Through our history curriculum, Commonwealth students learn how to be historians rather than passive absorbers of facts. They read and analyze primary sources from a variety of voices in their historical contexts. And they graduate knowing that a single textbook could never tell the whole story. Core to this discovery process is an annual research paper that hones students’ analytical and writing skills and challenges them to dig deep into a variety of sources on any topic of interest, whether it be debunking the “primitivist” myth around jazz, the impact of the breakup of the AT&T monopoly, or, as you will see in this recent example, where Siri '27 finds parallel fears over artificial intelligence in computer science today and the "automatic programming" introduced by Grace Hopper in the 1950s. In both eras, women programmers were—and are—the most vulnerable to extinction.

In March, 2026, Anthropic published a paper investigating the potential impacts Artificial Intelligence could have on the labor market. The findings were frightening: several high-status professions which had previously been seen as highly skilled and therefore relatively “safe” jobs were now in danger of AI-automation. 

Take, for instance, computer science, which for years had been considered one of the safest majors. Now, the Anthropic study warns, computer programmers are the most exposed to AI automation, with a whopping 74.5% observed exposure. While this is a rather shocking turn of events for those told that programming was safe from automation, it is actually not the first time programmers have faced an automation scare. In the 1950s, the field of “automatic programming” presented a similar threat to the programming profession. In particular, Grace Hopper’s invention of the “compiler,” which translated high-level human-readable instructions into low-level machine code, was viewed with the same mixture of fear, excitement, and skepticism with which we now treat AI. Hopper deftly navigated the anxieties and desires of the programming and business communities, promising a world in which programming could become more accessible without taking away existing jobs or devaluing the status of current programmers. Ultimately, the compiler had positive and negative implications for the programming community. For the most part, the computing industry grew, and the programming community thrived. However, just as the 2026 Anthropic study now predicts that workers in the fields most exposed to AI automation are “more likely to be [women],” the rise of automatic programming had the most disruptive effect on female programmers as the programming field became increasingly masculinized over the 1960s and 70s due to the elevation of the profession. As is often the case, new technologies like the compiler prompt cycles of “creative destruction”—an economics concept used to describe how new products and technologies can lead to economic growth and increased job opportunities while simultaneously having the unfortunate effect of disrupting certain industries and displacing certain groups.

When computers emerged in the 1950s, headlines began to discuss the potential consequences this new technology could have for American society. One New York Times article from 1950 predicted the dawning of a new “Robot Age,” in which machines would “think and do the hard work,” thus “free[ing] men to develop their real talents.” The author welcomed the rise of the “mechanical brain[s]” with enthusiasm—optimistically predicting that this new age would bring about a paradise, “carefully tended by robot workers,” in which even the weather bowed down to humanity and its machines: storms would darken the skies “only when forecast by robot calculator, and perhaps only when called for by robot weather service.” In this utopian future, there would be “no threats of war” or concerns over the “burning [of] irreplaceable geological resources” for energy; instead, humans would be free to live “abundant and secure li[ves], with their wants supplied cheaply by robot factories.” It was an admittedly “rosy” vision of the future—the same article also recognized that machines had the potential to bring suffering, “endanger[ing]” peoples’ lives through the use of “robot weapons in warfare,” or threatening Americans’ “jobs and pocketbooks” through the automation of industry. It is perhaps unsurprising that following the bombing of Hiroshima and Nagasaki, many were wary of new technologies “falling into the hands of ‘irresponsible militarists,’” or that after seeing the ability of the machines to do such “diverse chores as preparing utility bills, keeping tabs on inventory for stores and mail order firms, figuring insurance premiums, and making airline reservations,” many people were fearful that the computer would come for their own jobs next. 

