DEV Community

NJEI
NJEI

Posted on Edited on Originally published at njei-blog.vercel.app

Software is the brain, not the body

Originally published on my blog. I'm writing from Cameroon, so the examples are local, and I won't speak for other countries.

One of the most useful pieces of technology in Cameroon's recent history is a plastic pipe.

It runs from near the bottom of Lake Nyos up to a raft on the surface. In August 1986 the lake released a cloud of carbon dioxide that killed 1,746 people in a single night. The pipe removes the cause. Once it is primed, gas dissolved in the deep water bubbles up inside it, makes the column lighter, and keeps the flow going by itself. By 2019 scientists had concluded that one pipe could keep the lake safe on its own, with no outside power.

No screen. No code. No app. It is also the clearest example I know of what technology actually is.

What I mean by technology

Technology is any new way of solving a problem so that life gets better — for people, and for what people love: their animals, their land, their families.

Software is one branch of that. I understand why so many young people in tech use the two words as if they were the same. Code is the part you can start tonight, on a laptop, without asking anyone's permission. But it hides how much of the field lives somewhere else:

  • Agriculture. How a harvest is dried, stored and moved before it rots.
  • Health. How a heartbeat is measured in a village with no cardiologist.
  • Transport. How a road is built, marked and maintained so a trip takes the same time tomorrow as today.
  • Animals. How a vaccine stays cold on its way to a farm.
  • Water and power. How either one arrives when you turn the tap or the switch.

Every item on that list can be improved with software. None of them is software.

The word I'm borrowing

Engineers have a name for building physical and digital parts as one system: mechatronics. The word sounds heavy. The idea isn't.

A mechatronic system does the three things your own body does. It senses, it decides, it acts. A sensor reads the world — temperature, weight, moisture. A controller decides what to do about it. Something physical then moves: a motor, a valve, a pump. A thermostat works this way. So does the anti-lock braking in a car, and so does an irrigation system that opens a valve when the soil goes dry.

Mechanics is the body. Electronics is the nervous system. Software is the brain.

Software is the brain. A brain without a body is only a thought.

I should be honest about where I stand. I have spent my time in software and have very little background in mechanical or electronic systems, so I'm writing about the part I'm missing, and I may get details wrong. The pipe at Nyos has no electronics in it at all, so calling everything mechatronics would be a stretch. I'm using the word as a way of seeing: a working system has a body, and the body comes first.

Software automates what already exists

Software is automation. It takes a process that repeats and makes the repeating cheaper and faster. If the process doesn't exist, or exists differently every day, there is nothing to automate.

I made the same argument from another direction in A system must be predictable before it can be automated. A delivery app needs addresses that resolve and trips that take the same time twice. A farm app needs a harvest that is measured, stored and moved the same way each season. A clinic booking system needs clinics that keep time. You can write the code perfectly and the screen will work. The thing it was meant to control won't.

So when someone asks what to build, my first question has become: what will it control?

A test of an idea I like

There is a story I find tempting. Scarcity forces invention, and we haven't faced the kind of life-threatening pressure that makes a people build, so the foundations of technology are thin here. I believed a version of that. It does not survive the first fact I checked.

In 1986 Lake Nyos killed 1,746 people. Two years earlier Lake Monoun had killed 37. Both are in Cameroon. If danger alone produced technology, the pipe would have gone in the following year.

It went in fifteen years later. Feasibility studies began in 1995, the first permanent pipe was installed in 2001, and two more followed in 2011, with scientists from several countries studying the problem.

So the spark was never what we lacked. People here invent constantly.

What is slow is the carrying: the path from one person's idea to something funded, built, tested and trusted.

Take Arthur Zang. Cameroon has a few dozen cardiologists for more than 20 million people, nearly all of them in two cities. He built a touchscreen tablet that takes an ECG in a village clinic and sends it over the mobile network to a specialist. In tests by Cameroonian scientists it proved 97.7 percent reliable. He needed about $45,000. His family didn't have it, banks wouldn't lend it, and he learned electronics from a free online course.

The software was written here. The first thirty units were built by a manufacturer in China, and by late 2016 about seventy had been sold. He has said he hopes ultimately to manufacture in Cameroon.

That is scarcity forcing invention, and it is also the whole problem in one device. The idea was ready. The body behind it wasn't.

Why the money is slower than the idea

"Hindered" is the right word, not "limited." The money exists. It just prices us at a premium.

Africa gets less than one percent of global venture capital while holding about 18 percent of the world's people and roughly 5 percent of its output. That is moving — African venture funding went from about $100 million in 2015 to more than $3 billion in 2025 — but it is still small against the size of the continent. Deloitte, citing a 2023 World Bank report, says African SMEs often pay interest rates up to 22 percent higher than global peers because of perceived risk rather than actual default rates.

Perception cuts both ways, and Deloitte is careful to say that underestimating risk is costly too. I'm not arguing the risk is zero. I'm arguing it should be measured before it is priced.

I think this stigma lands harder on hardware than on software. Code can be tested for the price of a laptop and a data bundle. A pump, a sensor or a circuit board needs a prototype, a mould, a batch, a stock of parts. The more a thing costs to try, the more a risk premium hurts.

What I'd ask for

Three things, in order:

  • Fund the foundations first. Reliable power, testing and certification labs, and a steady supply of components. These come before the next app competition, and they decide what any app can do.
  • Put hardware in young hands early. Sensors, motors and microcontrollers in schools and workshops, the way laptops went in.
  • Measure risk before pricing it. Investors who price African risk from data instead of headlines will find the premium is smaller than they assumed.

I don't know how long this takes. I do know the pipe at Nyos took fifteen years to get from disaster to first installation, and it has been working since.

Next time someone tells you technology means learning to code, ask what the code will control. If the answer is something that doesn't exist yet, that's where the work is.


I write about building, systems and technology from Cameroon at njei-blog.vercel.app.

Top comments (0)