. Special circuits in each control panel report trouble instantly to a central unit, and the whole line is electronically stopped until the bottleneck is cleared. Then the same impersonal switches and relays speed things back up to normal again. The line will turn out seventy fully machined blocks an hour —with the expenditure of less than a third of a man-hour of human labor on each.

But what about the men who tend the machines of automation—what’s their reaction to this revolution they’re shepherding? Like the theorists, some see it as a boon, others as a curse. "It’s a completely different job, and a lot better,” says Robert Handley, a wellmuscled twenty - seven - year - old St. Catharines man who, after six and a half years as a machinist, is now a toolsetter, electronic style. "For one thing, you’re not just running one machine all the time, doing the same thing over and over.” To run machines that utterly dwarf him, Handley had to absorb training in electricity and what he calls "efficient operations”—in effect, the theory of mass production. Today he has more responsibility and makes more money (the average operator’s bi-monthly basic pay envelope is

five to ten dollars heavier than those of conventional machinists).

To Handley, the future seems bright: "I guess I’ll stay here for good.”

But Gordon Lambert, a husky heavyweight worker in the foundry where the blocks are originally cast, isn’t so sure. "Now,” he says, "all I do is push a button on an automatic machine and take out the finished product. They’ve taken most of the skill right out of the job.”

The machine has also knocked Lambert’s salary down. Formerly, a skilled coremaker like Lambert, working under a system of incentive bonuses, could average $2.02 an hour. Now, mainly because the machine is independent of human skill—but partly because the union, as a matter of policy, demanded abolition of the incentive system—the job pays a flat $1.85.

Lambert, in his capacity as an official of Local 199 of the United Automobile Workers, says the union is dubious about other aspects of automation as well. There’s a general feeling that without production planning the machines will turn out a year’s requirements faster, resulting in longer layoffs. And such hard-won union demands as seniority rules appear to offer less protection to long-service employees than formerly.

Blacksmith, Make Me a Sword

The United States has traveled further and faster in automation than has Canada, though we’re not far behind. This Second Industrial Revolution was probably born in the U. S. even before the Industrial Revolution we all learned about in school. The man most frequently nominated as its father is Oliver Evans, a Philadelphia miller. He decided in 1784 to assemble three types of power conveyors into a single line, and the result was the world’s first continuous-process flour mill. The Jacquard loom, another early example, was invented in France in 1801. Pattern control was achieved by means of punched cards not too unlike the ones that shuffle through modern business machines by the thousand.

By and large, though, it remained for the twentieth century, and particularly for World War II, to advance automation to its present status as a revolution, an intellectual fad, a bogey man and/or a ticket to unlimited prosperity.

It has been made possible by such things as the invention first of the radio tube, and later a gimmick known as a transistor. It was powerfully spurred by the need for gun-aiming devices that could keep up with fighters streaking across the sky at four hundred miles an hour. Most of all, though, its present shape is due to a new mathematical theory of communications and control, largely worked out by Norbert Wiener and his associates at the Massachusetts Institute of Technology.

This, while it goes far beyond anything you could handle with highschool algebra, is simply a look at the way information is received and used. It applies to machines as well as men.

Let’s consider an old-fashioned blacksmith. Let’s say he’s told to make a sword. He heats his steel and hammers it into shape. Now, it obviously wouldn’t be very hard for a good mechanic to put together a gadget that would dunk a piece of steel in a furnace for a while, withdraw it, lay it on an anvil, and in so doing trip a steam hammer or die to do the shaping. There would be only one thing wrong with such a machine—a lot of the time it wouldn’t work because it could not receive, or use, vital information. If the furnace went out, the gadget

would blunder precisely on, banging away at cold metal. If someone left a doir open in midwinter, reducing the efficiency of the forge, the dies would fall on metal not hot enough.

The human blacksmith, on the other hand, sees or hears or feels evidence of trouble and corrects it. The rapidly growing vocabulary of automation long ago found a word for this process, which is the heart of the new system. The engineers call it "feedback.”

