In its mission statement, Harvard College aims to “educate the […] citizen-leaders of our society” and create “conditions for social transformation” by imbuing its students with the skills and experiences to do just that: lead and transform. For centuries, Harvard has successfully fulfilled this promise, educating everyone from the second President of the United States John Adams to the real-life characters of The Social Network. More recently, though, prestigious undergraduate institutions like Harvard have had their value propositions questioned, with notable critics like Peter Thiel even offering prospective students money to forgo their degrees in favor of pursuing ideas that will create value. One of the most stinging criticisms came from a recent Harvard graduate whose article, “Harvard Creates Managers Instead of Elites,” argued that Harvard students are trapped in a hamster-wheel of “optionality” in careers such as management consulting because “Harvard has abdicated its directive to guide students to achieve great things.”
Despite its stated mission, Harvard is not teeing its students up to make the impact it promises. Harvard prides itself on exposing its students to a range of new ideas and skills through three curriculum requirements: General Education (Gen Ed), Expository Writing (Expos) and Language Requirement. Logic would suggest that the solution—or at least part of it—can be found by examining these requirements. Of the three, there is one in particular that should not be a requirement: the Language Requirement. Conversely, Harvard doesn’t require it’s students to know a skill whose importance increases every day: coding.. If Harvard wishes to fulfill its mission, the college needs to update its requirements to reflect the needs of a digital world. For starters, Harvard should replace the language requirement with a programming requirement.
To understand why Harvard should require it’s students to learn how to code, it’s helpful to break down why expository writing is important enough to be considered a requirement. The ability to communicate ideas well and clearly through writing is not only fundamental to being a leader or enactor of change , it’s also a basic skill that serves one well in life—be it transcribing the conversations of your iconoclastic teacher, complaining your priest is abusing his power, or even just shouting into the void that your university’s curriculum needs to change. The ability to write articulately and convincingly gives the writer leverage. What is leverage? In engineering, a lever is something that is designed to give its operator a mechanical advantage, like a pulley. Ancient Egyptians certainly couldn’t lift the enormous blocks that made the pyramid by hand; they used pulleys and other levers to gain a disproportionate advantage with respect to their strength. Similarly, knowing how to write articulately and convincingly gives you a disproportionate advantage because by doing so you are able to spread ideas far more easily and efficiently than somebody that just gets in unstructured dialogues. Leverage is anything that gives you a disproportionate advantage—capital, status, knowledge. This leverage is then compounded by media—something that has been increasingly accessible since the printing of the Gutenberg Bible, and is almost universally accessible today. Media allows you to self-publish, which means it’s totally permissionless, and it has minimal costs of replication, in that for each person to read what you have written costs you almost nothing.
This formula of writing plus media gives you permissionless leverage with no marginal costs of replication. That means I can write this essay and post it on my blog without anyone else’s permission, and anyone in the world will be able to consume my ideas. On top of that, I could even monetize my blog in a few minutes using tools like Substack or Ghost. Today, anyone who knows how to write well enough can use this method to drive change (and earn money) while they sleep, without answering to anybody. A fantastic example of this (and probably the poster boy for paid blogs) is Ben Thompson—who writes Stratechery, a blog on technology and media. Thompson has been writing the blog—which costs $120 a year—for almost eight years and has approximately 25,000 paying subscribers. For almost all of recorded history, writing plus media existed in a category of its own in the hierarchy of leverage. This is likely why Harvard introduced Expos back in 1872, in order to ensure each of the leaders it was shaping had the tools to take advantage of permissionless leverage with no marginal costs to efficiently affect impact. But for a few decades now, writing has not been alone. It has a fierce and often more potent competitor that Harvard must acknowledge: code.
Code is the other side of the “permissionless leverage with no marginal costs of replication” coin. Unlike writing, code actualizes the cliche “actions speak louder than words.” It allows the author to communicate their ideas and vision through instantiation. For the first time in human history, thinkers can not only tell people their hypothesis for how the world should work—no matter how crazy—but actually test that hypothesis at low cost by building something that reflects the hypothesis and letting the world use it. The perfect example of this is Bitcoin. Satoshi Nakamoto, Bitcoin’s unknown creator(s), had a radical hypothesis that they communicated in writing, and then released in code to test it. Vitalik Buterin and Gavin Wood realized that decentralized blockchains had even more potential to create a world of better governance and finance and demonstrated this by authoring Ethereum. Today, Bitcoin and Ethereum combined are valued at well over a trillion dollars and counting.
Had Satoshi, Vitalik and Wood only been able to pull the lever of the written word, these ideas would probably have been published in some sparingly-read journal and receded into obscurity because of their complexity and divergence from legacy systems. However, as that trillion dollar market capitalization shows, it is more powerful to see ideas instantiated—no matter how experimental—than it is to read a manifesto. For that reason, it seems obvious that Harvard, which is educating its students to be agents of change for those parts of society that may need re-invention, not just reform, should ensure they know how to pull both of the most fundamental levers for change, so they can go out into the world and decide which one to use and when.
