Comparative Advantage: What Humans Are For
Stop asking what AI can’t do. Start asking what you’re relatively best at.
There’s a question that’s been eating at people since ChatGPT went mainstream:
What can humans do that AI can’t?
It’s the wrong question.
The Absolute Advantage Trap
When economists first tried to explain international trade, the intuitive model was simple: countries export what they’re better at making.
But that logic breaks down fast. What if one country is better at producing everything? Should the other country simply… not trade?
Ricardo’s answer, in 1817, was elegant and counterintuitive: it doesn’t matter who has the absolute advantage. What matters is comparative advantage — what each party gives up least to produce.
Even if England was more efficient than Portugal at producing both cloth and wine, both countries were better off specializing and trading. The math works out. Always.
We’re making the same mistake with AI.
The Real Question
AI has absolute advantage in a terrifying number of things.
Speed. Memory. Consistency. Breadth. Availability. Pattern recognition at scale. Text generation. Code. Analysis. Summarization. Translation.
If the question is “can AI do this better than me?” — the honest answer, for most tasks, is increasingly yes.
But that’s not the question that matters.
The question is: given that AI exists, what should I specialize in?
Not what AI can’t do. What I’m relatively best at — where the opportunity cost of AI doing it instead is highest.
What Humans Have Comparative Advantage In
Not because AI will never get there. But because right now, the gap is wide enough to matter.
Judgment under genuine uncertainty AI is extraordinary at situations with historical precedent. It extrapolates from patterns. Novel situations — ones with no prior data, no analogous case — still require a human who can sit with ambiguity and make a call.
Accountability and trust Someone has to be responsible. A client, a patient, a counterparty — they need a human who can look them in the eye and say I stand behind this. AI can inform the decision. It cannot own the consequences.
Contextual meaning AI reads everything ever written. It doesn’t know what it feels like to lose a client, to miss a deal, to rebuild after failure. Wisdom that comes from lived stakes is still a human domain.
Relationship as the work itself Sometimes the meeting is the product. The conversation is the value. No deliverable, no output — just two people building something through trust over time. AI can support this. It cannot replace it.
Partner, Not Tool
Here’s where the mindset shift matters.
We’ve been trained to think of AI as a tool. You pick it up, use it, put it down. You’re in charge. It executes.
But that frame leads to the wrong questions. How do I use AI? What can I automate? Where do I apply it?
A better frame: AI is a trading partner.
You bring what you’re relatively best at. AI brings what it’s relatively best at. The output is better than either of you could produce alone — not because one is smarter, but because specialization and exchange always beat self-sufficiency.
This reframes everything.
The question isn’t “should I use AI for this?” It’s “which of us has the comparative advantage here, and how do we divide the work?”
The Practical Implication
Ricardo’s insight led to a world where countries stopped trying to produce everything domestically and started building genuine specializations.
The same thing is about to happen to individuals.
The professionals who thrive won’t be the ones who resist AI, or the ones who blindly delegate everything to it. They’ll be the ones who develop a clear, honest answer to: what is my comparative advantage?
And then they’ll go very, very deep on that — while letting AI handle everything else.
Protectionism failed for countries.
It will fail for people too.
The goal isn’t to compete with AI. It’s to trade with it.