Why tech needs a new kind of English major

AI can spit out an infinite number of words, but we need humans to use judgment, make meaning, and tell the story—even in a slide deck. I saw Spider-Man: Brand New Day last week. I struggled to enjoy it. I guess I hit CGI overload, although its 89% score on Rotten Tomatoes suggests others have a higher tolerance. Ditto LinkedIn. I visit the site at least once a day, hoping to catch up on smart things friends and colleagues are saying, but I can’t hear them over the whir of their robots typing. I’m not against artificial intelligence or special effects. I’m against the “artificial” and “special” saturating everything to the point that there’s no humanity left. Spider-Man became tedious when everything was so overwhelmingly manufactured that all the wonder was gone. LinkedIn becomes dull when everyone sounds samesy. The comparison isn’t perfect, but the problem is the same: Once production becomes cheap and abundant, attention gets expensive. Why? For one thing, as building a credible first version gets easier, making your product stand out gets harder. Or, as reported years ago in a Wall Street Journal essay, former Google and Twitter executive Santosh Jayaram argued that as more ideas became buildable, competition was shifting from engineering alone toward storytelling. We don’t need more writing. We need more people who know what’s worth saying. They will absolutely use AI to sharpen and extend their voices, but AI won’t be their voice. Tech companies need word people Every few years I realize I need to revisit something I wrote in 2014. Even then, the cost of delivering a new product had plummeted with the rise of open source software and cloud infrastructure. As the industry swam through oceans of products, it became increasingly difficult to help would-be customers understand why they should bother to try out product X versus product Y. In the midst of this sea of choices, I suggested that my next marketing hire might not have a traditional marketing background and, instead, I’d focus on a candidate’s ability to construct an interesting thesis and synthesize it in a few hundred words. We need people with the skill to articulate the why of a product, in other words. If I felt that way 12 years ago, imagine how much greater our need is today, in our age of AI. Every week, another AI company is born, raises a few hundred million at a multi-billion-dollar valuation, and starts competing for your attention. Meanwhile, scads of existing AI companies release new models that are even better than they were (checks watch) five minutes ago. As an industry, we’re spoiled for choice, without having a clue what we should choose. Again, we need more English majors. I realize that some English majors can’t write their way out of a paper bag, and some engineers write beautifully. The degree isn’t the point. I’m talking about people who know how to read closely, ask uncomfortable questions, find the argument buried in a pile of facts, and explain it so another human can act on it. The writing is simply where all that judgment shows up. AI made plausible prose cheap Large language models (LLMs) can write remarkably well, although they can sound tinny to those of us who have made a career out of communication. Still, “tinny” works just fine for many things. Give Claude or ChatGPT a product brief and ask for a first-call deck, launch post, executive email, or 20 social snippets, and it will crank that out within seconds. Nor did AI invent bad corporate writing. Companies have ghostwritten tedious thought leadership pieces or mindlessly dull press releases for eons. AI merely industrialized it. Research suggests the sameness isn’t imagined. In a 2024 Science Advances experiment, access to AI-generated ideas helped people write stories that readers judged more creative, especially when the writers were less creative to begin with. It also made the stories more similar to one another. More recently, a Nature Human Behaviour study analyzing more than 880,000 texts found that writing became less varied as AI use spread. In experiments on a subset of those texts, LLM rewriting generally preserved the substance while reducing variation in writing complexity by 21% to 50%. AI may make writing better overall, but it also makes it a bit beige (to use my son’s term). The scarce skill is judgment LinkedIn knows this. As I wrote earlier this month, the company is trying to suppress generic AI slop. In that article, I focused on whether some outputs, including sales collateral nobody wants, should exist at all. But today I ask a different question: Who can turn all this abundant output into something worth using? Call it “taste,” if you will, but it’s a plea for a person to show up in the writing. This is why “storytelling” can’t simply mean making product claims sound better. In AI, particularly, the stakes are too high. Customers are deciding where their data goes, which work to automate, whom to trust, and whether to bet years of organizational effort on technology that seemingly changes daily. A communicator who merely turns the product sheet into smoother sentences isn’t helping. We need people who can distinguish a benchmark from a customer outcome or a new feature from a reason to switch. They need to understand what the company can honestly promise and what it can’t (or shouldn’t!). A deck isn’t sales enablement Consider the generic first-call deck. Any decent LLM can produce one in seconds: market shift, three customer challenges, product architecture, impressive logos, and next steps. It could be awesome, as first decks go. It still might not be particularly useful. Maybe, for example, a seller doesn’t need that pile of slides. Maybe she needs a way to earn the next 30 minutes with a prospect. To understand how to help a seller do that, the product marketer needs to spend less time asking a machine to manufacture generic decks and more time listening to sales calls, talking with sellers, interviewing customers, watching where the existing pitch loses the room, and then tweaking it. AI can be hugely helpful in this work. An LLM can analyze call transcripts, find recurring objections, turn customer evidence into a talk track, adapt a core narrative for an industry, and help a seller prepare for a specific account. It can complement the human in the loop by giving force to her intuition and questions. That’s the new job for the “English major”: not producing more collateral, but translating between the people who build the product, the people who sell it, and the people who might buy it. In this case, you’d rightly call that product marketing, but it’s a more thoughtful product marketing than we often see. The focus is on the storytelling that stitches together all the other functions, such as competitive intelligence, product launches, sales enablement, and more. This emphatically is not a Luddite role. It’s very much AI-first, early and often. But the AI shouldn’t own the story; it’s a helper to guide and support the story with analysis and data. The goal isn’t artisanal, AI-free prose. We don’t need humans to type every word any more than a good movie needs to ban CGI. The best special effects serve the story and largely disappear. Once the effects become the story, the wonder goes away. We need a new kind of communicator to guide storytelling in tech companies that otherwise think their job is simply to produce 1s and 0s at an increasingly frenetic pace. Someone, in short, to help us appreciate the why in the midst of so much what.