Children ask "why" as one of the first things they learn to say. They don't ask for answers — they ask for the journey. We spend the formative years of our lives translating curiosity into knowledge and understanding, and most of what we call human creativity is downstream of that habit.
Consider three "what if" questions, the kind that seeded some of our species' biggest breakthroughs:
- What if we could trick our immune system into creating specific antibodies?
- What if we could exploit fluid dynamics to take flight?
- What if we could use light to draw transistors directly onto silicon?
The first led to vaccines. The second, to airplanes. The third, to photolithography — the process that made modern computer chips possible, and with them, essentially all of modern electronics. None of these breakthroughs came from a fast answer. They came from humans being unwilling to stop asking questions, and from the long, undirected exploration that follows.
A piece by Anne-Laure Le Cunff in The New York Times last week helped me understand why this matters at the level of biology. Curiosity, she writes, "puts the entire brain into a mode of heightened receptivity — not just for the specific thing you want to know but also for everything around it. Curiosity opens a window, and while the window is open, learning deepens across the board."
The window stays open only as long as the question remains unanswered. The instant you get your answer, the window closes. That's important, because the territory between your question and your answer — the detours, the tangent articles, the unexpected connection between two ideas — is where most of the learning actually happens. When physicists Arno Penzias and Robert Wilson detected a persistent hiss in their radio antenna in 1964, they could have written it off as equipment noise. Instead, they kept asking what it might be, and they ended up discovering the cosmic microwave background — the radiation left over from the Big Bang. That discovery didn't come from efficient retrieval. It came from sustained curiosity in the face of a half-understood signal.
Our tools are increasingly designed to close that window for us. AI assistants in particular are very, very good at giving us answers fast. Three seconds fast. Which is wonderful, when the answer is what you needed. But it's also the kind of speed that closes the window on curiosity before curiosity has had a chance to deepen.
This brings me to a writer named Craig Mod, who recently appeared as a guest on Every's AI & I podcast. Mod is, by his own description, a "mega-dork" about technology. He vibe-coded a replacement for Campaign Monitor, the newsletter platform that was charging him $6,000–7,000 a year for a product that hadn't innovated in a decade. (Nevermind that Anthropic's new Fable model audited his custom app, it told him, "Oh, you have 8 trillion security holes — shall I fix them?")
Mod is also a serious writer. He publishes books. He walks long distances and writes about them. He has a paid membership that funds his work. And he has a hard line: AI does not write a single sentence of his prose. Not a book. Not a newsletter. Not an article. He uses Claude as a research assistant — not asking for answers, but to gather and condense sources he can then evaluate himself. He uses AI to handle the infrastructure around the craft. The craft itself stays human.
Your work, after all, is a product of and a reflection of you.
I wear two hats that nobody pays me to wear. I'm a product-minded software developer. I'm a programmer at heart. My actual job pays me to be a technology leader at a public company, and nowhere in my job description does it say I need to write code or ship products. But these aren't hobbies — they're how I'm wired, and that wiring is what makes me better at the technology leadership work I actually get paid for.
When I put on my Product cap, I love vibe-coding — it gives me the ability to rapidly explore design choices, feature combinations, and approaches, with code quality and architecture taking a back seat to imagination. When I put on my Programmer cap, I want to be close to the architecture, close to the code itself, applying AI in deliberate, targeted ways: to diagnose problems, repeat patterns I've carefully crafted, and automate the tedious work that would otherwise pull me out of the flow. AI still writes code, but it does so at my side.
The pattern looks similar to Mod's. AI handles the infrastructure around the craft. The craft itself stays human. We live in a world far too polarized by AI, with diametrically opposed views. AI is either an infallible technological marvel, or a plague to be avoided at all costs. Neither side sees the full picture.
If you lean into the hype, you make Le Cunff's concern a reality: questions become endpoints, the territory between curiosity and understanding collapses, and the next Penzias-and-Wilson moment never gets the breathing room to emerge. If you dismiss AI entirely, you miss the very real opportunity it offers — to handle the parts of your work that aren't load-bearing for your identity, and to widen the room in which your curiosity can run.
Le Cunff's piece doesn't pick a side. It says the technology is doing real things to the territory between our questions and our answers, and the choice we make about that territory will shape the next generation of human creativity. Mod and my own experience say the same thing from the ground up. Use the tools, but own your craft. Care about what you're making. Care about why you're making it. Curiosity is too important to spend on expediency. The next photolithography breakthrough, the next impossible "what if" — those come from humans who kept asking "why" long after the answer was available. Don't let AI close the window on curiosity. Use it to widen the room.