To bring automation to his sword making, the blacksmith could install an electric furnace as his forge and it could be fatted with a thermostat that would "know” if the furnace wasn’t hot enough and "tell” another control to feed it more power. (This, incidentally, is one of the commonest forms of automation. The odds are that you have a thermostat in your own home, controlling the oil burner in your basement. If you look around an oil furnace, you’ll find at least three other feedback controls designed to prevent vanous forms of trouble.)

Other instruments in the blacksmith’s shop could read the temperature of the metal precisely, and tell the machine exactly when to withdraw it. Electric eyes could position it on the anvil and a connected circuit would trip the shaping die only when the metal was perfectly centred.

It all sounds expensive—and it is. But there are large benefits for the manufacturer in it. If he wanted to make swords in quantity, the blacksmith who had installed automation could far outpace his unconverted colleagues. His machine would assure uniformity of quality. His power die would not get tired; it would take exactly the same time to shape each sword. His machine would never come in on the morning after pay day with a hangover, nor would it waste time mooning over the cute little robot in the front office.

Let’s leave our blacksmith’s shop and consider a more complicated manufacturing process—the sort of thing you’d find in the automated factory of today. What, for instance, happens when a machine can’t use feedback to actually correct an error? Automation engineers get around that problem by arranging to have the ailing machine turn itself off and yell (electronically, of course) for human help. Nowadays, these occasions are rare.

Since the automated assembly line must handle not one but many operations, controls for the] whole process have to receive many kinds of information from a variety of locations. They must be able to take any one of several different corrective actions, depending on how they add up the information received. And that’s just what they do—add up. That’s because most complex controls involve, in larger or smaller form, the device known technically as a digital computer and popularly as an "electronic brain.”

The brain, which is variously regarded as a sort of Frankenstein monster that could eventually enslave mankind, and as a fool gadget that can’t even predict an election, is in essence nothing more than a glorified adding machine.

That doesn’t mean it’s to be despised. The glorification process has gone a long way and has equipped the brain with some abilities that are more than human. The larger ones have "memories” capable of storing thousands of single facts; unlike human beings, they recall any one of them instantly and infallibly. They can be coupled to more-than-human senses (X-ray vision, for instance), and will compare new information from these with a wide range of stored information at many times the speed of the human brain. They

There’s even a machine that collects tolls—cheat it and it calls the cops

do most of this by a special kind of addition, using a special number system devised for the purpose.

The bigger computers can solve mathematical problems that are literally beyond human capacity. One such, in atomic physics, was handled recently by International Business Machines. It, and others like it, involve seventy-two million separate operations. A man working with pencil and paper might finish one in about eight hundred years. The new IBM 701 calculator can come up with the answer in two hours flat.

These brains obviously give real promise of fulfilling the dream of the completely automatic factory. But they can replace clerks as well as laborers, and they are already outshining routine human effort in offices as well as in factories all over North America. General Electric uses a giant Univac calculator to make up the weekly payroll for the twelve thousand employees in its plant at Louisville, Ky. The brain does the entire job. It adds bonuses earned, makes income and medical-plan deductions, figures overtime—all the things a payroll clerk has to do to a pay cheque in midtwentieth century. It distributes all totals among the cost accounts of the company’s various departments. Then it writes out a cheque for everyone concerned, prints a payroll register, and reports ready for the next job. The whole complex process takes less than six hours.

You’ll find evidence of a similar set of operations in the neatly punched holes that decorate one corner of Canadian Family Allowance cheques. But that’s only one type of electronic brainwork. Suppose a big company decides to double its production. As any executive can tell you, the effect on inventory is a lot more complicated than simply ordering twice as much of everything and arranging to have it delivered twice as fast. Working out a new purchasing schedule can take a trained staff weeks—but a computer will do it in hours. The U. S. Navy, in fact, is now using the IBM 701 to procure about two hundred thousand different aircraft parts, and schedule orderly delivery to sixty-five shorebased operations at home, plus four battle fleets operating throughout the world. The number of desk-borne sailors needed to equal the machine would run to several hundred.

On the civilian side, the Prudential Life Insurance Company has a computer that will bill policy holders for premiums, figure agents’ commissions, calculate dividends, and work out all the statistics on which premium rates are based. Officials estimate that the brain will take over the work of two hundred human employees in one department alone.

