The Human Premium
There’s a quiet inversion happening in how we assign value to services, and most people haven’t caught up to it yet
For most of economic history, human labor was the default. You went to a person because there was no alternative. The tailor made your clothes because there was no factory. The tutor taught your child because there was no scalable system. The accountant did your taxes because software didn’t exist. Human delivery wasn’t a feature. It was just how things worked.
That default is dissolving. And as it does, something counterintuitive is taking shape: the human-delivered version of a service is migrating from commodity to luxury.
The Inversion
Think about what’s already happened with physical goods. A machine-made shirt costs almost nothing. A hand-tailored shirt from a craftsman in Naples costs thousands. The functional difference is marginal. The status difference is enormous. We pay the premium not because the hand-stitched shirt keeps us warmer, but because a human chose to spend hours on something a machine could do in seconds. The scarcity of that attention is what creates the value.
Now apply that same logic to services. Today, if you want financial advice, you talk to a person. If you want a medical opinion, you see a doctor. If you want your child educated, you send them to a classroom with a teacher. In most of these cases, the human isn’t a luxury. The human is just the delivery mechanism.
But what happens when an AI system can deliver 90% of that service at near-zero cost, available 24/7, personalised to the individual, and constantly improving? The functional gap between the AI version and the human version narrows. And in some cases, the AI version becomes objectively better on measurable outcomes.
At that point, the reason you choose the human version changes entirely. You’re no longer choosing it because it’s the only option, or even because it’s the best option. You’re choosing it because it’s human. Because there’s a person on the other side who chose to show up, who is present, who is finite.
That’s a luxury good.
Education Is the Clearest Case
I’ve been thinking about this most concretely in the context of education, because the gap between what AI could offer and what most students currently receive is already stark.
Imagine a classroom where an AI system, potentially even embodied in a humanoid form, is connected to an enormous knowledge base. It can monitor every student’s engagement in real time. It can detect confusion before the student raises a hand. It can adjust its teaching style, its pacing, its examples, all individualised per student. It never has a bad day. It never loses patience. It never teaches to the middle of the class while the top and bottom drift.
I think about my own education and how different it could have been. As I mentioned in a recent article exploring the opportunities and threats in educational AI, I remember learning about a concept biology and being curious about what drove that process at a deeper level, about the psychology of how living systems respond to their environment. But the curriculum didn’t connect those threads. A system that could detect that curiosity and pull on it, linking biology to chemistry to psychology in real time, would have changed the trajectory of what I found interesting.
Now, in that world, what does a human teacher become? Not obsolete. Something different. The human teacher becomes the version you pay a premium for. The one who sits across from your child and builds a relationship that has texture and warmth and unpredictability. The one who can share a story from their own life that an AI never lived. The one whose attention is finite and therefore meaningful.
That’s the inversion. The human teacher goes from the default to the premium tier.
This Extends Everywhere
Education is just the most emotionally legible example, but the pattern is general.
A human financial advisor who sits with you and understands the anxiety behind your questions, not just the math of your portfolio. A human therapist whose own experience of suffering informs how they listen. A human chef who decides, based on instinct built over decades, to add something unexpected to a dish. A human doctor who notices something in your expression that doesn’t show up in your bloodwork.
In every case, the AI version will likely be more consistent, more available, and in many measurable dimensions, more accurate. But the human version will carry something the AI version cannot: the weight of a finite life choosing to spend its time on you.
That weight is what luxury has always been. We’ve just never had to think of basic service delivery in those terms.
The Counterargument Worth Taking Seriously
There’s a reasonable model where the human doesn’t get replaced or elevated to luxury status, but instead gets augmented. A teacher enhanced by AI tools who can focus entirely on the social and emotional dimensions of learning while the AI handles content delivery and assessment. A doctor who uses AI diagnostics but spends their time on the human judgment calls and the bedside manner.
I think this hybrid model is real and will exist. But I also think it’s a transitional phase. The augmented human is still more expensive than the fully automated version. And as AI systems improve at handling the social and emotional dimensions too (they’re already getting better at detecting emotional cues, at adjusting tone, at simulating presence), the gap that the human fills in the hybrid model shrinks.
What remains, eventually, is the irreducible fact that the human is human. Not that they’re better. Just that they’re real, and present, and mortal. And that will be enough to command a premium, the same way a hand-painted canvas commands a premium over a perfect digital reproduction.
The Uncomfortable Part
There’s a darker edge to this that’s worth sitting with. If human-delivered services become luxury goods, then access to human attention becomes stratified by wealth in ways it currently isn’t. Today, most children have a human teacher. In the future I’m describing, most children might have an AI teacher, and only wealthy families pay for the human version.
We already see early versions of this. Wealthy families hire human tutors while everyone else uses something like Khan Academy. Wealthy patients get concierge doctors while everyone else sits in a waiting room. The pattern isn’t new. But AI could accelerate it dramatically by making the non-human version so good, so available, and so cheap that the human version becomes a genuine luxury rather than a marginal upgrade.
Whether that’s dystopian depends on whether the AI version is actually good enough. If an AI teacher genuinely provides a better educational experience for most students than the average human teacher (which is a real possibility, given the constraints most teachers operate under), then the stratification might be less troubling than it sounds. The luxury isn’t necessarily better. It’s just more human.
What This Means for How We Think About Work
If you’re someone who delivers a service, whether that’s teaching, advising, creating, or caring for people, the question isn’t whether AI will be able to do what you do. In most cases, it will. The question is whether the human-ness of how you do it becomes the thing people value.
This isn’t about “soft skills” in the corporate sense. It’s about the fundamental scarcity of human attention in a world where artificial attention becomes abundant. The craftsman who takes three days to make a chair isn’t competing with the factory on efficiency. They’re competing on a completely different axis: the fact that a person made this, and that means something.
The same shift is coming for services. The sooner we understand that, the better positioned we are to think clearly about what human work means in the next decade.
The human premium isn’t about being better than the machine. It’s about being human in a world where that’s no longer the default.
We had this conversation in a group last week, it’s a very interesting discussion and curious to see the development