Indeed, as cartoons from this period illustrate, feelings towards computers were mixed. During the 1940s and 50s, there were intense debates over whether computers were truly capable of automating human intelligence. One 1946 Captain Marvel comic titled “Captain Marvel and the Incredible Calculator” embodied this dispute, featuring a competition between fictional mathematician “Professor Sneeyer” and the titular “thinking machine.” In general, the comic seemed to land on the side of the machine: the computer was capable of calculating a complex problem which took Professor Sneeyer “almost 8 months” to complete in a mere “ten seconds,” proving that machines could replace even distinguished mathematicians with their rapid calculations. Even in the second round of competition, in which the machine at first appeared to erroneously predict that an impending storm would “cause no damage” to the coast, the machine ultimately won out on account of its omniscient knowledge that Captain Marvel would step in to save the town:“I knew that the storm would hit the coast, but I also knew that Capt. Marvel would fly there and save the farmhouse!” While it was ultimately a resounding victory for the all-knowing machine, the comic was not entirely enamored with the “incredible” capabilities of computers. At one point, a mouse caused the machine to short circuit, almost killing Professor Sneeyer, illustrating how computers were seen as finicky and unpredictable. At other points, the computer struggled with seemingly simple tasks—for example, it initially refused to acknowledge that Captain Marvel existed, even though it was seemingly obvious to everyone else in the room. And, of course, although the computer was personified throughout the comic, it also seemed to have little regard for human life during the prediction of the storm’s effects—what would have happened if Captain Marvel was not there to save the day?

Another cartoon from 1952 explores this growing skepticism and distrust in the computer in greater depth. In this depiction, the computer is a giant box, haphazardly adorned with a smattering of buttons and levers. A group of scientists huddles around the gargantuan machine in the secret laboratory, excitedly declaring their success with wrangling the giant “electronic brain”: "We've eliminated the last 'bug', gentlemen... the human element... it'll now do an income tax return without blowing a fuse!"

Computers were often portrayed as miracle-machines, with mysterious, inner workings. This is evident in the cartoon’s signage, which issues vague warnings like “RESEARCH KEEP OUT” or “CAUTION: BRAIN AT WORK.” However, many were beginning to grow suspicious that the machines were really just a fragile jumble of parts—for example, the cartoon features random instructions to “oil” the computer using “alphabet soup” or to avoid “ask[ing] the brain what’s going on in Korea.” These details suggest that the machine was so fragile it might combust if its strange, fussy setup were altered. Indeed, as the caption suggests, the computer was so delicate that it struggled to complete even the most basic tasks, like computing “an income tax return without blowing a fuse.” These sentiments were not limited to comics and cartoons: in a similar fashion, many papers were less fearful and more skeptical of the hype surrounding the computer, warning that despite the frequent humanization of machines by the press, in the end, computers “d[id] not ‘think’” and should only be considered as “tools”—useful, time-saving ones, but certainly not magical “wizard[s]” or living beings capable of “sensing” and “thinking,” as some of the more colorful articles claimed. Some even went so far as to declare the hulking computers “moron[s],” which “balk[ed] at adding two and two”—a strange balance between humanizing the computers and yet still denigrating their intelligence. Computers had entered the public consciousness, and people were unsure exactly how to react—should they be fearful or hopeful, emotional or impassive? Would machines improve society, or did they spell humanity’s doom?

This was the climate in which Grace Hopper marketed her new invention—the “compiler”—which she hoped would deliver on the papers’ hopes that the computer would eventually free man from drudgery. In her 1952 paper, titled “The Education of a Computer,” Hopper claimed it was her goal to “replace, as far as possible, the human brain by an electronic digital computer”—particularly in the case of computing, which was, in fact, not “wizard[ry],” but the result of large amounts of human labor. Programmers not only had to carefully map out the mathematical solution to a problem, but also tediously translate the solution into machine code. Early on, this labor was divided into two professions: the “programmer”—a scientist or mathematician who did the high-level problem-solving—and the “coder”—a clerical worker who did the manual labor of inputting the programmer’s instructions to the computer, which could involve manually adjusting switch and plugboard settings, or encoding the instructions onto paper punch cards or punch tapes which would then be run by the computer. Hopper herself recognized the distinction between these two types of labor during the 1954 Symposium on Automatic Programming for Digital Computers: “A ‘programmer’ prepares a plan for the solution of a problem,” she explained, whereas “the coder [...] reduce[s] this flow chart to coding.” Since coding seemingly required less skill, coders were lower-ranking compared to programmers, and were often female, whereas most programmers were male. As Jennifer Light notes, the work of female coders was frequently diminished—male lab leaders often took credit for their labor, leaving the women merely “defined by their lab leader [...] or by their machines” rather than their own individual contributions. Despite their low-level designation, many of these coders increasingly took on much of the work that was supposedly designated to the high-level programmer. In the case of the “ENIAC girls”—who did ballistics computations during World War II—female coders had to become intimately acquainted with the mathematics and hardware of the machine in order to do their job correctly. For example, Betty Jean Jennings, a leading coder on the ENIAC project, claimed that the girls “learned to diagnose troubles as well as, if not better than, the engineer.” Similarly, the role of the high-level “programmer” had become increasingly entrenched in the coding process. As Grace Hopper noted in “The Education of a Computer” (1952),  the programmer had become burdened by the “dull labor of writing and checking programs.” Because of the intertwined nature of programming and coding, the division between “programmer” and “coder” had become increasingly blurred. 

