Machining to CNC: How Automation Redefined Skills and Wages

Machining, historically a blend of art and science, has undergone significant transformations with the advent of Numerical Control (NC) and Computer Numerical Control (CNC) technologies. These advancements have sparked debates about the "de-skilling" of craftsmen, a topic that warrants a deeper exploration.

The Beauty of Machining

Machinists traditionally operated powered machine tools to cut, shave, or grind metal, a distinct process from casting, forging, rolling, or stamping. Their role extended beyond mere metal cutting; they were planners, interpreting two-dimensional blueprints and visualizing them as three-dimensional parts to anticipate and prevent issues. This required a strong grasp of geometry, trigonometry, and algebra.

Beyond theoretical knowledge, machinists relied heavily on "feel." They could sense minute changes in dimensions, as small as half a thousandth of an inch, and adjust machine speeds and feeds based on the material and cut. An 1887 book, "The Complete Practical Machinist," highlighted the intricate and deceptive nature of lathe work, emphasizing the need for judgment, perception, and watchfulness. An untrained eye or ear might perceive a tool as working correctly, while a skilled machinist could detect underperformance or erratic operation that could lead to malformed products or injuries.

Machining was also a physically demanding and dangerous profession. The shop floor was often noisy, smelled of oils and greases, and was littered with metal chips, posing risks of severe injury. This challenging environment fostered a distinct masculine culture.

Becoming a Machinist

In the first half of the 20th century, machining was a leading skilled manufacturing job for American men. The most common path to becoming a machinist was a formal apprenticeship, typically lasting four years and 8,000 hours, or five years and 10,000 hours for tool and die-makers.

One apprentice at GE in the 1940s described his training: six months of basic shop training combined with three nights a week of classes covering shop math, college physics, blueprint reading, and electricity. The subsequent 3.5 years involved rotations on various tools, starting with the drill, for about six months each. Upon completion, apprentices were certified as "journeymen," receiving tools or a cash bonus, a pay raise, and, notably, becoming qualified for marriage.

However, this apprenticeship route wasn't universal. Some machinists were self-taught, gaining expertise through shop experience and home study. There were also courses, akin to modern coding boot camps, designed to teach machine shop skills. Additionally, some individuals became "semi-skilled" tool operators, specializing in a single tool with less formal education and more on-the-job training.

Machinists found employment in various shops, from "job shops" producing small runs of specialized items to large production shops manufacturing high volumes of identical products.

The Evolution of Tools and Skills

In the mid-1800s, machinists, like other craftsmen, often made their own tools. Arthur Briggs Farquhar, an American industrial magnate, recalled in his 1922 memoir that machinists of his youth needed to perform almost any type of metalwork, including chipping and filing to high accuracy, and were expected to make and maintain their own tools.

Industrialization, however, brought about a shift. New automated tools changed the relationship between workers and their craft. Factory workers no longer owned or made their tools, nor did they dictate their work pace. As specialized machinery emerged, incorporating the skills of craftsmen, jobs became simplified and segmented. The need for a few highly skilled craftsmen was replaced by a larger workforce, each trained for a specific, narrower task. For example, riveting a boiler, which once required nine skilled craftsmen, could be done by one skilled worker and eight unskilled laborers with a hydraulic riveting machine.

This specialization, while economically efficient, drew criticism for potentially destroying invaluable craft knowledge, cheapening labor, and creating a divide between those who conceptualize and those who execute. In 1883, machinist John Morrison testified to the US Congress that the trade was becoming increasingly subdivided, reducing some roles to "laborers' work." Numerical Control, in this context, represented another step in this historical progression, but its integration of metalworking and data promised a more radical transformation.

The Rise of Numerical Control (NC)

Numerical Control (NC) technology uses pre-programmed instructions to guide machine tools without direct human intervention. Machine programmers translate drawings into coordinates, which are then fed into an NC controller via paper or plastic tape.

Invented in the 1950s, NC took time to gain widespread adoption. Early versions were primarily used for Air Force and aerospace applications, often exceeding the needs of commercial and smaller job shops. Initially, shop owners hoping to "de-skill" their workforce with NC were disappointed. The "hard-wired" technology was inflexible, with limited functions and difficult upgrades.

Surveys from the mid-1960s indicated that NC still required operators with significant skill and technical knowledge, especially during initial implementation. The high risk of costly downtime meant that top staff were assigned to operate these machines.

NC's Significant Advantages

Over time, NC technology improved, revealing significant advantages in producing complex parts with high accuracy. It offered substantial savings through reduced rework, faster production, and shorter setup times. One NC drilling machine could replace three conventional machines, and one NC milling machine could replace two or three older ones, leading to higher utilization rates and better return on investment.

NC also produced higher quality workpieces, enabling the creation of parts previously impossible with conventional equipment. Batch production became more economically feasible, allowing for the efficient manufacture of small to medium quantities of items.

Further advancements saw NC tools combined with robot arms to create industrial cells, where robots handled loading and unloading, servicing multiple tools in rotation. These cells, often chained together with conveyor belts, enabled companies to achieve both mass production and product flexibility, a concept pioneered in Britain and popularized in Japan.