On the other hand, the same certainly does not hold for languages. Proficiency in a language other than English is a valuable skill, not only in that it opens the door to opportunities abroad and different cultures, but also because learning and maintaining another language is good for your brain. I do not mean to disparage foreign language study: German language and literature is part of my joint concentration, and anybody who knows me will be sure to tell you how much I enjoy studying German and the high regard in which I hold the German department at Harvard. Languages are just not required for Harvard to fulfill its mission, and so they needn’t be a requirement for its students. It isn’t necessary to speak a foreign language in order to be a citizen-leader. Information, by way of technology, is more accessible than it ever has been, bridging the gap between languages and cultures. One need only look at the burgeoning start-up sectors in Africa or countries like India, with its extremely varied languages and cultures, to demonstrate mastery of a language is no longer necessary to incite change or create an impact.
“The language requirement is outdated,” says a member of the Harvard class of 2020, who took Spanish to fulfill his language requirement. “Having to take a course I knew I wouldn’t continue meant I truly got nothing out of it.” This low regard for the language requirement leads to its perversion amongst a large number of the cohort who do not place out and find themselves having to take a year-long beginner course. Each undergraduate has a friend, or even a group of friends, in a similar position. In order to expend the least effort in fulfilling their language requirement, they choose a language they might have taken in high school or one which they have absolutely no intention of continuing to study.
Of course, there are a number of students who have always wanted to learn a language, and who undertake the study of French, Arabic, and perhaps even Zulu or Pidgin. They might even choose to study it further through their concentration or as a language citation. Perhaps the skill will be of use to them in their future careers, for research, or just to read their favorite book in the original language. But most of this subset will always have wanted to study a language, and so would be likely to take one up at college even if there were no language requirement. In fact, many students choose to continue to study a language, or even pick up a new one, despite having placed out of the language requirement.
Out of my blocking group of seven, all of whom placed out of the language requirement, four of us have continued to take courses in Harvard’s language departments, of which two are learning new languages. One of us is concentrating in a language, while each of the other three are considering earning language citations or secondaries. After all, we are the students that are relevant to the language department: the ones who have a genuine interest in pursuing a language. The language departments’ existence is certainly not predicated on the four bored students in the back of a ten-person introductory class who are only there because they have to be. In fact, as any professor especially in the era of Zoom will attest, disengagement is contagious, and so it follows that having the disengaged students who are merely in introductory language classes because they have to be is counter-productive and actually hampers the learning of those who are genuinely interested. How on earth Harvard presumes these poor souls learning how to say “ball” and “mother” in complex languages will translate to a “transformative” experience for these students and their communities is beyond me. This issue is exacerbated by the opportunity cost of so many students having to take 2 semesters of a course which will be useless to them—unless they’re at a restaurant in Berlin and order a coffee in German to impress their attractive server. The student now has only 29 of 32 course slots (in the average case) free, 28 if they are placed into the year-long Expos program. Wouldn’t it be better to replace these two courses for everyone with the study of something that gives you a universal, invaluable skill?
This is where a coding requirement comes in. Unfortunately, it is not as simple as making Introduction to Computer Science (CS50) mandatory. For starters, a serious problem with CS50 is that around 40% of its students take it SAT/UNSAT. As any Harvard student knows, taking a course SAT/UNSAT, although intended to allow you to explore a new field with lower stakes, ultimately just results in under-prioritizing the SAT/UNSAT course in order to devote more time to letter-graded courses, which—to put it bluntly—actually matter. I took CS50 and the main thing I got out of it is the ability to not entirely fraudulently put “Familiar with Python, C++ and JavaScript” on my resumé. Even if Harvard were to make it mandatory that people take CS50 letter graded—although this would be a start—it wouldn’t be enough. “I actually don’t think I gained any practical skills from CS50,” remarks an Economics concentrator from the Class of 2022, “because I felt there was a large discrepancy between the lectures and the problem sets […] Also, they tried to stuff too many languages into one semester, which is not nearly enough time to develop a solid enough grasp of that language.”
Taking it a step further, CS50’s way of introducing coding, with its merch, club fair, and other bells and whistles is outdated. That approach might have worked ten or fifteen years ago when the CS department was trying to attract students to programming, a quirky skill that had made people who sat in the same seats as you billionaires and might do the same for you too! But CS50 is now mainstream—it is even more popular than Economics 10a, the introductory course that was until recently the most popular course in the catalog. On top of that, programming has quickly become an indispensable skill—one that forward-thinking people recognize will be essential for problem-solving not just to add more value in future jobs, whatever they might be, but also in life.
Perhaps the best solution is a more clinical course, which, like Expos, is high intensity, harshly graded, and offers sections that target different concepts so undergraduates can select those that interest them and are relevant to their fields. An engineering student from the class of 2022 points to more focused and in-depth versions of boot camps, such as those offered in Harvard’s School of Engineering and Applied Sciences for the programming language MATLAB. Although these boot camps are short, he emphasized how their focus and subject orientation could be invaluable for students, especially in STEM or the Social Sciences. “Essentially I think that an improvement could be the provision of more of these sorts of bootcamps with narrow scope for different disciplines and different languages,” the student says.