With the invention of her first “A-0” compiler, Hopper hoped to free the programmer from the menial task of translating flow charts to machine code by relegating the labor to the computer itself, allowing the programmer to “return to being a mathematician.” This essentially took the low-level coder out of the equation: “It is this function, that of the coder, time-consuming and fraught with mistakes, that is the first human operation to be replaced by the computer itself,” Grace Hopper declared in the 1954 Symposium on Automatic Programming for Digital Computers. Hopper did not mean to present automatic programming as a job-stealer, though she was excited by the prospect that “computers [could] do [certain] job[s] for us.” Rather, much like the idealistic newspaper articles which hoped computers would “free” man, Hopper was optimistic that automatic programming could be the “answer to a programmer’s prayer.” The programmer’s job was quite safe as far as Hopper was concerned—compiling routines “only eliminate[d] the clerical part of the job,” whereas the programmer’s job—“the process of analysis”—“[could not] be assisted by the computer itself.” In fact, Hopper was reluctant to describe her work as “automatic programming”—instead, she preferred the term “automatic coding,” which made it clear that high-level programmers would not be subject to automation or devaluation. Tying back to the original goals expressed in her “Education of a Computer,” Hopper believed that the compiler could be used as a tool to enhance the programmer’s abilities and free them from tedium—its primary purpose was to “assist[...] the programmer,” not replace them. 

While Grace Hopper had argued that high-level programmers would be protected from automation, she made no promises for the safety of low-level coders—a group which was largely female. As mentioned earlier, the distinction between “clerical” coding and “intellectual” programming had gendered implications. While women had gained access to a wider range of jobs during World War II as a result of the draft and the subsequent shortage of male labor, many of these newly accessible occupations still inherited gendered stereotypes. In the general job market, women were still deterred from high-status, technical occupations—even into the 1960s, American job advertisements were separated by gender, with a stark division between the professions offered to women and men. As female programmer Jean Sammet described it, “under [the] 'Women's' [section] there was no such thing as engineer, mathematician, programmer, or anything like it. It was clerks, teachers, housekeepers [...] So I learned to look under the 'Men's Help Wanted.' [section]...” Within the programming field, women were often steered towards lower-level clerical jobs, in keeping with traditional gender hierarchies. For example, even when women scored higher on the “aptitude tests” which were used in the hiring process in the 1950s and 60s, William H. Enneis of the U.S. Equal Employment Opportunity Commission noted that companies often prioritized upholding these gender roles over following the results of the tests: “For example, in one company all the keypunch operators were women and all the programmers and computer operating personnel were men! This was so even though the mean scores for the women were higher than those for the men on each test being used to assess aptitude for computer work.” As a result, many women occupied clerical coding jobs, and were thus vulnerable to replacement by Hopper’s “automatic coding.”

