The complex future of work means everyone will need a diverse skill set

A colleague once lamented about a controversy in their academic department (computer science) involving a potential new hire. The candidate in question was formally trained in physics but had a strong record studying information theory, which was of interest to the department. I asked what the fuss was about, to which the colleague replied: “Some folks don’t think physicists should be in a computer science department.” The candidate wasn’t offered the job. 



There are many problems with these sorts of dustups. In this specific case, the biggest crime is the overreliance on an umbrella category (“physicist”) to mark someone’s appropriateness for a job. The departmental protest was not about the candidate’s skill set or accomplishments, but simply about the nominal category-fields through which they were trained. 



Unfortunately, the example is hardly out of the ordinary in many industries (with higher education being an especially large offender), a manifestation of an age-old question: How do we slice the universe of knowledge into pieces that organize expertise into meaningful bundles? And what do we mean by “meaningful”? Is it the structure that allows us to be the most productive, to generate the most ideas, or that which works the best administratively?



The differences between these definitions might seem minor at first glance, but they are profound in practice. And I believe that the future of work will include a courageous engagement with these questions—and a radical redrawing of disciplinary boundaries—as the current state stands in the way of progress, broadly defined.



Where do disciplinary boundaries come from? Historians and philosophers of science have long studied how certain fields came to be. The stories behind the birth of mechanical engineering are different from the ones behind classics or American studies. But they share some features in common: Few of these fields were built to consider how the world will look in the future. Alternatively, the reason why we have the labels we do is mostly because of historical path-dependence. We do it because it is what they did (in the past), and someone before them, and so on.



As they exist, disciplinary boundaries are the opposite of future-proof: They are nearly guaranteed to be obsolete, as their inventors didn’t have the foresight to consider how the world would change (understandably); old departments of engineering couldn’t have known about nanotechnology, or gene editing (though science-fiction writers did). Even sociology wrestles with new identities, driven by computational and big data methods used to study social systems.



The examples are numerous: We now know that photosynthesis has properties of quantum mechanics. “Bioengineering” and “biomedical engineering” have become separate fields, the former sometimes referred to as “applied molecular biology.” Computational linguistics leverages large data sets relationships between languages to infer patterns behind how languages evolved. Faculty who study misinformation sit in schools of business , social psychology, applied mathematics, and biology . 



In fact, the main challenge with disciplinary boundaries is that the pace of innovation surpasses our administrative capacity to put it into bins. This problem is a sibling of a ubiquitous tension in many industries: The demands of the professional world exceed the ability of institutions (e.g., higher education and vocational training institutions) to supply candidates with the requisite skill set. This played out visibly during my own career.



During the 2000s and 2010s (as I was leaving graduate school), “ data scientist ” was plastered on every other job ad, and not only in Silicon Valley and on Route 128 but in many industries. The problem then was that “data science” wasn’t a formal field. There were virtually no undergraduate degrees, graduate programs, or training certificates. Candidates who fit the bill came from all sorts of backgrounds, but had cobbled together a mixed quiver of arrows from statistics, computer science, and applied mathematics (like many, I acquired them through mostly informal channels, as my graduate program didn’t teach or require these skills). 



Though the challenges associated with the current canon are hard to deny, must we necessarily remove boundaries? Surely, there are valid defenses of existing borders around disciplines. They organize skills and personnel into bundles that allow us to train people along structured pathways, and allow us to emit legible Bat-Signals for the talent necessary to grow our enterprises.



When a company is searching for people who can expand its quantum computing division, it knows to ask for someone “with an MS or PhD in physics, with a focus on quantum mechanics.” This makes the act of running a business or laboratory simple: If we want to build a quantum computing company, we can eliminate the applicants with mechanical engineering degrees; they can’t quite possibly help us develop quantum computing technology as well as someone with formal training in quantum mechanics. 



Or can they? 



Of course they can. And many lucrative industries have been built by “outsiders”—people with no formal training in areas, who transformed them still. In addition, there are people who thrive specifically in being able to advance across different areas of expertise (several chronicled in David Epstein’s 2019 bestseller, Range ). So frequent are these cases that they are no longer outliers, and can just as readily be described as closer to the standard model of transcendent actors and innovators.



Because of this, it becomes difficult to defend status quo barriers around disciplines, when so many successful people do not fit. One solution resides in a reconciliation between the practicality offered by existing nominal boundaries and a flagrant disregard for their essential nature. We can just admit they might work for our professional purposes, but they ain’t real. 



Returning to the computer science department that we started with: Keep the computer science department, but recognize that the definition of “computer science” is necessarily dynamic and will require a departure from the norms that might have existed even a decade ago. One can acknowledge that and still hire people under an administrative umbrella, teach the necessary courses to undergraduate students, and run a cohesive department. 



The second reason to maintain the current gates around areas of expertise? Discovery seems to be moving along just fine: Artificial intelligence, robotics, medicine, and several other industries continue to invent at record pace, with the potential to improve human life in various ways.



But this argument suffers from no reasonable counterfactual: As far as we know, invention could be faster had we rethought boundaries of the past. There are several exceptions to the standard canon, generally think tanks and invention-driven subdivisions within major corporations (e.g., Bell Labs , the Santa Fe Institute, Microsoft Research). While many have been successful (even spectacularly so), they are too small in size or number to generate a proper comparison to the standard.



The good news is that multidisciplinarity is becoming the flavor of the month in select institutions. Corporations hire consultants to facilitate cross-disciplinary conversations. Those with doctorates in philosophy and rhetoric now drink the Google office espresso, with the same high salaries as their computer science colleagues. We are entering a moment. 



Even higher education seems to be rising to the occasion. Colleges and universities are dedicating entire “schools” to computing , rather than housing as single departments. The idea here is that computing is a way of thinking, learning, and living, rather than a single discipline that lives alone. Colleges/schools of computing are built to avoid the puny imaginations that often dominate department-level decisions.



Other examples are even grander: Under the leadership of its intrepid president, Michael Crow, Arizona State University, the nation’s largest public university, has embarked on a mission to rethink the role of higher education. This ambition has impacted all corners of the university: from unusual department structures (that do away with traditional disciplinary boundaries) to an emphasis on trying to educate as many students as possible from many life and educational backgrounds. The Arizona State example is not without its detractors and controversies, but it is also praised for its innovation . 



Could a larger movement to vaporize formal boundaries between disciples help us with broader challenges, such as climate change, conservation, growing inequality, and its health consequences?  One can argue that these are the areas where hackneyed thinking, protected by silos and traditions, is especially regressive. There is good news here, as new tools are shedding light on some of these problems, like pandemics.



Sophisticated data science methods—often led by quantitatively and computationally trained scientists—are increasingly used to address modern problems like inequities in pandemic susceptibility in native communities , the role of pandemics in shaping social institutions like prisons , and socioeconomic inequality in influenza surveillance patterns .



These are but a handful of examples of how diverse skill sets can be applied, not only to the wealthy but also to the potential benefit of marginalized populations. They are just the beginning, but offer examples of how multiple skills can inform problems long thought to be intractable. 



So, while many institutions already acknowledge the need to cultivate people who have multiple skills, the status quo is still structured around old notions of what domain expertise is. And the solution relies less on any pointed destruction of boundaries and more on a global embrace of the idea that they don’t describe the natural world; they are, at best, heuristics that teachers and students use to structure their teaching and learning.



But the future will require thinkers who not only transcend disciplinary boundaries, but reinvent them, toward a world with solutions we haven’t come up with for problems that our current boundaries have prevented us from identifying.