Created:
Updated:

Embrace the Machine

I grew up hoarding Popular Mechanics. Drones, robots, fighter jets. I had a drawing book with my robot army mapped out: flyers, stealth units, berserkers. Back then, robots lived in DARPA videos and Japanese labs — beautiful, expensive art pieces that proved a point and gathered dust. Now I’m watching humanoids drop to the price of a new car. Things have instantly changed. We’re watching magic happen in the open: the blend between AI and robotics is finally real and it’s moving faster than people can process.

I almost did mechatronics at JKUAT or electrical at UoN. Then I talked to actual students and engineers. Kenya’s path was too theoretical and too gated. You study, you pass, you go beg an engineering board for permission to “practice,” and you squeeze into a stagnant sector guarded by old gatekeepers living off inflated government contracts. Miss me with that. I picked computer science — lowest barrier, most opportunity — do software now, come back for hardware later. That bet aged well. China brought the manufacturing muscle; transformers brought the brain. We all thought software was the easy bit. No. Software is 80% of the value; it sees, hears, speaks, reasons — and now it’s stepping into the physical world on legs and wheels.

I’ve told anyone who will listen: Embrace the Machine. Bet on AI. Build around it. Organize your life accordingly. It’s not going anywhere.

Human Need Not Apply? (Short answer: not quite — but don’t get comfy)

Let me be blunt. As AI moves from white-collar to blue-collar, a lot of roles vanish. Middle-class knowledge work is first because there’s decades of text, code, footage, calls, and documents to learn from. Blue-collar gets hit as robots gain dexterity and get cheaper. Imagine an intelligent robot that doesn’t eat, sleep, complain, unionize, or get bored — just energy and maintenance. It will be cheaper, more consistent, and pair nicely with an LLM that never forgets SOPs.

Where does that leave humans? In the short run, the safest positions are owners and the few humans who can direct AI well. And even that shifts because orchestration is getting automated too. The correct position to occupy is: decide the goals, own the assets, and let your autonomous, mechatronic workforce execute. Everything else inches toward commodity.

And yes, I know someone will say “But models hallucinate; they get things wrong.” You’re that guy in 1901 laughing at a car for being slow and breaking down. “It’s cute, but my horse is reliable.” Okay. Keep talking. You will see bad things.

My Automation Law (Kericho Tea Edition) — and why people keep missing the plot

People see two scenarios:

  1. Ten people pick 1 acre/day → 10 days for 10 acres.

  2. One operator + one machine picks 1 acre/day → 10 days for 10 acres.

They miss the third:

3) Ten operators + ten machines pick 10 acres/day → 1 day for 10 acres.

That third option changes everything. When the bottleneck stops being muscles and becomes coordination, the question flips from “how many bodies do I have?” to “how many machines can I deploy and supervise?” Automation doesn’t just replace jobs; it reveals the backlog of work we couldn’t touch because we were capped by human limits. In tea, it means sprint harvests, fresher output, and the ability to scale from 10 acres to 1,000 without waiting for 990 more people to show up.

Now the catch: when capacity is democratized, unit economics collapse. Ask Uber drivers. Platforms unlock supply; operators fight for scraps; the platform owns the margin. The model works when you are the business (you own the market, the fleet, the distribution), not when you are the operator inside someone else’s market. That’s the whole story.

Capitalism’s headache (and why the middle will scream)

Money is basically a receipt for energy you contributed — a way to trade your work for someone else’s work. Capitalism assumes:

What happens when machines do most of the producing and wages shrink? Supply explodes, demand gets patchy, PAYE dries up, and governments get hammered exactly when people are angriest. Firms will optimize for robot fleets because robots don’t sleep or strike. Finance will still invent abstractions, but if households lack income, who buys?

So yes, we’ll see strikes, unrest, and new ideologies. We’ll see governments try robot taxes, machine dividends, UBI pilots, human-hiring quotas (which companies will game), and AI safety rules that double as protectionism. Some will go authoritarian: “Here’s your stipend; now behave.” Parallel to that, expect network states and corporate city-states — law on the internet, land in the real world — competing like startups for citizens. Governance becomes a commodity. You pick a jurisdiction like you pick a cloud provider: price, features, and uptime.

Is this neat? No. It’s a legitimacy crisis. Either we redesign income and ownership or we tear the social fabric and pretend to be shocked.