Although coders were the ones truly threatened by Hopper’s “automatic coding,” many programmers were still fearful that the compiler would threaten their highly-coveted positions. Over the course of the 1950s and '60s, programmers had found increasing power and status in corporations as computers began to gain traction in business applications. Due to the rising importance of computers in business, coupled with the relative scarcity of “skilled” programmers, programmers enjoyed high salaries and special status within their companies. Moreover, since programming was viewed as a secretive “black art”—a mysterious craft which required innate creativity and genius—it was generally accepted that programming was “the kind of work that is called creative [and] creative work just cannot be managed,” so programmers were given a high level of independence. The problem with the compiler—specifically, Hopper’s second “B-0” business compiler—was that it threatened the programmers’ exclusive claim to talent; after all, if anyone could communicate with the mysterious machines in English languages like FLOW-MATIC, why should programmers be paid such high salaries or given such special positions within the company hierarchy? Many managers would certainly like to rid themselves of the lofty “computer boys” if given the chance. If the compiler were to successfully open up programming to a wider range of users, programmers would no longer be able to act like “high priests” (as Grace Hopper liked to call them)—even they could become vulnerable to the automation and job loss feared by the papers. 

Of course, not all programmers were resistant to the compiler. The community was “divided into two camps—the ‘space cadets’ and the ‘neanderthals’”—labels coined by the outspoken computer scientist Herb Grosch. While the “neanderthals” were critical of the compiler’s threat to their jobs and clung to the old world of programming, the “space cadets” embraced automatic programming as the future of the software industry, sharing Hopper’s hopes that the compiler would make programming easier and more efficient while also elevating its status. Among those in favor of automatic programming, many believed that the field would elevate programmers to be able to compete with engineers in terms of status and prestige, since compilers added an easier to use “‘second level’ of computers” for which programmers were the new experts. Many female programmers were particularly drawn to this promised elevation of programming; while women were largely excluded from the “masculine” field of engineering, they had a significant foothold in the computing industry (even if they were mainly confined to lower-status roles). These women hoped that the elevation of programming would, to some extent, uplift all programmers—even if complete gender equality in programming wasn’t on the table, female programmers could still stand to gain higher status and professional recognition independent of their status within the computer industry. 

Those in the “space cadet” camp attributed the complaints of the “neanderthals” to selfish motivations. The “space cadets” accused the “neanderthals” of greedily “guarding [their] skills and mysteries” and resisting progress by willfully portraying the development and adoption of the compiler as a fruitless pursuit. According to the “neanderthals,” the compiler had major flaws—especially in terms of run-time costs. Although a mathematician equipped with the A-0 compiler could solve a problem over 17 times faster than a team of three people using traditional programming methods, some critics noted that the final “run-programs” produced by the compiler were slightly less efficient than those produced by skilled programmers. A single extra minute of run-time, they argued, could add up to thousands if the program was run repeatedly; moreover, at the time, computer hours were far more costly than programmer hours, so the effect of the extra run-time would be even more detrimental. Later on, of course, hardware would become much cheaper than labor, but during 1952, these were concerns which Hopper took seriously when developing the next iterations of the compiler (the “A-1” and “A-2”).

Indeed, Hopper was careful to address the hopes and fears of each camp. Catering to the hopes of the “space cadets,” Hopper argued that the compiler would free programmers to work on complex problems that were actually worth their time, rather than wasting effort on “arithmetical chore[s].” Indeed, a brochure for FLOW-MATIC—the language for Hopper’s “B-0” business compiler—claimed that the FLOW-MATIC system would “free[...]” skilled programmers from “clerical drudgery to do more creative work.” Importantly, although it would become easier to train non-programmers to use the “FLOW-MATIC method of programming,” opening up positions for non-technical users, there would still be a need for a select class of “highly skilled career programmers.” Thus, the “neanderthals” had little reason to fear job loss or the devaluation of the profession.

Of course, Hopper was not only addressing the programming community when defending the compiler—she also had to convince business leaders that automatic programming was worth the investment. When speaking to a business audience, Hopper actually played up the fact that the compiler would make companies less reliant on programmers, thus helping to save on labor costs and improve efficiency. In a 1953 report for Remington Rand in which Hopper sought funding for the compiler, Hopper warned that the cost of programming labor was rising exponentially as companies relied more and more on a short supply of skilled programmers, and that this trend would only get worse over time as computers became increasingly important in the business industry. Indeed, reducing programming time was paramount in a business setting, she argued, where processed information was only useful if delivered fast enough to inform an important decision. The compiler would solve all of these problems: replacing costly programmers and making the programming process more efficient and accurate by reducing human error. The 1957 FLOW-MATIC brochure echoed these arguments, claiming that the business compiler would not only achieve “greater coding efficiency” but also remove the need for a “large staff of trained programmers,” since “the method of writing pseudo-code [could] be easily taught to clerical workers.” Business compilers in particular also had the added benefit of making computing more accessible to businesspeople by allowing them to “transmit their ideas from system flow charts directly into the running programs”; moreover, since everything was in an English language rather than symbols, FLOW-MATIC would finally “break the communications barrier between programming and management groups.” While Hopper had previously resisted the idea that the compiler would lead to job loss among the programming community, here, she utilized those same ideas to appeal to the interests of businesses.

