Face Value: Trust, Liability, and the Human Face in an Age of Perfect Copies

Chapter 1

The Call From Your Daughter

You do not wire the money. You hang up, and you call her back on the number you have had for years. She answers from her kitchen. She is fine. She has no idea what you are talking about.

Sit with what just happened, because it is the whole book in one minute. The voice was perfect. Nothing in the sound of it told you the truth. What saved you was not your ear. It was a rule you reached for without thinking: when the signal can be faked, go around the signal to something that cannot be. You went back to a number you knew. Next year, for a larger sum, you will want more than that. You will want to see her.

For most of history a voice was proof. If you heard your mother on the phone, it was your mother. Nobody had to teach you that. It was true because making a convincing copy of a particular human voice was hard, slow and expensive. The difficulty was the guarantee. You never saw the guarantee, because it was built into the world.

That guarantee is gone. A few seconds of someone talking is now enough to make them say anything. The same is true of a face on a video call, a signature on a page, the tone of an email, the style of a text message. Each of these was trusted for one reason only. It was hard to fake. When it stopped being hard to fake, it stopped being trusted, whether or not anyone announced it.

When a signal costs nothing to produce, it stops carrying information.

This is not a technology story. It is a story about information, and it is very old. A signal tells you something only when it is easier for an honest sender to produce than for a dishonest one. A message anyone can send says nothing about who sent it. The moment AI made every voice, face and writing style cheap to produce, all of them quietly became messages anyone can send.

What is left

So the question this book asks is simple. When everything that travels down a wire can be copied, what still cannot be? What will people trust?

My answer is the oldest one there is. They will trust a face. Not a picture of a face. The face itself, in a room, at a time, attached to a body that can only be in one place at once. They will trust the person who shows up, who can be found again tomorrow, and who has something to lose if they lie.

That sounds like nostalgia. It is not. It is what happens to trust when you follow the logic all the way down. Every other signal has a price that AI is driving toward zero. Presence has a price that physics sets, and physics does not negotiate. You cannot be in two rooms. You cannot get back the hour you spent driving to see someone. You cannot send a copy of yourself to shake a hand.

What this book will do

In the chapters that follow I want to lay out why that is true, from first principles, and what it means for how we work, buy, hire, raise children and build businesses over the next ten years.

Part One explains why trust has always been a matter of cost. Part Two explains why trust needs a face: someone you will meet again, and someone you can hold responsible. Part Three follows the money: when AI makes the work cheap, value moves to the scarce thing beside it, and that thing is an accountable human. Part Four is practical. It is about living face-first, as a worker, a business owner and a family. Part Five makes predictions and puts them in a ledger, so you can check me.

The thesis fits in a sentence. A copy costs nothing. A face costs a body, a place, and an hour you cannot get back. That cost is not a flaw in the human face. It is the reason the human face is about to become the most valuable thing in the economy.

Chapter 2

Expensive on Purpose

Why does a peacock drag around a tail that makes it slow, visible and easy to catch?

For a long time the tail was a puzzle. In 1975 the Israeli biologist Amotz Zahavi offered an answer that sounded backward. The tail is valuable because it is a burden. Only a healthy bird can afford to carry it and still survive. A sick bird with a magnificent tail would be eaten. So the tail is an honest advertisement of fitness, and it is honest precisely because it is expensive. Zahavi called it the handicap principle.

Two years earlier, an economist named Michael Spence had found the same idea in a completely different place: the job market. An employer cannot see how capable you are. You can tell them, but everyone says they are capable. Spence showed that a credential can carry information when it is cheaper for capable people to earn than for others. The degree does not have to teach you the job. It only has to be harder to get if you are not the kind of person the employer wants. The cost is what does the talking.

The rule underneath

Put those two ideas side by side and you get a rule that runs through all of human trust:

• A signal is believable only when it is costly to fake.
• The cost has to fall harder on the liar than on the honest sender.
• When the cost disappears, the signal does not become false. It becomes empty.