These sorts of courses would introduce students to programming languages like Python, R, or Stata, and do far more than encourage enrollment in other Computer Science courses. “For higher-level courses coding experience is taken as a given. It’s generally not even listed as pre-requisite and does not have much support,” he adds. Thus, these courses would also enhance the learning experience (its “transformative”-ness) by eliminating significant barriers to entry to courses in other departments, such as Statistics, Biology, or Engineering, that are coding-based. Another avenue could be a more general introductory course that lands somewhere in the middle of CS50 and CS1, which offers a slower-paced introduction to computer science, suggests a recently graduated Computer Science concentrator.
This is not an accusation that Harvard has virtually no Computer Science department or that it is abysmal. This argument is simply criticizing a failure on Harvard’s part. Of course, two semesters of programming won’t turn a Harvard first-year into Mark Zuckerberg, but two semesters of Expos hardly yields rising Simone de Beauvoirs either. But just as it is impossible to take a humanities course without having to write essays, it will soon be impossible to take STEM or even social science courses without having to use at least basic R—if this isn’t the case within the next five years, then Harvard will be outdated. Any humanities person getting halfway through college and realizing they don’t understand basic programming is analogous to a STEM major getting halfway through college and realizing “I’m halfway through college and I don’t even know how to write a damn essay.” It is a skill so essential that Harvard should provide a failsafe, just as it does with Expos.
The most common response to this argument from other Harvard students is: “Surely I can just hire someone who knows how to code.” For those who say this and also defend the language requirement, get this: you can also hire a translator, and guess what? They cost less. For those who understand why the language requirement is outdated but still pose this question when faced with a coding requirement: this might be true for businesses and entrepreneurs who have been established for a long time. These are (in most cases) people who graduated college before the widespread adoption of mobile phones; they generally have the lever of capital, reputation, experience or some combination of the three to back them up. A hedge fund manager like Ray Dalio can hire quants that code instead of doing it himself because he has the bandwidth to pay them mouth-watering salaries so that he may use their leverage for his capital gain. However, for our digitally-native generation, setting up any business, especially one that is looking to achieve scale and have a great impact, will need technical know-how. It need not be exhaustive: knowing the fundamentals of any system, be it programming or even rocket science, helps form a framework to conjure ideas, understand what works and why, and make key decisions regarding scaling and functionality in an informed way—Elon Musk isn’t building the rockets himself, but listen to him speak and you know he has a good enough idea about what’s going on. In fact, it’s more than reasonable to assume that even Ray Dalio has a general understanding of the code his programmers are writing. If the aspiring founder of a high-potential business is hiring someone to build the product for them because they can’t code the fundamentals, the person they are hiring will flip them the bird, steal their idea, and just do it themselves (for corroboration, watch The Social Network). Just as writing a really good essay requires a grasp of constructing sentences and then targeted learning can teach the rest, so building software requires a fundamental understanding of code.
“Another argument for this point is that all Harvard students know how to speak at least one language, whereas not all Harvard students know how to code,” says Oskar Schulz ’22, an Economics concentrator in the class of 2023. “As long as you can speak one language, whether it’s English or American Sign Language, you can communicate: after that, it’s easy to go from one to N easily, whereas with coding, if you don’t know how to code, you have to go from zero to one. That’s a far greater jump.”
Schulz is referring here to the model most easily found in Peter Thiel and Blake Masters’ book Zero to One. The basic argument is that replication is easy, whereas the initial jump from no knowledge to enough knowledge to replicate to more knowledge is the real barrier to entry in any system of understanding. Once somebody understands what grammar is, they can learn (with varying difficulty) the grammar of other languages. However, if somebody understands neither basic grammar nor its function, learning a language is much more difficult, as they will not be able to replicate with reference to a “1”. In fact, it is only because one spends the first eight to ten years of their life learning and studying their native language that they can use it as a base layer for a foreign language. In our lifetimes technological progress is going to continue to accelerate such that code will be as much of a base layer for life as language has been for the last two millennia—in much of the world, it probably already is. Societies will adopt blockchain technologies, Elon Musk will put chips in our brains and Google will drive us all around . Code will, of course, be at the foundation of this. If each of us at leading institutions aren’t taught this base layer now, we will not even be able to keep abreast with these developments such that we have enough knowledge to have an educated opinion about these technologies. This is a problem because, if history is anything to go on, a few of us might hold public office one day. As any antitrust hearing with a big-tech company CEO reveals, it’s imperative for public officials to have some understanding of technological developments in order to make good policy decisions about them.
Ultimately, no coding requirement will be perfect. To say so would be part idealistic and part stupid. Like with the language requirement, there will be people who slack here as well. Indeed, every couple of years a few English or Philosophy concentrators will come along who will never use the skills learned again, but at the absolute worst it will be a net-neutral outcome to the current state of the language requirement—for this to come to pass, the value-proposition of understanding basic code will have to diminish over the coming years, while the world’s average English proficiency, which is currently rising, will have to decrease. I doff my hat to anybody who genuinely believes in this scenario because they are living in a past century.
Virat (virattalwar@college.harvard.edu) is surfing poorly, failing at keto, and pretending to read Proust while taking time off from Harvard.