When speaking to a business audience, Hopper also had to be careful to evade the skepticism surrounding the compiler. While business leaders had no reason to fear that the compiler would threaten their job security—in fact, there was reason to believe it might actually prove to be a boon to their power struggle with the new class of computer programmers—many people were reluctant to believe that computers could do anything beyond simple arithmetic, much less compiling their own instructions. It wasn’t just an issue of skepticism about efficiency: as with the paper which described computers as “moron[s]” or the cartoon which presented the computer as a fragile jumble of parts, many people simply didn’t believe in the capabilities of the machine. As Hopper put it, “It was totally hopeless [...] it was obvious to everyone that computers could only do arithmetic; they couldn’t write programs.” To combat this skepticism, Hopper tried to frame the compiler’s work in simple, mechanical terms, explaining that the machines “weren’t really writing a program, they were only piecing one together.” This framing also served to reassure programmers that the compiler was only automating the simple, clerical elements of programming. Nevertheless, many people “just didn’t understand” how the compiler could achieve this fantastical feat. Hopper was well aware of the limitations of the computer—despite her humanizing language in the marketing of the compiler, she was well versed with the inner workings of the computer from her time working at the Harvard Computation Lab and claimed to have “always seen computers as a tool, as much as a screwdriver or a lever or anything else”—but some still viewed her compiler as mere wishful thinking. As a result, when marketing the compiler, Grace Hopper had to walk a fine line, delicately balancing her use of accessible, humanizing language with pragmatism to avoid appearing carried away by the optimism surrounding the computer. 

Luckily, Hopper, from her time as a teacher at Vassar, was well equipped to demystify the compiler for her business-oriented colleagues while also maintaining her credibility. Rather than using intimidating technical jargon, Hopper talked about the computer using metaphors which would be accessible to a general audience. In “The Education of a Computer” Hopper started by explaining the computing process through the metaphor of a “production line”—comparing the way the machine processed raw inputs to produce information to the way a factory might “produce[...] an automobile, a rail, or a can of tomatoes.” By using the familiar image of a factory, Hopper made the computer seem more approachable. At the same time, by presenting the operation as a rigid, mechanical process she illustrated that she was not under the misguided impression that the compiler was introducing true intelligence to the computer. Hopper used a similar rhetorical trick at the end of the paper, personifying the computer as a “college sophomore” who she jokingly hoped would soon “be accepted as a candidate for a graduate degree.” Even when humanizing the computer in this way, Hopper still did not elevate it beyond the position of a student—it was certainly not a “wizard,” like the press claimed. There were exciting developments brewing in the field of automatic programming, but ultimately, Grace Hopper was quick to point out that “the computer [was merely] an extremely fast moron” and that it was simply not “possible to walk up to a computer, say “write my payroll checks,” and push a button.” Indeed, Hopper was always careful to remind her audience that “such efficiency [was] still in the science fiction future.”

Hopper’s proficiency with language also helped her when it came to code-switching between a technical and non-technical audience: “I could switch my vocabulary and speak highly technical for the programmers, and then tell the same things to the managers a few hours later but with a totally different vocabulary.” Hopper was highly cognizant of the competing interests of managers and programmers. Indeed, it was an interesting puzzle to market the compiler—for how could one democratize programming without degrading a class of programmers who had built their status upon exclusivity? The marketing campaigns for automatic programming walked a fine line between promising managers lower programming labor costs and inciting the fears of programmers—for example, one 1955 Remington Rand advertisement pictured a manager tearing up code sheets, freed from relying on programmers now that “UNIVAC Tells Itself What to Do!” 