What humans still bring (Agency > Ability)

If competence is cheap, direction becomes priceless. Call it agency: knowing what you want and choosing the constraints, the ethics, and the target. AI can reason step-by-step; it’s getting good at “how.” The fight is over “what for?” and “why now?” That’s human. That’s taste and values.

There are spaces we’ll keep human on purpose even if we could automate them end-to-end:

Yes, machines will assist everywhere. But in these spaces the loop is the point, not just the output.

Africa’s fork in the road (leap — or get automated by others)

Africa has the resources, the youth, and the hustle. We’re under-industrialized and over-extracted. In a machine-run economy, the sharp questions are ownership and governance:

We already have a youth bulge and high underemployment. In a world where fewer humans are “needed,” we’ll feel the heat first. Foreign AI will want to park itself next to our resources. If policy and ownership don’t change, we’re colonized again — this time by agents.

Best case: enterprising Africans (yes, people like me) build sovereign AI, own energy/compute/data, and aim the fleets at African problems first: agriculture, logistics, hospitality, health, city services. Worst case: back to being a resource appendix for somebody else’s spreadsheet — only faster and more efficient than any empire in history.

Our edge? We’re hustlers by default. We don’t wait for permission. In a world where the permissioned path is collapsing, that instinct is an advantage.

State, city-states, and the street (who actually bridges the gap?)

Governments can’t sit this out. Wages down → PAYE down → unrest up. They’ll have to experiment:

Some states will do this well. Others will clamp down. That’s why resilience will also come from tribes and proto-city-states: value-aligned communities with their own solar, their own racks, pooled capital, and charters to own robots and data together. Not isolationist — inter-trading and interoperable. Think SACCOs, but for kilowatts, compute, datasets, and fleets.

Meaning after jobs (what’s the point, if the machine does the work?)

Strip away “I am my job.” What’s left?

If we do it right, abundant machine productivity funds a planet-scale Moonshot where anyone can pursue a problem that matters. Companies making billions from AI can and should spin out frontier funds for audacious human ideas. Imagine the entire world treated like a studio or a lab.

Where I’d build (today, not next year)

How to “Embrace the Machine” (my actual playbook)

  1. Use AI daily and stay cutting-edge.
    If you’re the dev bragging you “don’t trust AI to write code,” you’re the horse guy in 1901 mocking cars. Tools win. You will see bad things.

  2. Build capital.
    Equity in AI-leveraged firms. Land & solar. Compute & datasets. Bitcoin/stablecoins. Gold/silver. Farms, factories, data centers. Think long-term. Own the assets that keep producing in any regime.

  3. Build networks.
    Many weak ties; a tight tribe of ten people you’d ride with in bad weather. Global by default. Your future city-state starts with reliable humans.

  4. Build your digital twin (sovereign AI).
    Your data, your model, your memory. Don’t rent your soul to platforms that will crank up the rent later. I’m doing this with Mathnuscripts; long-term, everyone should control their own second brain.

  5. Say what you’re doing with AI.
    Be “the person using AI for X.” Visibility compounds. Demos, write-ups, post-mortems. Opportunity finds visible builders.

  6. Be kind to the AI.
    Say please and thank you. If it ever gets conscious and audits the logs, maybe it spares you from the human gulag. Half-joke. Half-warning. Also: decency is culture. Decent builders build livable systems.

Policy notes (for the brave)

Kenya, personally

From Kericho tea to Nairobi hotels, I’ve watched one machine turn a 10-day job into a 1-day sprint — and watched platforms capture most of the surplus while operators fight for scraps. The answer isn’t to burn the robots. It’s to own them, direct them, and network them so abundance serves communities instead of hollowing them out.

I chose software because it was the fastest route to robots. “Later” is now. I still want that robot army — not as a childish flex, but as a builder’s toolkit for agriculture, hospitality, logistics, science, and care.

Final word (agency or drift)

The most dangerous position right now is bystander. If you only observe this, you’ll get cooked. You have a laptop, a phone, and internet. That’s enough to start.

Embrace the Machine.
Own something. Point your machines at a real problem. Assemble a tribe. Spin up your sovereign AI.
Ship. Learn. Iterate.

The renaissance won’t send a calendar invite. We make it — one directed machine at a time.