'Mastery Is Not To Be Found In Tools Of Efficiency'

On AI and the usefulness of education

The following is a keynote speech delivered at Villanova College

There’s a phrase that used to get a lot of mileage back when I was in high school, and that phrase was “when am I ever going to use this?”

Calculus? When am I ever going to use that, I would think to myself. Biology, I would not be using that, I was sure of it. I asked the question about Shakespeare and world religions, about geography and physics. Some people would be using that, sure. But me? No.

Today, the unfortunate fact of the matter is that you could ask this question of many more things once considered more straightforwardly useful. How “useful” is it to be able to closely read and understand long or difficult texts when you can probably suck out most of the useful tidbits with a little help of AI? How useful is it to know how to program, design, or write, when billions of dollars are being marshalled by the world’s biggest companies to build tools that code for you, create images, videos, and designs for you, that write your emails, reports, birthday cards, wedding toasts, and eulogies?

AI is probably going to get better at doing most of this and more. We seem to be closing in on the day when this question — “when am I ever going to use this” — could be asked of basically anything we learn.

But in the years after high school it has occurred to me, in my wisdom, that this question is the wrong one. As your teachers have probably explained to you, the value of your education is less in your daily lessons than in learning how to think. So, what should we be thinking about in a world intent on building machines designed to do more and more of our thinking for us?

*

I graduated from this school in 2009. At that point, I wasn’t really sure what I wanted to do. I knew I kinda liked writing, but I did not really see myself as a “writer.” I went to journalism school because I thought maybe I would get into TV or radio. It was a way to give shape to a series of vague interests. I thought it might be useful.

I ended up really liking university, and I still hold onto this romanticized notion of it. Of course, there was all the fun stuff I’m sure most of you are looking forward to: the parties and the freedom and whatnot. But I also enjoyed classes. I got to pick the ones that I actually wanted to go to, listen to professors I found smart and interesting, and got to read books written by people whom I would come to regard as geniuses. There’s a kind of magic in that, in being able to follow them as they laid out how they think about the world.

I fell in love with magazine writing, especially the stuff that was coming out in the early 2000s. There were these writers with unmistakable voices and inventive ways of storytelling. Some of them lived like rock stars, getting paid hundreds of thousands of dollars to write six stories a year about whatever they wanted for Vanity Fair, Harpers, or the New Yorker. They had ridiculously sized expense accounts, travelled to any corner of the world that interested them, and had readers waiting for them to publish basically anything — a deep expose into a slightly aged news story, or some sweeping personal essay of an otherwise ordinary experience. People are still reading them decades later.

I remember when I was studying at Carleton in Ottawa, the faculty used to tell us that 20 years before our time, the most prestigious publications and outlets used to line up on convocation day to scoop up all the top graduates, offering them comfortable lifetime gigs at the Globe and Mail or the CBC. That would not be the case for us. We got a handshake and about 26 months of unpaid internships before some of us landed full-time positions in media, and many of the rest moved on to roles in public relations and marketing.

After the requisite internship circuit, I became a staff writer, then an editor at a well regarded magazine for physicians until — unsettled by seeing colleagues laid off and feeling my optimism about the future of the industry fade — I moved on. I also dearly miss magazine writing and would trade those years for nothing, but I have no regrets about the career that decided to pursue in advertising. I am fortunate to work at a great agency, with great people, and we do work that I’m proud of.

My education was useful. It was where I began to learn what good writing looks like, and it gave me an opportunity to discover the people and ideas that would shape my professional life; it was where I found something that I wanted to master.

*

AI today is most useful in what I call “the proficiency gap.” This is an almost universally held view among professionals that AI can do a bunch of these other jobs, which I don’t know how to do, but of course it can’t do my job, which I do know how to do.

If you don’t know how to code but you want to build a website, you can “vibe code." If you need to put together a report but you aren’t a designer, AI tools can do that too. If you’re not the best with words, an AI-written draft of the email you're struggling with is probably going to look pretty good.

But with a different set of eyes, AI-produced work doesn’t look quite as hot. I know next to nothing about coding, but I know our director of programming is a long way from being able to outsource coding tasks to non-coders using AI. Our designers are not especially happy when they are forced to work with a file a client has designed with AI, or asks them for an absurdly specific image that they can "just make with ChatGPT."

Now, both the coders and designers I work with have no problem using AI to write, but I do not. That is because AI writing is the worst. This is not just because the writing itself tends to be cliched and awful in dozens of other ways, but because it is hollow. It is a pantomime of what you could say so you don’t actually have to say anything. It undermines the fundamental function of any act of communication. I don’t care if you don’t think you can write, or can’t spell, or don’t understand subject/verb agreement, or whatever. I care what it is you’re trying to say. And when you use AI to write, neither of us really know what that is.

Your sincerity is also immediately in question. Am I really sorry if I hand off the task of writing it to AI? Am I really thankful if I cannot be bothered to take the time to thank you myself? How confident can I be in the contents of this email if I needed ChatGPT to co-write it with me?