Several firms make electronic systems that displace not only elevator operators, but also the starters who used to control traffic. Computers have replaced clerks to register and allot train and plane reservations. One company recently demonstrated a robot toll collëctor for bridges. It will make change, count the number of cars going through and balance the cash at the end of the day. If any mere human being thinks he can drive past this electronic watchdog without paying, the machine slyly takes a photograph of his rear license plate and passes it along to the local constabulary.

From these examples, it’s evident

that automation may turn up almost anywhere. In most cases, the only possible limiting factor is economic: the employer or manufacturer must determine whether a routine operation can be repeated often enough to cover the original cost of the machine.

In spite of such things as automatic corn huskers and cotton pickers, farming is generally considered immune from complete labor displacement because of the relatively short season and the widely varied tasks that must be done. Complete automation is unlikely, too, in several small service occupations and in the manufacture of specialty products in limited quantities.

Inherent in successful automation is the absolute necessity for long production runs, plus an assured and preferably expanding market, if the enormously higher costs of the new machines are to be spread thinly enough to give us real benefits in lower prices. Conversely, any firm that misjudges the market, or whose sales staff can’t drum up business in sufficient volume, can easily face bankruptcy as a result of a decision to go automatic.

With huge amounts of capital tied up in machines, the fact that a breakdown anywhere along the line will tie up the whole line makes stoppages much more costly. In fact, automation demands a new way of thinking about manufacturing. It works best in continuous-process operations; it is with maintaining the whole process or "flow” that even specialized departments must now concern themselves. (This, incidentally, is a major reason why the oil industry is further into automation than any other. When you deal with actual liquids, a "flow” concept of manufacture is a natural.)

Everything May Be “Free”

It may not be enough to automate only the factory. For some industries, this would be something like putting an Orenda jet engine into the Wright Brothers’ original airplane. The rest of the business—planning, distribution and marketings—may have to be largely automated, too. Certainly, these functions will at least have to be thought of, and treated as, extensions of the continuous process going on in the factory.

One suggested consequence of all this is that North Americans may have to be unsold on the rapid style changes we have learned to love in our gadgets. A five-year freeze in design of, say, washing machines would make automation really pay off in lower prices. Some companies are already casting about for new methods—and among them, believe it or not, is the idea that the way to an assured market might be to give appliances away free.

Well, not exactly free. What the consumer would buy would be a longterm "service contract,” probably calling for monthly payments. The "service” might be almost nonexistent, but his contract would call for automatic "free” replacement of his washing machine every three or four years with a later model. This would allow maximum use of automation for threeto four-year production runs and maximum careful planning to cut costs.

The mere thought of introducing total automation into industry without thorough planning first being done, has drawn cries of alarm from North

American labor leaders. CIO president Walter Reuther, testifying before a congressional committee this year, demanded that the United States government do some planning immediately before the country drifted "aimlessly into dislocations and disruptions, mass unemployment and catastrophic depression.”

In Canada, a similar demand was made last June in the House of Commons by Colin Cameron, the CCF member for Nanaimo. Citing the example of a new seventeen-million-dollar pulp-and-paper mill in his own riding which employed only one hundred men, he called on the Department of Labor to make a full-scale survey of automation’s growth and impact. Social Crediters agreed with him that the new system held disturbing possibilities.

The debate, which was inconclusive, also brought out the other side of the argument—that automation can only lead to a higher standard of living. William Hamilton, a Montreal Conservative, even had a brush with Mr. Speaker when he described critics of the new age as "barnacles on the backside of progress.”

Outside the House, the optimistic view was summed up in the same month by J. R. White, president of Imperial Oil. "Automation,” he claimed, "does not destroy jobs. Instead, it creates them.”

Both sides in the running controversy can buttress their arguments with figures. Labor leaders point to the U. S. steel industry, which is turning out as much steel this year as it did in 1953—and doing it with seventy thousand fewer men. In the auto industry, though cars are being turned out at a record rate, the American membership of the UAW has not risen appreciably in five years.

The unions also point out that many skilled craftsmen are being replaced by machines. In the packing industry, the removal of hides from slaughtered animals used to demand high manual dexterity. Now an automated system largely developed in Canada does the whole job after a semi-skilled worker makes a single incision. Where it formerly took ninety-eight top-rated hide strippers to skin 110 steers per hour, the latest figure at Canada Packers in Toronto is forty-seven men to maintain the same rate.

