I’ve been thinking a lot about Design for a Better World, the latest book by Don Norman — and how quietly uncomfortable it feels when you allow its ideas to sit with you for a while.
Norman doesn’t open the book by talking about screens, interfaces, or usability heuristics. Instead, he asks us to look around. To notice the world we’ve built. And once you do, it becomes hard to ignore how artificial most of it is.
Cities engineered for cars instead of people.
Houses disconnected from climate and geography.
Condominiums with perfectly trimmed green grass — the kind you recognize from TV shows — even when that grass has no relationship whatsoever with the local ecosystem.
Everything optimized for comfort, predictability, and control.
But then comes the question that designers often avoid because it doesn’t fit neatly into a roadmap or KPI:
What happens when nature doesn’t behave the way our artificial systems expect?
We call weeds a “problem.”
We label insects as a “plague.”
We deploy fixes to remove friction from systems that were never designed to be alive in the first place.
Rarely do we stop to ask whether those “problems” are simply signals that we changed the environment first.
This line of thinking brought me back to a story from Brazil that mirrors Norman’s argument almost too well.
If you’ve watched the movie Rio, you probably remember Blu — inspired by the arara-azul-de-lear, a rare and striking macaw native to the Caatinga region in the state of Bahia. Beyond animation and pop culture, this bird plays a crucial ecological role: it helps disperse seeds that sustain the Caatinga’s vegetation.
The Caatinga itself is a unique biome. It covers most of the Northeast of Brazil and exists nowhere else in the world. It’s semi-arid, highly specialized, and deeply interconnected. Yet today, 46% of the Caatinga has already been destroyed, largely due to livestock farming and agricultural expansion. Even more alarming: only about 2% of the Northeast of Brazil is officially protected area.
As natural food sources disappeared, particularly the licuri palm trees that produce the small coconuts the arara-azul-de-lear depends on, the birds turned to what was left. Cornfields.
The result? Farmers began labeling the birds as a “praga” — a pest — not because the species changed, but because the system around it did.
Attempts to “fix” the problem often made things worse. Introducing common bees, for example, seemed like an intuitive solution for pollination. But those bees are not native to the Caatinga. They attack the birds’ nests, disrupting reproduction instead of supporting the ecosystem.
After years of work by biologists and conservationists, small reserves now exist to protect the species. Still, artificial progress continues to collide with natural systems:
- Government programs like Luz para Todos, designed to bring electricity to remote areas, have unintentionally electrocuted hundreds of birds every year. Energy companies are now forced to redesign parts of the electrical network to reduce this damage.
- Livestock farming continues to advance into fragile land, accelerating desertification.
- Monoculture farming replaces biodiversity with efficiency — and then treats the consequences as externalities.
There is even another layer, rarely discussed outside scientific circles. A portion of arara-azul-de-lear pairs are same-sex couples. They build nests, form long-term bonds, and copulate — but do not produce eggs. This is not a flaw. It’s a biological reality. But in a system already under pressure, it adds yet another challenge for species preservation.
None of this is about villains.
Each individual decision made sense in isolation. Electricity brings development. Farming produces food. Infrastructure drives economic growth. But together, these decisions created cascading effects no single actor fully owned.
That is precisely the warning Don Norman is making.
We can no longer afford to design in silos.
Not products.
Not cities.
Not financial systems.
And certainly not artificial agents and AI systems that act autonomously inside complex environments.
Design today is no longer local. Every system we introduce reshapes the context around it. The real risk isn’t bad intent — it’s narrow thinking.
The uncomfortable shift Norman proposes is this:
Designers are no longer responsible only for what they intend to create, but for what their creations enable, disrupt, and displace over time.
So the question I keep coming back to is not “Can we build it?”
It’s “What ecosystem will this change once it exists?”
I’m curious to hear from you:
👉 How are you rethinking product strategy, system design, or AI development to account for second-order effects, not just immediate outcomes?
Let’s talk.