Wealth Follows Scarcity

In 1900, a skilled machinist in a factory town was paid well because there were not many of him.

By 1960, there were far more machinists, better trained, using better machines, producing far more output per hour.

Their real wages had risen, but the money in that industry had moved to the people who owned the plants, the patents, the distribution contracts, and the land under the buildings.

Nothing about the machinist's skill got worse.

The scarcity moved.

Part I ended with a blunt answer: you were being paid for scarcity, and the scarcity has moved.

Part II asks the obvious follow-up. When intelligence is cheap, what is still expensive?

Before naming the answers, you need the frame that generates them, because a list you have memorised is worth very little and a frame you can apply yourself is worth a great deal.

Price Is a Report on Scarcity, Not on Merit

A price is not a measure of how difficult something was or how much it deserves to be rewarded.

A price is the number at which the quantity people want equals the quantity available.

Difficulty and scarcity used to travel together, which is why the two got confused.

Medicine is hard to learn, so few people could do it, so doctors were scarce, so the price of their time was high.

The difficulty was real, but it was not what paid; it was only the mechanism that produced the scarcity.

Once a machine approximates an output at near-zero marginal cost, the difficulty of the human version stops producing scarcity, and the price stops following it.

This is why "but it took me ten years to learn this" is an accurate statement about your past and a useless statement about your future income.

Substitutes and Complements

Economists sort inputs into two relationships, and the whole of this Part rests on the distinction.

A substitute is something that can be used instead of another input, so cheapening one lowers demand for the other.

A complement is something that must be used alongside another input, so cheapening one raises demand for the other.

The rule that matters: when the price of an input falls, the value flows to its complements.

Cheap engines were terrible news for the market in draft horses and excellent news for the owners of oil fields, refineries, road networks, and repair shops.

Cheap electricity did not enrich the operators of factory line shafts, whose job disappeared, but it enriched everyone selling motors, appliances, wiring, and generating capacity.

Cheap computation did not enrich people who were fast at arithmetic; it raised the value of proprietary data, of software distribution, and of the firms sitting between a user and a market.

The pattern is the same each time: the substituted input fell in price, the complements rose, and ownership of the complements is where the durable money ended up.

What AI substitutes for

It substitutes for the production of routine cognitive output: first-draft text, translation, summarisation, standard code, standard analysis, standard design, standard research.

Not perfectly, and not everywhere, but well enough and cheaply enough that the price of the human version of the same output falls.

What AI is a complement to

Anything the system needs in order to be useful but cannot supply for itself.

It cannot supply a customer who has heard of you, or the trust that makes someone accept an answer without checking it.

It cannot supply the private data that makes a general model specific to one industry.

It cannot supply electricity, land, buildings, chips, or a body in a room, nor the capital that buys them.

It cannot supply a person who is legally and reputationally accountable for the decision.

It cannot supply the ownership of the workflow it runs inside.

Those seven gaps are not an arbitrary list, and the next lesson gives each of them a name.

Rents: Why Some Income Lasts

Economists use the word rent for income that comes from controlling something scarce rather than from the cost of producing it.

A landlord in a city with restricted building collects rent because there is only so much land near the centre.

A person with a twenty-year reputation in a niche collects rent because that reputation cannot be manufactured quickly.

Rents persist in proportion to how hard the underlying thing is to copy, and copying is the fastest form of competition.

So the durability of any income stream is roughly the durability of the barrier around whatever it rests on.

A skill learned from a public curriculum, applied to produce an output a machine can now approximate, has almost no barrier left.

A customer list, a licence, a power substation, a brand people trust, or a contract that names you as the responsible party each still have one.

The Copy Test

Here is the test to carry out of this lesson.

Ask of anything you own or do: can this be copied at near-zero cost, by someone who does not already know what I know?

If the answer is yes, its price is heading down, regardless of how good it is or how long it took to build.

If the answer is no, ask what exactly is stopping the copy.

There are only a handful of honest answers, and each one is a form of scarcity.

Legal barriers: a licence, a patent, a regulated approval.

Physical barriers: land, a building, a machine, a grid connection, a body in a particular place.

Informational barriers: data nobody else has, a process nobody else has documented, a relationship nobody else has.

Social barriers: trust, reputation, a network position, an audience that already pays attention.

Accountability barriers: someone must sign, and the signature carries liability that a system cannot carry.

Control barriers: you hold the account, the contract, the integration, the customer login.

And capital itself is the barrier that can buy several of the others.

Work through the test on a real example.

Suppose you write technical documentation at $65 an hour.

The writing itself is copyable at near-zero cost, so the price of the writing falls.

But the six-year relationship with the compliance officer is not copyable, the sign-off naming you as the competent reviewer is not copyable, and your archive of four hundred approved documents in that industry is not copyable.

The income was labelled "writing" but three of its four load-bearing parts were never writing at all.

That is the usual discovery, and it is usually good news.

Why This Means Owning, Not Just Doing

There is a trap inside all of this, and it catches careful people.

You can identify the right complement, become excellent at supplying it, and still capture almost none of the value, because being the complement and owning the complement are not the same thing.

A mechanic in 1925 was a complement to the cheap engine and was paid a wage.

The person who owned the garage, the parts inventory, and the corner lot was also a complement, and kept the residual.