I know some of you are bored listening to me. I get it. But I promise you it could be worse. If I did not have enough respect for you to write this speech myself — if I had AI crib some well-worn material from other famous alumni day speeches so I wouldn’t have to think quite as hard about what I wanted to say to you — you would be able to tell.

It took me a few years even after AI came around to appreciate this, so I don’t really expect you to at this point. If I had an essay-writing machine in high school, I would have been using the essay-writing machine.

But the thing is, you shouldn’t. I know this is a hard sell, so let me put it this way: I am a professional writer and basically nobody cares what I have to say. I am only slightly more influential than a homeless guy shouting into a paper bag. Getting people’s attention is rare, and holding it for any period of time is rarer still, which is why I cannot explain to you what a privilege it is to have an audience. As a student, you have an audience, by default, in your teachers — people who are thoughtful, who care about you, and who are professionally trained. By the time you reach the end of school, it may be the last time you ever have an audience like that. You might as well tell them what you think.

*

Some people will tell you the most useful thing you can learn today is how to use AI. There is some truth to that. AI literacy is essential if only to help us understand how the ground is shifting beneath our feet.

The other thing these particular people will tell you is that AI is eventually going to free all of us from menial work so we can focus on higher-order tasks. That’s the important stuff. Optimizing your schedule? Done. Research? It’s combed through the journals and periodicals for you so you can get to answers quicker, figure out what questions to ask next. Reporting? It’s synthesized the relevant information for your review before you circulate it. Strategy? Here are a handful of options based on all the available data and your past preferences.

The current frontier in AI technology is what is known as agentic AI, programs that operate independently as they carry out chains of more complex tasks. I do not know if there will ever be a super-intelligent AI that replaces the sum total of human economic output so we can spend our days in a state of perfect leisure, reading poetry, tending our gardens, frolicking, etc. But I know for certain that almost all of us will be giving up more and more of these little decisions to save our precious mental bandwidth for the bigger ones.

This is the promise of AI: that the really meaningful work happens on the other side of the work it is doing for you. But I don’t think that’s true. Ezra Klein, an American journalist, said in an interview recently that when he started his career, his greatest competitive advantage over his older professional peers was his willingness to do things like reading budget reports — these simple 30-page documents — that everybody else found “too boring.”

Being able to do difficult things — including boring or tedious things — is far more valuable than being able to use a tool that is designed to be easy to use. The world is not exactly awash in job applicants who write things like “good at Googling” on their resume. To the extent that using AI will become a skill, it will be in your ability to parse its outputs and to lend it a point of view. But I promise you, you will not have a point of view to lend if you are so often willing to give up your opportunities to develop it. Mastery is not to be found in tools of efficiency.

Of course, sooner or later, you will find yourself in situations where efficiency matters more. When you are a successful scientist who simply does not have the time to apply for 30 research grants in a day, you won’t need to listen to me about how it’s better to be doing that stuff yourself. Just be sure to invite me back when you’re giving your keynote here in a few years.

*
​
There is a long history of moral panic about new technology. The internet was seen as both the great democratizer of knowledge but also as something that was going to make us dumber. We fretted about television; the constant barrage of low-brow stimulus, junk food for our brains. The radio was feared as a tool of propaganda and mass hysteria. Even writing. In Plato’s Phaedrus, Socrates tells the story of the Egyptian god who invented writing and gifted it to humanity. The king responds: “This discovery of yours will create forgetfulness in the learners’ souls…They will appear to be omniscient and they will generally know nothing.”

These examples are typically raised to show that fear of AI is overblown. Years after the introduction of the internet, and TV, and radio, and the written word, we’re all…fine. Right?

All of these critiques are related to the same phenomenon that has been creeping through history, and none of them are wrong. There is not a person in modern society who has the capacity for oral storytelling that our ancestors did. In the 19th century, coal miners would read Thucydides, and Charles Dickens’ readership in America was comparable to the number of people who watch the Super Bowl. Today, even someone like me, the guy standing here asking you to read books and write your own essays, has a difficult time reading one unless my phone is in another room. You watch enough YouTube shorts and eventually movies are going to feel harder to get through.

But each one of these technologies has also brought about new forms of human connection and art, and new means to expand our intellect. We want simple answers. We want clarity around which AI tools are going to enhance our abilities and which will diminish it. AI good, or bad? But the truth is that nobody can give you those answers because it’s not just one question. It is a series of questions that you are going to be asking yourselves every day from here on out.

When you read Shakespeare or study calculus or the human cell, it is not because you are expected to use these things, exactly. You’re probably not going to make a living reciting Shakespeare in the park. Most of us are never going to wrap our heads around the equations that hold up our bridges and buildings, or understand the mechanisms by which medicines save lives. But we learn about the ideas at their foundation because there exists in you the possibility that one day you will decide they matter to you, and this possibility must be nurtured for its own sake. Once you stop asking whether you're ever going to use this, that, in my experience, is when the real learning starts. 

 It has been my pleasure to address you all here today, and I wish you nothing but the best.