Midlife MavericksI carried thirty years of remorse over choosing economics instead of engineering. Then Reed Hastings opened his mouth.
At fourteen I spent days clutching the course-options form, unable to decide between law and economics. I remember the pen stopping mid-word, and the spot where I was supposed to put the cross.
Law pulled me by the human side. Economics by the numbers. And I, who always loved philosophy, literature, and that stubborn feeling that understanding people was the most useful thing you could learn, ended up choosing economics.
I’ve always thought engineering and physics would have been better choices. Every time I crossed paths with a colleague who’d done MIT in data science or a doctorate in applied statistics, I felt that small thing in my chest telling me I should have shut up the humanist side and learned to code.
This week I read something that blew that whole story up.
Reed Hastings, founder of Netflix and former CEO of the company that rewrote how two billion people spend their evenings, was interviewed by Reid Hoffman. Hastings has a master’s in artificial intelligence from Stanford. From 1988. Three decades before most of us even knew what it was. Today he sits on the board of Anthropic.
If anyone knows this AI cycle from the inside, it’s him.
And what he said, into the microphone, without softening anything, was: “If I had a three-year-old today, I’d bet everything on emotional skills.” He also said we’d spent twenty-five years telling kids to learn to code and that we’d been wrong. That maybe we’re now entering a rotation back to the humanities, to the combination of history and literature. And also that: “Salary doesn’t follow value. It follows scarcity.”
The man with a Stanford AI master’s, sitting on the Anthropic board, saying publicly that if he had a small daughter today he wouldn’t put her in front of a coding course. He’d put her in front of the humanities.
I spent three decades carrying remorse and even embarrassment about having chosen social sciences and not something more “serious,” with proper stochastic calculus and theoretical physics, which I always found unbearable. And now Reed Hastings says it was the other way round.
How did we get to this point in 2026, with language models shipping code better than any junior engineer and entire audits being done in minutes by software that costs less than a hundred dollars a month? The reality is that in 2026, what’s left for humans is precisely what the technology still does badly. Picking up when a client says they’re aligned and actually isn’t. Reading what someone can’t put into words. Knowing when to push a meeting forward and when to let the silence do the work. The ability to read between the lines.
At the root of it, what the AI sector knows and generally doesn’t say out loud. What’s becoming rare is what many of us women bring from the factory and were trained to hide. The eye. The ear. The humanistic training underneath the technical training, the thing that for decades looked like a weekend luxury and actually turns out to be the scarce asset in the market. And salary follows scarcity.
At fourteen I didn’t choose wrong. I was thirty years early.
Comment below. What “soft skill” did someone tell you was a weekend luxury that you now charge people for? I read every response.
This article was first published in Portuguese in my weekly column Oh pá, não me lixem! for Executiva.