While Hopper was likely not directly involved in these advertisement campaigns, she was clearly aware of the tensions between management and the programming specialists. Even within her own company, relations between technical experts and the sales division had soured as the “EMCC technical people” had become resentful towards the sales team—so much so that Herbert F. Mitchell Jr. decided to send a letter snidely congratulating the Sales Division on their “brilliant approach of preparing a presentation for customers that was so fantastic, bizarre, and childish that the EMCC would be shocked into the position of demanding to undertake [their responsibilities]” or else risk being made “the laughing stock of the entire professional computer field.” Many managers, on the other hand, were frustrated by the arrogant programmers who had displaced them in the corporate hierarchy. Increasingly, it seemed that one side had to give: Hopper’s new invention would inevitably alter the landscape of the computer industry—the question was, who would pay the price?

While it seemed impossible for Grace Hopper to deliver on her contradictory promises to both programmers and managers, ultimately the compiler did, in fact, make programming more accessible to business users while also elevating the programming profession. Generally, the programming community benefitted from the compiler. While managers had previously hoped that automatic programming might allow them to finally rid themselves of programmers, after initial fears of automation had subsided, it became obvious that the compiler was not stealing programmers’ jobs—programmers were in high demand throughout the 1950s and 60s, so much so that the subsequent labor shortage prompted a so-called “Software Crisis” in the 1960s. As one 1967 Fortune magazine article described it, "Competition for programmers has driven salaries up so fast that programming has become probably the country’s highest paid technological occupation [...]. Even so, some companies can’t find experienced programmers at any price.” Clearly, the “experienced programmers” in question had no reason to fear an imminent shortage of job opportunities or the devaluation of the programming profession. Automatic programming did not remove the need for human problem-solvers—as Willis Ware observed, “All the programming language improvement in the world will not shorten the intellectual activity [...] inherent in the programming process.” In fact, as Nathan Ensmenger points out, automatic programming languages also offered the opportunity to elevate programming in academic status by isolating the theoretical aspects of programming from the drudgery of coding. Much like Grace Hopper and the “space cadets” predicted, the compiler did not make programmers obsolete, but instead enhanced their value (at least until 2026). There was still a clear conception that programmers needed to be highly skilled—job advertisements of the time typically required prospective programmers to hold a college degree, with a preference for science and math majors.

At the same time, programming also became more accessible in certain regards. While it was still certainly a high-status profession—with some advertisements lauding electronic data processing as “one of the fastest growing fields in the USA” which provided “jobs with ‘sky’s the limit’ futures”—it was now possible for those without intense programming experience to eventually become programmers, with the right training. For example, in 1957, Sperry Rand’s FLOW-MATIC brochure was already promising that new users could be trained to use the FLOW-MATIC system “in just a few days,” and by 1961, IBM was offering courses in basic programming without any prior education requirements. Because of the perceived “shortage of capable programmers” during the 1960s, companies were eager to hire programmers at a cheaper rate. Though they could not rid themselves of the trained specialists required to operate compilers, they could (and did) attempt to hire and train less-skilled workers—an effort which was largely made possible by innovations in automatic programming. 

To this end, some companies attempted to bring women into the programming workforce, since women could be paid less, although the methods used to make the field seem more welcoming to women often relied on patronizing stereotypes. For example, IBM launched an ad campaign titled, “Meet Susie Meyer,” which claimed that even “a young girl with no previous programming experience” could “find happiness handling both commercial and scientific applications.” Of course, such advertisements did little to fight the stereotypes surrounding women in programming, much less the wage disparity between male and female programmers. Rather, these female-programmer advertisements were often used merely as rhetorical devices to illustrate the newfound ease of programming. As Nathan Ensmenger puts it, “The subtext of appeals like this were non-too-subtle: If pretty little Susie Meyer, with her spunky miniskirt and utter lack of programming experience, could develop software effectively in PL/1, so could just about anyone.” Despite their immense contributions to the field of programming—including Grace Hopper’s own pioneering work on the automatic programming which made such ease of use possible—women were frequently presented as inexperienced and incapable when it came to the high-level elements of programming, and so were often subtly discouraged from entering the field. 