The NC Deskilling Debate

These changes, initially slow, accelerated throughout the 1970s with improvements in computer technology. By 1980, 25-30% of machine tools sold in developed economies were NC-enabled, though the installed base was lower. The benefits, however, pressured companies to adopt these expensive machines.

A key source of productivity gains came from increased utilization and reduced labor. Work shops could replace two or three experienced conventional machine operators with fewer NC operators. While NC operators required training, it was significantly less than the four to five years needed for an apprentice machinist. A 1982 paper on Swedish machine shops noted that NC operator training could be completed in 6-12 months. One firm reported hiring 22 NC operators instead of 44 lathe operators, and another replaced 63 qualified machine tool operators with just 21 NC operators.

Shop owners found that NC allowed for better tool utilization, especially in developed countries like Sweden where finding skilled labor for night shifts was challenging. NC operators could work unconventional hours, keeping machines running.

The perception that the "NC operator" job was less skilled than that of an old-school machinist was widespread. A 1980s survey of Canadian manual machinists found that nearly 80% believed NC machining required "less skill." Those who had actually operated NC machines were more divided, while managers and NC programmers generally leaned towards the "less skill" argument, though less emphatically.

From NC to CNC

The 1970s saw the emergence of Computer Numerical Control (CNC), enabled by microprocessors. CNC allowed for the creation and editing of instructions directly within a computer. This development surprisingly redistributed some design power back to operators. Early NC systems created a clear division between the programmer in the office and the operator on the floor, with the operator primarily responsible for setup and supervision.

CNC machines, with their manual data input controls via keyboards and screens, empowered operators to edit designs, fix issues, or even create their own designs. This didn't require extensive programming knowledge, and machinists who learned these skills found them complementary to their existing conventional machining expertise.

Bill Bowman, a machinist who experienced the CNC revolution, recalled that employers increasingly expected machinists to program their own machines. Learning these new CNC skills boosted his wages; in 1989, he earned about $30,000 annually (approximately $81,000 today) for CNC-enriched work.

Wages and Skill Shift

Despite the perception of de-skilling, wage data suggests a more nuanced picture. A 1989 study comparing 1981 Bureau of Labor Statistics data found that Class A machinists earned $9.72 per hour, while NC operators earned $9.51. Class A machinists still earned more than Class B and C machinists, who handled more repetitive tasks and earned $8.54 or $6.41 per hour. Interestingly, some NC operators were recruited from Class B and C machinists, resulting in a pay raise for them.

However, wages are a broad metric influenced by factors like location, unionization, and work-life quality. Bowman's experience further illustrates this complexity: as more people learned CNC skills, his wage boost compressed, and years later, despite doing more CNC programming, he earned $20,000 annually (about $9 an hour) after layoffs.

This suggests that while NC operators may not be "unskilled," there was a shift in required skills, with deep mastery in one area being de-emphasized in favor of other competencies.

System Supervision: A New Skill

One of the new skills that emerged with the NC revolution was the ability to oversee and supervise these complex systems. Proper machine monitoring required an abstract understanding of how the system worked. When breakdowns occurred, operators needed to diagnose the problem, which was often not immediately obvious.

As one operator explained, diagnosing a breakdown involved thinking, retracing the incident, identifying signs of failure, and understanding the underlying logic, rather than just pressing buttons. This highlights the continued need for practical machining knowledge, even in an automated environment. A mid-1990s survey of Taiwanese CNC operators found that 80% had experience in both machining and programming.

A Hurco operator in a US field study described how a drill working too hard would "let you know" through sound and finish, prompting a correction in the program (slower speed or larger spot). This demonstrates how classic machining instincts were still vital, albeit expressed differently.

Furthermore, NC operators needed to understand how various tools interacted within the overall system. Increased automation and interconnection led to more complex, layered problems, requiring collaboration with other experts for debugging. While the job shifted from tactile, physical work to more monitor-based supervision, reviewing parameters and responding to messages, the underlying need for skill remained.

A 1991 study by Jeffrey Keefe, analyzing 30 years of Bureau of Labor Statistics surveys, found that overall skill levels declined by only about 1%, essentially zero. However, this average masked significant changes in job composition, with some roles disappearing and new tasks migrating to NC operators, creating considerable turbulence in the workforce.

Conclusion: Parallels to AI and Programming

The evolution of machining with NC and CNC offers fascinating parallels to current developments in AI and programming. Early AI coding tools, like GitHub CoPilot, focused on autocomplete and line insertions. More recent AI coding systems are becoming "agentic," a level of autonomy not yet seen in modern CNC machines.

A February 2026 article by Ryan Lopopolo at OpenAI, detailing a product built by a team of AI agents monitored by humans, resonates with the historical shift in machining. This raises questions about the future of programming, currently a dominant high-paying job, much like machining once was.

While programmers will likely still need to read and judge code to diagnose problems, the method of fixing them might shift from directly editing code to instructing an AI agent. This suggests a change in the expression of skills rather than a complete loss. However, skills will undoubtedly migrate, some jobs will decline, and new fields will emerge, creating an "NC operator of the AI coding era."

The constant takeaway from this historical analysis is the necessity of continuous learning, as exemplified by Bill Bowman's experience. Change is inevitable, and adapting to new technologies and acquiring new skills will be crucial for navigating future transformations in the workforce.