what small nations must build

Small nations in the intelligence economy.

Edited with Sofia

Notes on Context, No. V

My grandfather worked in a fish cannery on Maio island, Cabo Verde.

Back then, the island's work held together. Fishing fed the cannery, the cannery fed trade, and trade gave the next boat a reason to go out. People knew how to do things, and there were ways to turn that knowledge into work.

Walking the Canton Fair this month, I kept thinking that intelligence might bring that back. Machines are getting more capable, and the expertise to design, plan and run them is getting easier to reach. For the first time, small nations might build industries at their own scale.

I still believe that. But a cheaper robot doesn't make the boat arrive on time.

A model can help plan production. Someone still has to finance the equipment, maintain it, secure the inputs and find the customers. Intelligence changes some of the costs. It doesn't change the whole business. In Cabo Verde, the World Bank names unreliable, expensive inter-island transport as a brake on growth. Software can help us work around a constraint like that. Removing it is a different kind of work.

That's where the small-nation argument gets interesting.

Small countries have always needed capabilities that are expensive to keep at a small scale. Technical expertise. Good administration. Planning. Maintenance. The ability to understand a problem and keep working on it after the first project ends.

AI makes some of that affordable to small teams. A local business can handle more of its own planning and customer service. A municipal office can spend less time digging through documents and more time solving the case. Someone in the diaspora can help build something at home without moving their whole life back. Each change is small. Together, they change what a place can do.

But access isn't ownership. In The Commanding Heights, Will Manidis and Joe Roberts argue that intelligence will be everywhere, while the wealth it creates settles with whoever controls how it reaches the economy. We'll get the same models as everyone else. That doesn't decide who owns the customer, who improves the service, or who gets paid.

Take tourism. A visitor asks an agent to plan a few days on Maio. It finds a guesthouse, a guide, a restaurant and a morning on the beach. The island gets easier to discover.

Now follow the money and the knowledge. Who keeps the listings accurate? Can the guesthouse change its offer? Can the guide talk to the customer directly? Does the local business learn enough to do better next time? Can it survive if the platform changes its terms?

Those answers decide how much capability stays on the island.

I used to say context can't be downloaded. That's only half true. Context can be recorded, organized and learned, and anything that can be learned can be taken. The people who know a place best need ways to put that knowledge to work before someone else does.

That's what the digital should be for: a tool that helps a place decide, organize its resources and create its own opportunities. At Maioazul, we try to build it so each piece feeds the next. An event brings work to local businesses, better information helps people find them, and the income pays for the next round.

Some of that should stay open. A reliable ferry schedule helps the guesthouse, the resident and the visitor alike. Keeping it accurate is real work, even if nobody can own it. Customer records, land decisions and public money are different. There, we need to be clear about who has access, who can act and who answers when something goes wrong.

The same goes for autonomy. Cabo Verde will keep using technology built elsewhere, as it always has with knowledge, equipment and partners. The real question is whether we understand the systems we depend on, can maintain what matters and can switch suppliers when we need to. Autonomy is having choices and the skill to use them.

Being small can help. On an island, the person with the problem and the person who can fix it are rarely far apart, and a good fix shows up fast. But small places also run short of resources, concentrate power, and watch projects die when their founder or political sponsor leaves.

Continuity doesn't happen by itself. It has to be built.

So I'd measure progress in ordinary things. A business serving more customers. A resident making fewer trips to finish a process. A young person learning to maintain a system. An organization that keeps getting better after its builder has gone.

My grandfather's cannery belongs to an earlier chapter. The lesson still holds. Knowledge becomes productive when people can organize around it, and development lasts when the ability to do the work stays with them.

Intelligence can be imported. Capability has to be built.