In manufacturing, highly trained die cutters are losing jobs because of machines that will make as many perfect copies as you want from a single master die. Even the master die may not be cut by a human being much longer. There’s a million-dollar machine being installed at the Convair plant in California which is capable of eighteen different machining operations. All that’s needed is a blueprint of a newly designed part, which no one has ever made or even seen. Engineers can punch out instructions on a tape which will tell an electronic brain what tool strokes are required to make it. Unlike most other automated equipment, which is single-purpose, this machine can switch from one kind of part making to another as fast as instruction tapes can be changed.

Electronic brains have also moved in on the baking industry, with taped formulas for turning out bread, cake, or pretzels—including bending them. In addition, thousands of housewives are using ready-mix baking products which are blended and packaged by automated equipment. The whole development is tolling a louder and louder knell for the old-fashioned baker.

Multiply cases like these by scores, and the alarm expressed by union leaders over the dangers of advancing into an automated age without adequate industrial planning is under-

8tandable. Some big businessmen, on the other hand, quote figures that tell a different story.

E. H. Walker, president of McKinnon Industries, points out that his company increased its employee roster from 710 in 1929 to 5,416 last spring —and that over the same period, more and more machinery, some of it automated, has been used. R. M. Robinson, general manager of Canadian General Electric’s electronics division, says that in spite of such developments as printed circuits and automatic soldering, staffs in CGE’s electronics plants have doubled over the past five years. And in office accounting, where machines have made their most dramatic advance, U. S. figures show seventy-one percent more people at work in 1950 than there were in 1940.

How can these apparently contradictory figures be reconciled? There are three main factors which tend to explain them, though their meaning is hotly disputed. And always it is to be remembered that they apply only temporarily—to the twilight period between non-automation and full automation.

First, and simplest, is the breakneck rate at which North Americans reproduce themselves. The birth rate, plus immigration, creates a steadily expanding market. During recent years, while machines more than doubled the output of our meat-packing plants, the population of Canada grew by two million. Automation merely kept pace with new demand; there were no major layoffs. "Eventually, though,” warns S. S. Hughes, assistant director of the United Packinghouse Workers, "you reach a saturation point in this kind of thing.” Machine productivity is still growing—and there’s a limit to the amount of beef a person can eat.

A second factor is invention. Thousands of people are hard at work today making things nobody wanted—or had even heard of—ten years ago. Businessmen also point to the fact that every new product tends to generate new subsidiary jobs. Two favorite examples are television, with its hosts of servicemen, and the automobile, which has put thousands of people to work doing everything from road building to running gas stations and designing voluptuous radiator caps.

The third factor is automation itself. While it replaces unskilled workers and some manual craftsmen, it demands specialists. Right now automation faces a shortage of engineers in particular. It will need them by thousands over the coming decade. At the University of B. C. last spring, it was reported that ten jobs were available for every man graduating in applied science.

All the experts agree that this is one of the major shifts caused by automation. There may be a shrinking demand for unskilled workers but there are new—and more highly paid—opportunities for men with a technical and scientific education. This immediately suggests that our high schools may have to abandon their current de-emphasis on "hard” subjects like algebra and physics if graduates are to have enough background for later job training. And if the Second Industrial Revolution speeds up (which it gives every sign of doing), we may need large-scale retraining programs for people already at work. To this end some companies are already conducting courses for their present employees.

But there remain unanswered questions. Why, for instance, should we bother with automation at all if we end up with just as many people working, though at different jobs, as there were in the first place? The answer to that is that even if we

re-employed all our present working force, the new machines would still turn out goods so much faster and with such uniformly better quality that we’d have a vastly higher standard of living.

Well then, what about this question of re-employment: Will the new de-

mand for technicians and draftsmen balance the shrinkage in unskilled labor? In any one industry, the answer seems to be "No.” Will the growth of small subsidiary service organizations that don’t need automation mop up the people displaced? To that one you’ll find as many opinions as there are experts.

The biggest question remains: Will we re-employ all our present working force and future additions made to that force via the birthrate? The experts part company sharply in hot dispute. They agree only that automation is going to cause dislocation—jobs disappearing in one place and opening up in another.