Both were on the right side of the technology, and only one of them accumulated anything.

This is why the course runs laborer to architect to owner, and why its second half is about ownership rather than about skills.

What the Three Readers Do

Maya

Maya, 38, runs marketing at a mid-sized software company on a salary of about $145,000, and her team has gone from nine people to four in two years.

She applies the copy test to her week and finds that campaign production, where the four remaining people spend most of their time, is fully copyable.

What is not copyable: her relationships with the two largest customers' marketing teams, and the fact that she is the only person in the building who knows which past campaigns converted and why.

That second item is proprietary data sitting in her head and in a spreadsheet nobody else reads, and she notes that it may be the most valuable thing she owns at work.

Tom

Tom, 47, spent twenty years as a freelance translator and technical writer, earning about $90,000 a year until the work was largely automated and his income fell to about $38,000.

The copy test confirms the bad news in one line: translating an equipment manual is copyable at near-zero cost, which is why his rate collapsed.

It also surfaces three things that are not: his knowledge of which regulatory phrasings pass inspection, his archive of two decades of accepted documentation, and the fact that a manufacturer can put his name on a review as the accountable human.

Tom's problem was never that he lost a skill; it was that he was selling the one part of his work with no barrier around it.

Leo

Leo, 24, earns $52,000 in a customer-success role at a logistics startup, has $2,000 saved and $18,000 in student loans, and uses AI tools all day.

His first pass at the copy test is discouraging, because almost everything he personally produces is copyable, which is what being two years into a career looks like.

His second pass finds two real things: he sits between the company's customers and its product team, which is a network position, and he can see which complaints repeat, which is data.

He also holds the asset the other two do not, forty years of compounding time, and that is capital he can only spend by starting now.

Worksheet

1. List the five activities that produced most of your income last year, as tasks, not as a job title.
2. For each, answer the copy test: can a competent stranger with current AI tools reproduce this output at near-zero marginal cost?
3. Where the answer is no, name the barrier: legal, physical, informational, social, accountability, control, or capital.
4. Name the input in your work whose price has fallen most in three years, and estimate the percentage fall.
5. Name three complements to that cheapening input that your own work already touches.
6. For each of those three, mark whether you are the complement or the owner of the complement.
7. Estimate what share of your income came from rents (hard to copy) rather than production (easy to copy).
8. Write one sentence on what happens to your income if the copyable share halves over three years.

Common Mistakes

Treating difficulty as a moat

People defend their position by describing how hard their work is.

Difficulty was never the source of the income; scarcity was, and difficulty only mattered when it produced scarcity.

Assuming the complements are always technical

The most valuable complements to cheap cognition are usually unglamorous: a customer list, a licence, a location, a reputation, a signature that carries liability.

Chasing the newest technical skill often means chasing the input whose price is falling fastest.

Confusing being early with owning

Being the first person in your firm to use a new tool well produces a short-lived advantage and no asset.

Within eighteen months everyone in the building has the tool, and the only person still ahead is the one who converted the early advantage into a relationship, a dataset, or an equity stake.

Mistaking a high salary for a strong position

A salary is a claim on your own labour that ends the day you stop, and its size tells you what your labour is worth today, not how scarce it will be in five years.

Waiting for certainty about timing

Nobody can tell you how fast a given capability will be commoditised, and anyone offering a date is guessing.

Direction is more reliable than speed, and direction alone justifies shifting a small share of effort and surplus toward things that cannot be copied.

The RW Finance Perspective

The question "what is scarce here, and what stops someone copying it?" is not a career question we invented for this course.

It is the central question in business analysis, and it has been for a century.

When an analyst asks whether a company has a moat, they are applying the copy test to that company's earnings: what stops a competitor doing the same thing and competing the profit away?

The answers match this lesson: licences, patents, switching costs, network effects, scale, brand trust, control of a physical bottleneck, and access to data nobody else has.

That is why the Stock Quality Flower gives durability of returns as much attention as the size of returns, and why a Company Page is organised around understanding the business before discussing the price.

A business with high returns on capital and no barrier is a business whose returns are borrowed from the future, and the same is true of a person.

Applying the frame to yourself first makes you a better judge of companies later, because you will have practised the durability question where you have the most inside knowledge.

The next lesson, The Seven Scarcities, takes the seven gaps sketched here and gives each one a name, an explanation of why it stays scarce, and a description of how an ordinary person acquires a small stake in it.

Key Takeaways

• A price reports scarcity, not difficulty, effort, or merit, and the two were only ever correlated by accident.
• When the price of an input falls, demand and value flow to the complements of that input rather than to its substitutes.
• AI substitutes for routine cognitive production and complements distribution, trust, proprietary data, physical assets, capital, accountable judgment, and control rights.
• Durable income is rent, and rent lasts exactly as long as the barrier that prevents the underlying thing from being copied.
• The copy test asks whether something can be reproduced at near-zero cost by a competent stranger, and a yes means the price is heading down.
• There are only a few honest barriers to copying: legal, physical, informational, social, accountability, control, and capital.
• Being a complement to a cheap input earns a wage, while owning the complement earns the residual, and only the second accumulates.
• Direction is far more predictable than timing, which is enough to justify shifting effort and surplus now rather than waiting for certainty.
• The durability question you apply to your own income is the one a serious investor applies to a business.