The Artificial Person

A company is sometimes described as an artificial person. The phrase is usually legal: a company can own property, enter contracts, borrow money, employ people, sue, and be sued. It has a name and a history that remain intact even as the people inside it change. It can survive its founders. It can accumulate obligations and make promises about a future none of its current employees will personally see.

I think the phrase is useful beyond the law. A company really does acquire something like a life. It consumes capital and attention. It develops habits. It remembers some decisions and forgets others. It can become disciplined, confused, useful, extractive, or vain. Most importantly, it acts in the world through products. The product is where the artificial person makes contact with everyone else.

This leads me to a simple view: a company should exist to make something useful. The thing does not need to be physical, and usefulness does not have to mean necessity in the narrowest sense. Software can be useful. Entertainment can be useful. A payment network, a laboratory, a restaurant, a game, or a logistics service can all produce real value. But there must be some point at which the company stops describing itself and does something for another person.

The thing itself

Companies are surrounded by representations of themselves. There is the pitch deck, the brand, the valuation, the cap table, the press release, the roadmap, and the story told to investors. These things are not automatically bad. A company needs language to coordinate people and capital to operate. A plan is a representation of work that has not yet happened, and every serious project begins with some version of one.

The problem begins when the representation becomes more important than what it represents. A high valuation is treated as proof of a good company. Attention is treated as proof of demand. Fundraising is treated as proof of progress. The company becomes successful on paper before it has become useful in practice.

A product interrupts this confusion because it can be tested. Does it work? Does anyone use it? Does it make an expensive process cheaper, a difficult task easier, or a previously impossible action possible? Will someone choose it when the launch is over and no one is watching? These questions are less glamorous than questions about market size, but they are closer to the truth.

This is not an argument against ambition. The largest companies often need years of research and capital before their products become practical. A biotechnology company may run experiments for a long time before it can sell anything. An infrastructure company may need to build expensive systems before the first customer can use them. Usefulness can be delayed. It cannot be permanently replaced by the promise of future usefulness.

One clear thing

I also think a company should be simple, although simple is easy to misunderstand. It does not mean that the underlying work must be easy. Biology is not simple. Manufacturing is not simple. Financial systems are not simple. The company can be working on a hard problem while remaining simple in purpose.

A simple company can answer basic questions without changing the subject. What does it make? Who is it for? Why is it better than the available alternative? What must become true for the company to work? Complexity inside the product may be unavoidable. Confusion about the purpose of the company is not.

Simplicity creates constraints. It prevents every new technology from becoming a strategy and every possible customer from becoming a market. It makes it easier to reject work that is interesting but irrelevant. It also exposes failure. If the company has one clear purpose, it is difficult to hide from the fact that the purpose is not being achieved.

Complicated stories are attractive partly because they make failure hard to locate. A company that is simultaneously a protocol, a community, a marketplace, a financial asset, a social movement, and an infrastructure layer can reinterpret almost any outcome as progress. If the product is not used, perhaps the community is growing. If the community disappears, perhaps the token price is rising. If the price falls, perhaps the protocol is early. The number of explanations increases while the amount of evidence stays the same.

The company on paper

Capital markets create a second version of a company: the company as a financial object. This version is made of claims about future cash flows, risk, ownership, and probability. It is necessary. Without a way to price uncertain futures, many difficult projects would never receive the money required to begin.

But the paper version and the operating version are not the same thing. The paper version can change very quickly because expectations can change very quickly. The operating version remains attached to slower facts: laboratories, customers, software, supply chains, employees, machines, permits, and time. A company can double in value during a week in which almost nothing about its product changed. It can also become much more useful during a year in which its valuation falls.

Finance is most valuable when it helps the operating company do real work. It pools savings, distributes risk, funds experiments, and allows people to build before all the required resources are already in hand. It becomes dangerous when movement in the paper version is mistaken for movement in the world.

I call this mistake the paper game. The score is measured in valuation, volume, price, or the cleverness of the financial structure. The score may be consequential—paper claims determine who controls resources—but it is still a score. A company ultimately has to leave the game and encounter reality.

The closed circle

This is the source of much of my dislike for crypto and Web3. I do not mean that cryptography is useless, that distributed systems have no legitimate applications, or that every project in the category is fraudulent. The narrower criticism is that many crypto businesses form a closed loop in which the financial representation is also the product.

A token is issued. The token becomes valuable because people expect the network to become valuable. The network appears valuable because the token has a price. Rising prices attract users whose activity is then presented as evidence of demand. Often the main thing those users want to do is acquire, trade, lend, stake, or speculate on more tokens. Capital moves, software runs, and fees are paid, but it can remain difficult to identify the external problem being solved.

Traditional finance can become self-referential too. It produces its own abstractions, incentives, and games. The difference is not that shares and bonds are inherently real while tokens are inherently fake. The relevant question is what sits underneath the claim. A share can represent ownership in a company that makes medicine, moves freight, or sells useful software. A bond can finance a factory or a power plant. A financial instrument earns its seriousness from the activity it makes possible, not from the sophistication of the instrument itself.

My objection is therefore not primarily aesthetic or ideological. It is about the direction of causality. A useful product creates value, and a financial claim may be built on top of that value. In the closed loop, the claim is created first and everyone waits for usefulness to emerge underneath it.

Beyond the chart

When I say that companies should be concrete, I do not mean that every company should manufacture an object. I mean that its value should become visible in an outcome outside the company's own accounting. A good database reduces the cost and risk of storing information. A good design tool changes what a person can make. A good insurance product allows someone to survive a loss that would otherwise destroy them. These are concrete effects even when the product is entirely digital.

Concrete products also create honest constraints. In aquaculture, an elegant presentation cannot correct poor water quality. Software cannot claim that a sensor is working when it is not. A system has to operate under physical conditions, with real costs and real failure modes. The facts eventually become unavoidable.

I value that kind of constraint. It forces a company to distinguish between what has been said, what has been built, and what has actually worked. It does not eliminate exaggeration or bad judgment, but it makes both easier to discover.

The reason to exist

An artificial person has no automatic right to attention, labor, or capital. Its existence has costs. People spend years inside companies. Investors give up the use of money elsewhere. Customers reorganize parts of their lives around products. Communities absorb the consequences when companies fail or behave badly.

The return for those costs should be more than a successful narrative. The company should make something worth having, improve a process worth improving, or discover something worth knowing. It should be possible to point to the value without first opening a chart of its valuation.

This is a demanding standard, and many young companies will not meet it immediately. They are experiments. Some will be wrong, and failure is not evidence that the attempt was dishonest. The important distinction is whether the experiment is trying to produce value in the world or merely trying to produce the appearance of a valuable company.

A company can be ambitious, strange, and technically complex. It can take a long time to work. But its purpose should remain intelligible. The artificial person should eventually be able to answer the same question asked of any person entrusted with resources and responsibility: what did you do that was useful?