Revenge of the English Majors

For decades, if we wanted a student to have a great chance of a successful (and prosperous) career, we generally steered them toward STEM. That was where the jobs and the money were. The humanities were something you studied if you could afford to, or if you were so passionate about it the money simply didn’t matter to you.

Reality supported this advice for a long time. Coming out of college, a computer science major typically earned more than an English major. Many schools, noting this, shifted their money to the technical side.

I felt this dynamic in my early years as well. Initially I wanted to be a philosophy major, believe it or not. What I couldn’t figure out is how that degree would ever earn me a living. So I instead went into electrical engineering. I don’t regret that decision at all. Personal computers hit the scene when I was a teen, and I was interested in them too. But I do vividly remember feeling the tension between the humanities on one side and technical learning on the other, with the latter seeming like the best way to get ahead.

Learning computers and engineering early did help me. I believe that a large part of my business success was being a step ahead of most people on understanding how computers work and how to use them.

In case you hadn’t noticed, tech has been changing rapidly these past few years. While the machines I came up on in the 1980s and 1990s rewarded people who could handle math and logic, that has become less and less true since the day LLMs showed up in our lives. As of right now, the technology that matters most in business today is language based, and you get results out of it by writing and talking to it in plain English. The better you are at that, the more it gives back. I think it’s clear that we’re in the era of the revenge of the English major, where language skills matter more than math.

Working with AI models turns out to be mostly a writing and reading problem, with a little bit of philosophy mixed in. You have to articulate what you want clearly, and give the LLM the proper context in the right order. That is a skill in itself. Then there’s the reading, parsing, finessing of what you get back. LLMs will often hand you an output that sounds sure of itself but it is anywhere from flat-out wrong to ever so slightly, uncannily off. The knowledge worker of today must be able to understand and catch these things, then iterate using more articulation through language. Anybody who studied the humanities has done years of this already. They know how to read closely, and how to rewrite a sentence until it finally says what they mean. It just never occurred to many of them that they’d be doing so in cooperation with an AI model.

Of course the shift can be overstated. Gifted technologists will always be prized. People still need to build and maintain these systems, to understand the technical side of what they output when necessary. But even the gifted engineer who can’t explain what he’s building loses a bit of his edge in the AI world. The most valuable employees are increasingly those who can get technical and write a clean paragraph. What is changing is that the market increasingly pays for being good with words as its own skill set, even in technical arenas. No longer is it just a nice extra.

The common scenario of a young person who loves to read and write but keeps being told there’s no living in it may be on its way out the door. The tools everyone is primarily using now were built to be communicated to in human language. The people who use that language well have a new, and growing, advantage.