Build to Think: How Machines Shape the Way We Know
By Nathan Waterhouse ยท 2025-07-30
You learned to ride a bike as a child. You can probably still do it. But here's the thing: physicists still argue about exactly how bicycles work. The gyroscopic effect? Not enough to explain stability. Trail and rake geometry? Part of it, maybe. We've been cycling for over a century, yet the full physics remains surprisingly contentious.
This should be unsettling for anyone who thinks innovation flows from understanding to application. We like to imagine it works like this: first we discover principles, then we build things. But what if we've got it backwards?
The Cart Before the Horse (And Why That's OK)
At 21, studying Product Design, I watched people colonise early virtual worlds and chat rooms. They were doing something profound, building new kinds of relationships, new ways of being present. But ask them to explain what they were doing, and they'd shrug. The doing came first. Understanding lagged behind.
This pattern shows up everywhere once you start looking:
- We fermented beer for millennia before discovering yeast
- Sailors navigated by stars before understanding celestial mechanics
- Doctors prescribed aspirin for decades before anyone knew it inhibited COX enzymes
- The Wright brothers achieved flight while physicists "proved" it was impossible
There's even a name for this in the philosophy of science: the machine theory of science. It suggests that technology doesn't follow from scientific understanding. Often, it's the reverse. We build things that work, then scramble to explain why.
This isn't new thinking. Sociologist Edgar Zilsel traced how Renaissance craftsmen's practical knowledge eventually forced theoretical science to catch up. Historian Thomas Hughes showed how technologies develop their own momentum, dragging society and understanding in their wake.
The steam engine is the classic case. For decades, these machines pumped water from mines while scientists scratched their heads. The engines worked, but nobody understood the principles. It was only through tinkering with these mysterious devices that the laws of thermodynamics emerged. The machine taught us the science, not the other way around.
When Building Is Thinking
Designers have a phrase for this: "build to think." You've felt this if you've ever:
- Sketched an idea and suddenly seen its flaws
- Written a draft that revealed what you actually meant to say
- Prototyped something and discovered the real problem
Your hands know things your head doesn't. The act of making isn't just executing thought; it's a form of thinking itself.
This is oddly comforting in our current moment. We built GPT before we had a theory of intelligence. Now we're poking at these language models like those 18th-century engineers poked at steam engines, trying to derive principles from a working mystery.
But Here's Where It Gets Strange
So far, so good. We can do things before we understand them. But there's a second, weirder idea lurking here:
Tools don't just help us do things. They change what we see.
This isn't about capability. It's about perception itself.
The World Through Instagram Eyes
Think about the last sunset you watched. Did you see it, or did you see potential content? That mental shift ("this would make a good post") isn't you using Instagram poorly. It's Instagram reshaping what a sunset is.
This happens with every powerful tool:
- Spreadsheets didn't just calculate; they made us see relationships as rows and columns
- GPS didn't just prevent getting lost; it atrophied our mental maps, changed how we experience cities
- Email didn't just speed up mail; it made all communication feel urgent
- Dating apps didn't just expand options; they turned romance into a marketplace
- Clocks didn't just track time; they created "lateness" and made us think of hours as currency
Before the microscope, there was no microbiology. Not because we lacked the word, but because the microbial world literally didn't exist in human experience. The tool created a new category of reality.
Philosopher Martin Heidegger (see enframing) saw this coming. For him, technology wasn't just instrumental; it was a way of knowing, a mode of revealing. Each tool doesn't just help us see the world. It reveals which world we're living in.
The Double Whammy of AI
Which brings us to this moment, where AI is delivering both revelations at once.
First, the familiar pattern: We built it before we understood it. GPT works, but we're still arguing about whether it "understands" language or "thinks" or just does very clever pattern matching. We're back to riding bicycles we can't fully explain.
But the second shift is already happening: AI is changing what we see.
- Writing is becoming collaboration rather than solo creation
- Coding is shifting from syntax to intent
- Expertise feels less like possession and more like curation
- Creativity itself is being unbundled into components we didn't know existed
A colleague recently said, "I can't write a first draft anymore without feeling like I'm doing it wrong." That's not tool misuse. That's a tool reshaping what writing is.
From Product Design to Epistemology
Looking back at my 21-year-old self watching those early VR pioneers, I realise I was glimpsing something fundamental. Those pixelated avatars weren't just representing people. They were creating new categories of presence, new ways of being together.
McLuhan saw it coming: "We shape our tools and thereafter they shape us." But it's deeper than shaping. Our tools don't just change what we can do. They change what's there to be done. They don't just solve problems; they reveal what counts as a problem.
The Question for Leaders
If you accept this double insight, the strategic implications are profound:
Innovation isn't always about knowing more. Sometimes it's about building something that works and figuring out why later. Your next breakthrough might come from tinkering, not theory.
When you adopt a new tool, you're not just gaining capability. You're changing your organisation's perceptual field. What new realities will become visible? What old ones will fade?
This is particularly urgent with AI. Every prompt you write, every workflow you automate, every decision you delegate is quietly rewiring not just what your organisation can do, but what it understands itself to be.
Build to Think, Think to See
At Adaptive Edge, we often talk about navigating change. But perhaps the deeper challenge is recognising that our tools aren't how we respond to change. They are the change.
That prototype your team is building, that AI system you're training, that new platform you're implementing: these aren't serving your vision. They're creating the conditions for new vision.
Sometimes the next insight won't come from thinking harder. It will come from building something that lets you think differently.
And sometimes, you won't even recognise what you've built until it's already changed you.
The bicycle still works, even if we can't quite explain why. The question is: what are you pedalling towards, and what kind of balance are you learning along the way?