A U. S. analyst of statistics, Peter Drucker, after studying the effects of population growth, predicts an actual scarcity of labor over the next twenty years. J. A. Calder, past president of the Canadian Manufacturers’ Association, sees a "period of readjustment,” after which everyone will be "resettled in gainful employment.” On the other hand, a massive new study of the American economy forecasts a fiftypercent rise in unemployment in the United States by 1960.

Much of the unemployment—if it comes—will be what critics of automation call "hidden.” It’s hard to find many cases where automation has resulted directly in a layoff. The machines move in one by one and, since there’s always some turnover in the labor force, people who retire or girls who leave an office to get married are

simply not replaced. An unskilled worker quits one job and finds it difficult to get another because machines have moved in on his field. Or a young high-school graduate finds that none of his education has fitted him for the skilled jobs which seem to be the only ones available. Even where a completely new automated factory —an addition to the nation’s previous capacity—is opened, fewer new job opportunities are created than was formerly the case. As one critic says: "It’s not the people who are laid off. It’s the ones who just aren’t hired.”

Some prognosticators have gone so far as to paint a future in which there is ever-increasing productivity and prosperity for those with jobs, while the new machines squeeze a few more people out of the golden circle ¿very year. Every worker displaced, they point out, is also a consumer removed from the market automation needs.

On this point, M. M. MacLean, acting Deputy Minister of Labor in Ottawa, says: "Our statistics ... do not distinguish between technological unemployment and unemployment due to other causes. The statistics for recent years, however, do seem to show that the most persistent single cause of

unemployment in Canada has been the seasonal reduction in activity in many industries each winter, and that such seasonal unemployment is much greater than any technological unemployment which is likely to have existed recently.”

MacLean says that the Labor Department is constantly studying employment levels and has recently given some thought to automation.

"At present,” he points out, "we do not know how rapidly automation will be introduced*into the various industries, or what the effect will be on employment, or whether [ it j will he serious enough to call for special measures, such as re-training programs. Since the problem still lies chiefly in the future (although perhaps not very far in the future), we have no statistics at present which throw much light on it.”

The department concedes that the effects of automation may be different in some ways from those of earlier mechanization, and that they may come more rapidly and thus create greater problems. "We have so far,” j MacLean concludes, "very little to go on.”

Labor unions, on the other hand, have already made up their minds: they welcome anything that makes labor less laborious but, they insist, there must be the most careful planning in introducing the new technology to industry so that dislocations are kept to the minimum. Ted Silvey, educational director of the CIO, told a conference this summer that his organization "has definitely taken the position that automation is a blessing.” He added: "We must plan to use it

so that it blesses mankind.” The unions see no reason why industry should not continue along the same road it has followed ever since machines first began to extend human output. Union plans are to take their slice of the new productivity in the form of better pay and security (the guaranteed annual wage is a current example) and in more leisure through shorter hours.

This was clearly indicated in miniature this year at the Holmes Foundry in Sarnia, where engine blocks are turned out for Ford of Canada. Automation raised output from 600 engines a day to nearly 1,000 —and simultaneously, according to union figures, the labor force was cut from 467 to 260. The UAW promptly demanded a cut in the work week from 48 to 40 hours. After a three-month strike, it got it, with pay equivalent to 44 hours at the old rates.

For the future, George Burt, Canadian director of the United Auto Workers, says, "The thirty-six—or thirty—hour week will be one of our next objectives.”

How soon it will come, Burt isn’t prepared at the moment to guess. It will all depend on the effect of automation and how fast it makes itself felt. Like other major changes in UAW policy, Burt says, this one will probably start in the United States "and we’ll be right behind our American brothers.” The other unions in the CIO-CCL "big three”—the Steelworkers and the Packinghouse Workers—are also agreed on shorter hours.

The central dilemma of automation can be summed up in a single anecdote, possibly apocryphal. Walter Reuther, the story goes, was touring one of Ford’s new automated plants with a company executive. Looking at the machines, the executive remarked, "You’re going to have quite a time getting them to join your union.”

"Yes,” Reuther is supposed to have replied, "and you’re going to have quite a time trying to sell them cars.” ir