As a young man, I never missed Newsline on NTA on Sunday nights. Frank Olize’s voice belonged to the weekend as surely as Sunday rice. If you remember that voice, welcome to the club: our knees now forecast rain better than the weather service. And if you remember what NTA showed before Newsline existed, I greet you with both hands, ma and sir.
Every so often, Newsline featured a young Nigerian genius. One had built a car from scrap metal in his father’s backyard. Another had built a plough or helicopter. A commissioner would promise support, and the audience would applaud from their television sets. I never heard of these inventors again. The genius was real. The path from prototype to product was non- existent.
That gap matters more now, because I believe the next wave of artificial intelligence is physical: machines that sense, decide and act in the real world. The world is currently dazzled by autonomous vehicles and humanoid robots dancing on conference stages and competitions.
Nigeria needs something humbler: functional machines aimed at its everyday problems.
Last week, I stopped at a roadside vulcanizer to pump my tires. His compressor was the usual kind: a small petrol engine driving a piston pump on a steel air tank. New units sell for roughly
N180,000 to N950,000 in Lagos markets, and he buys petrol every day to run his. He works in daylight, in one of the street’s lowest-paid trades. I asked myself why such a basic machine burns petrol when an energy source shines directly above it. I was perturbed about why the machine has remained the same for over twenty years, maybe longer.
A solar compressor is not glamorous, but nearly every bus stop needs one. Add a sensor that reads the tire and stops at the right pressure, and it becomes the humblest form of physical AI.
The catch is economic: over a year, it must cost the vulcanizer less than petrol, or he will not buy energy for his compressor ever again.
At the grander end, Zipline drones built in America serve over 1,300 health facilities and about six million people in three Nigerian states and some parts of Africa. TerraHaptix, a company led by Nigerians, recently raised $52 million to build drones and other physical AI for security – one of Nigeria’s current domestic challenges.
Other Nigerian proof points exist. Koolboks rents solar-powered refrigerators to traders.
PowerStove, makes a smokeless stove that generates its own electricity and tracks its use, so the company can sell carbon credits and share the income with users. It has sold over 200,000 stoves. Each founder built a business model, not just a machine. The Newsline era missed this because inventors or politicians are not automatically great managers of enterprises.
Nigeria is not starting from zero. With the United Nations Development Programme, the Federal Government has begun a network of university innovation pods, across the six geo-political zones. I have visited the UNDP-backed hubs in Maiduguri and Lagos, and there are others.
They are good. But buildings and equipments is not enough.
A hub that waits for applications will never meet the boy with the scrap-metal car. Hubs must hunt in workshops, markets and laboratories. They must run all year, untouched by strikes and semester calendars, and be led by people who have built businesses, products and protected intellectual property. The first pods are all in universities, yet polytechnics, where hands-on engineering lives, belong at the centre. The vulcanizer’s compressor or food grinding machine would make a fine first project: design it, produce it, and train young technicians to maintain it in every town. Count success in products shipped and jobs created, not ribbons cut.
Universities everywhere should also feel uncomfortable, because the old bargain is breaking.
Stanford economists found that since generative AI spread, workers aged 22 to 25 in the most
AI-exposed occupations have seen a 13% relative decline in employment, while older colleagues held steady or grew. Brynjolfsson’s explanation should alarm every vice-chancellor and policy maker. Experienced workers hold tacit knowledge learned on the job, the tricks of the trade that are never written down, while the young compete on textbook knowledge, exactly what AI copies best. And textbook knowledge is what the lecture hall sells.
In Nigeria, an estimated 600,000 graduates leave universities and polytechnics each year, into an economy that absorbs too few of them. If the jobs are not there, universities must help create them. Workshops and real customers build the tacit knowledge machines cannot copy. In an age of Liquid Knowledge, a degree that prepares a student for one fixed job prepares them for a world already gone. What a university should become deserves a larger debate, and I will open it in a coming column.
A Yoruba proverb teaches that ?w?´ ?ni la fi n tún ìwà ara ?ni ?e: one mends one’s own fortunes with one’s own hands. Physical AI is, quite literally, a matter of hands. The next young genius on Sunday television deserves more than applause and failed promises. She deserves a workshop, a partner, a mentor, and a market. Let the next drone over Katsina be built in Daura.
AI prompt of the Week: Build physical AI for your town
This week, design the machine that fixes it a problem in your community. Open any free AI assistant, i recommend Blueprint (blueprint.io), a tool that turns a single prompt into a buildable hardware plan with wiring diagrams, a parts list and assembly guides. Write this [personalise the details within the bracket]:
‘I live in [your town], Nigeria. The hardest manual task here is [describe it]. Design a simple, low-cost machine to help. Electricity is unreliable, so prefer solar or manual power. Use parts I can buy in a Nigerian market or online. Give me: how it works in plain words, a parts list with estimated prices, a wiring diagram if it uses electricity, step-by-step assembly, the main safety risks, and how a small business could sell or rent it in my town.’