In this way, as with many tales of creative destruction, innovation came with a cost, and in this case, it was primarily female programmers who paid the price, despite Hopper’s personal belief that talent for programming transcended gender. During the 1960s and 70s the field of computing became extremely masculinized. In an ironic reversal of the early feminization of the field, one article in the March 1967 UNIVAC Newsletter described Hopper’s own pioneering role in the field of computing as an “inva[sion]” of “man’s sacrosanct world of technology,” neglecting the fact that, in the early days, computing may very well have been described as a ‘woman’s world.’ Although the article was positive with regards to Grace Hopper’s “insight [and] foresight” and meant to use Hopper as “evidence of the role women [could] play” in the “complex business world,” it is clear that by the 1960s women were uncommon, and perhaps not fully welcome, in the computing community. These sentiments were backed up by statistics: in 1968, women made up only a mere 8.73% of the computer science industry. 

This was, in part, due to the impact of automatic programming—in the process of elevating the programming profession, it became necessary to defeminize the field, as femininity was typically associated with unskilled labor. Indeed, it was generally believed that the presence of women in the field of software brought prejudice against the programming profession as a whole—for example, one female programmer, Judy Clapp, noted that many of her colleagues believed that “‘Engineering was much harder, much harder, than what ‘we women did’—than software.” As Janet Abbate notes, the exclusion of women from software engineering forums like the NATO Software Engineering conference held in Garmische, Germany was telling: “The Garmische organizers and their followers were distancing themselves, consciously or not, from an earlier era when programming had been seen as women’s work.” While the Garmische conference might have been only a subconscious attempt to elevate programming, according to Guilherme Heurich, managers also made explicit efforts to masculinize the field in order to enhance the status of programming: “Computer work had to become more attractive [...], with better prospects and, crucially, a better perception of what the work entailed. It needed to look like a man’s job, not just something that any former ‘factory girl’ could do.” To achieve this, Heurich argues that 1960s advertisements began catering to the male gaze—depicting young, attractive women in mini-skirts operating the massive machines and using feminine names for the computers to make computing seem more appealing to men, while also reducing female programmers to pretty, but ultimately insignificant, accessories of the office.

Hopper actively fought the idea that women were incapable of participating in programming—after all, her compiler, in theory, should have made programming more accessible to women. In one 1967 Cosmopolitan article, Hopper, using her trademark metaphors, compared programming to “planning a dinner,” arguing that by virtue of their “patience” and “ability to handle detail,” women were already “‘naturals’ at computer programming.” Hopper was not alone in trying to make the field seem more welcoming to women—this same article also claimed that although the computing field was admittedly “overrun with males,” there was actually “no sex discrimination in hiring” and “women [were] found at all levels of work involving computers”—not just the clerical labor of “key punching [...] machine cards” but also “highly skilled job[s]” like “systems analysis.” Others who supported female programming used similar metaphors to argue that women were equally—if not better—suited to programming than men. For example, many, including Hopper, connected programming to stereotypically feminine roles, such as that of a mother or a teacher. 

Despite the gendered prejudices Hopper faced throughout her career, Hopper refused to believe that women could not succeed in computing. In her early days at the Harvard Computation Lab, her boss, Howard Aiken “had not wanted a woman officer,” and although he eventually came around to Hopper, he still held on to gendered views: “Grace was a good man!” Later, during her time at Remington Rand, Hopper knew that “there were not the same opportunities for women in larger corporations like Remington Rand” and watched as her colleague, Betty Holberton, made “tremendous contributions” that were never “properly appreciated.” Still, Hopper maintained that women were able to reach equality with men in programming—in one 1980 interview, she claimed that women “ha[d] done quite well” in computing and predicted that “they [would] go further as time [went] on.” Despite Hopper’s stubborn belief that gender equality in programming was within reach if female programmers could simply prove themselves by “doing a better job,” ultimately, the future did not hold equal opportunities for female programmers. The introduction of the compiler automated away many of the low-level clerical jobs which had allowed women to gain a foothold in the field and prompted the masculinization of the remaining high-level elements of programming as programmers and businesses sought to raise the status of the programming profession. These changes affected the next generation of female programmers: in the decade or so following that interview, the percentage of computer science degrees earned by women dropped from 36% to 28%, and in the decade after that, it dropped yet again to 21%. More recently, in 2018-2019, that percent remained stagnant at 21%. 

