Skip to content
Fritz Gerald ZephirinMenu

Social

Talent Is Evenly Distributed. Leverage Is Not.

4 min read

Start with a claim I should define before anyone can object to it. By “talent is evenly distributed,” I don’t mean every résumé looks the same everywhere — education, infrastructure, and opportunity differ brutally across places. I mean the raw material: the latent capacity for skill, judgment, and excellent work appears in roughly the same proportion in Port-au-Prince as in Palo Alto. What differs is everything around the person — and that everything is what I mean by leverage: tools, capital, rails, networks, and the chance to be in the game at all.

The strongest evidence for this view isn’t a slogan; it’s what happens when leverage arrives somewhere it wasn’t. In 2007, Kenya got M-Pesa — payment rails on a feature phone. The Science study that followed the rollout found access to mobile money lifted roughly 194,000 households — about two percent of Kenya — out of poverty, with the strongest effects for female-headed households. No one got smarter. No talent was created. A rail arrived, and ability that had always been there could suddenly compound. That is the cleanest controlled demonstration I know that the binding constraint was never ability.

Three gaps are now closing at once, which is why this decade is different.

The tool gap. A developer in Lagos and one in London now open the same model. Open weights made frontier-adjacent capability a free global commodity — by 2026, the most-downloaded models on earth come from a lab most Americans couldn’t name. And the field evidence says AI tools help the less-experienced most: 34 percent productivity gains for novices against single digits for experts. Read that finding geographically and it’s quietly radical — leverage compresses precisely the gaps that credentials used to enforce.

The access gap. Hiring across borders used to require a multinational’s legal department; now employment is becoming a protocol — the legal and payroll plumbing is commodity, and the country someone lives in is becoming metadata. This is the business I’m in, so discount accordingly; but what I watch daily is hard to argue with: given the same tools and competent local management, geography stops predicting performance. The variable that matters is never the passport.

The settlement gap. None of it works if the money can’t arrive. The World Bank still prices sending $200 home at over six percent on average — banks at fifteen. Every point of that friction is a tax on exactly the people with the least slack, which is why paying people reliably, everywhere isn’t an ops detail in this thesis. It’s the load-bearing rail.

Now the uncomfortable part. This argument is usually dressed as “talent arbitrage,” and arbitrage has an extractive reading: pay less for the same work, pocket the spread. I want to name that directly, because the extractive version is both wrong and self-defeating. The point of even distribution is that the spread closes — leverage flows to ability, wages converge, and the arbitrage, by succeeding, ends itself. A business built on the spread staying open is short its own thesis. The durable version is built on matching — role, person, time zone, price discovered honestly — and on being the rail that closes the gap, profitably, faster. Where convergence finally settles is a question I can’t answer and won’t pretend to; that it has started is visible from where I sit.

I sit in Miami, which is not incidental. Miami is a hinge city: US capital on one side, LATAM and Caribbean talent on the other, and a population that personally embodies the connection. Geography still matters — but its job changed. It used to determine what you could do; now it shapes what you can see. From here you can see both the capital that doesn’t know where the talent is, and the talent that never lacked anything but the rails.

The fair objection: maybe AI leverage compounds for elites instead — complementing the already-skilled until gaps widen, not close. The novice-gains evidence cuts against it so far, but population-level data will settle it, and this essay is hostage to that data. I’ll take the bet, though, because I watch it pay daily. The world is not short of talent. It has never been short of talent. It has been short of rails — and we are finally, unevenly, laying them.

Sources & further reading