Ironically, then, despite her well-intentioned goals for the compiler, Grace Hopper left behind a complicated legacy. While much of the programming community was unaffected by the rise of automatic programming and the compiler generally contributed to the growth and democratization of the computing industry, Hopper’s invention also had the unintended effect of pushing women out of the programming space. While Hopper was highly perceptive and skilled when it came to defending the compiler to various audiences, and generally predicted the positive effects her invention would have on the programming profession, it is often difficult to predict all of the potential consequences a new innovation might have on society. Frequently, the effects themselves are not clear cut: in this case, the exclusion of female programmers was quite subtle—most women were not explicitly discouraged from pursuing programming, and yet programming had still come to be considered as a ‘man’s world.’ This transformation of the programming community mirrors much of the same “creative destruction” triggered by Artificial Intelligence today. Just as the Anthropic study warns that women are already disproportionately affected by this development, the story of the compiler demonstrates that shifts in professional identity are often tied up with gender roles and reminds us that technological innovations often have inequitable social impacts. Even tools designed to increase accessibility and efficiency can have unintended, yet drastic consequences. In this light, the “creative destruction” of Hopper’s compiler serves as a warning for the impending age of AI automation: new technologies aren’t “wizard[s]” or “moron[s],” but agents of double-edged change—sparking progress and growth at the expense of those who are shut out of new professional identities.

Bibliography

• Abbate, Janet. Recoding Gender: Women's Changing Participation in Computing. Cambridge, MA: MIT Press, 2012. 
• Beyer, Kurt W. Grace Hopper and the Invention of the Information Age. Cambridge, MA: MIT Press, 2009. Accessed January 31, 2026.
• "Changing the Curve: Women in Computing." UC Berkeley School of Information. Last modified July 14, 2021. Accessed April 12, 2026. 
• Ensmenger, Nathan L. "Letting the Computer Boys Take Over: Technology and the Politics of Organizational Transformation." International Review of Social History 48, no. S11 (2003): 153-80.
• Heurich, Guilherme Orlandini. "Computing Gender." In Coderspeak: The Language of Computer Programmers. UCL Press, 2024. 
• Hopper, Grace. Interview by Angeline Pantages. In Oral History of Captain Grace Hopper. Computer History Museum, 1980.
• Hopper, Grace. "Automatic Coding for Digital Computers." In The High Speed Computer Conference. Baton Rouge, LA, 1955. 
• Hopper, Grace. "Automatic Programming Definitions." In Symposium on Automatic Programming for Digital Computers, edited by Navy Mathematical Computing Advisory Panel. Washington, D.C.: Office of Naval Research, Department of the Navy, 1954. Accessed March 26, 2026.
• Hopper, Grace. "The Education of a Computer." Proceedings of the 1952 ACM National Meeting (Pittsburgh) (ACM '52), May 2, 1952, 243-49.
• Introducing a New Language for Automatic Programming : UNIVAC FLOW-MATIC. Sperry Rand, 1957. Computer History Museum, ​​
• Light, Jennifer S. "When Computers Were Women." Technology and Culture 40, no. 3 (1999): 455-83.
• Mandel, Lois, and Henry Grossman. "The Computer Girls." Cosmopolitan Magazine, April 1, 1967, 52-56.
• Massenkoff, Maxim, and Peter McCrory. "Labor Market Impacts of AI: A New Measure and Early Evidence." Anthropic. Last modified March 5, 2026. Accessed April 12, 2026.
• Uctu, Ramazan, Nadide Sevil Halici Tuluce, and Mustafa Aykac. "Creative Destruction and Artificial Intelligence: The Transformation of Industries during the Sixth Wave." Journal of Economy and Technology 2 (November 2024): 296-309. 
• Williams, Kathleen Broome. Grace Hopper: Admiral of the Cyber Sea. Annapolis, MD: Naval Institute Press, 2004.