That last line is the one people miss. A cheap signal is not a lie. It is noise. A beautiful letter used to tell you the writer cared enough to spend an evening on it. Now it tells you nothing about the evening. It might still be sincere. You simply cannot tell from the letter.

A short history of expensive things

Look back and you will see that every trusted signal in human history was trusted for its cost.

The wax seal on a letter was trusted because the signet ring was hard to copy and stayed on one person’s finger. The handwritten signature was trusted because forging a hand takes skill and practice. Letterhead meant something when printing was costly and a fake business could not afford it. A phone call from a familiar voice meant something because nobody could synthesize your grandmother. A video meant something because you could not fake a moving, speaking face without a studio.

Each of these signals worked for a while and then failed in the same way. A technology arrived that made it cheap to copy, and the trust drained out of it. Printing made letterhead cheap. Photocopiers made documents cheap. Image editing made photographs cheap. Each time, people moved their trust one step up, to whatever was still hard to fake.

Trust has never lived in the message. It has always lived in what the message cost.

What AI does to the list

AI is not one more rung on that ladder. It takes out most of the remaining rungs at once. Text, voice, face, video, style, tone, handwriting, the particular way a friend phrases a joke. All of it is now cheap, and getting cheaper every month.

That is why the moment feels strange. People are not used to losing several signals at the same time. They are used to losing one and moving up. This time, when they look up for the next rung, most of the ladder is gone.

What remains is the thing that was always at the top, the thing all the other signals stood in for. A person, physically present, whose time is finite and whose face is attached to a life that can be found.

Expensive on purpose

There is an old habit in human affairs of making important things deliberately costly. Weddings are expensive. Funerals require travel. Treaties are signed in person. Deals are closed over dinner. We call it ceremony and we sometimes think it is waste.

It is not waste. It is signaling. Flying across the country to be at someone’s wedding says something no card can say, because the card is cheap and the flight is not. The cost is the message. Zahavi would have recognized it at once.

In the coming decade we are going to rediscover ceremony, not because people become sentimental, but because ceremony is the last place where the cost cannot be faked. When a message costs nothing, the visit becomes the message.

Chapter 3

The Ladder of Copies

Imagine a ladder. Each rung is a way of proving who you are and that you mean what you say. At the bottom are the cheap rungs. At the top are the expensive ones.

For most of the last century, the ladder looked something like this, from bottom to top: a typed message, a printed letter, a signed letter, an email from a known address, a phone call in a known voice, a video call, a meeting in person.

Most of daily life ran on the middle rungs. You did not need to meet your bank. A phone call in your voice, with a few questions answered, was enough. You did not need to see a new client. An email from the right address, a call, maybe a video meeting, and the work began. The top rung, being in the room, was reserved for big moments: hiring, closing, marrying, burying.

Copies climb the ladder

Every technology of copying climbs this ladder from the bottom. First it makes the lowest rung cheap, then the next, and trust retreats upward ahead of it.

For a long time the climb was slow. It took centuries to go from forged signatures to forged documents, decades to go from forged documents to edited photographs. AI climbs fast. In a few short years it has reached the rung of the voice and the rung of the video call. The middle of the ladder, where most of the economy actually lives, is the part being taken.

That is the important point. AI is not threatening some exotic corner of trust. It is taking the rungs where ordinary business happens: the call to the vendor, the email from the boss, the video meeting with the client, the voice of a family member asking for help.

The middle of the ladder, where most of daily life lives, is the part AI is taking.

Figure 1

The ladder of copies

1. A person in the roomHolds
2. A video callFakeable
3. A familiar voice on the phoneFakeable
4. An email from a known addressFakeable
5. A signed letterFakeable
6. A printed letterheadFakeable

Why the top rung holds

The top rung is different in kind, not just in degree. It holds because of three facts about bodies that no software changes.

First, a body is in one place at a time. When you are in my office you are not anywhere else. A copy can be in a million places. A person can be in one. That exclusivity is a cost, and costs are what make signals honest.

Second, presence spends time you cannot refund. If you drive forty minutes to see me, those forty minutes are gone from your life. You could have spent them on anything else. Economists call this opportunity cost. I call it proof. The only way to show someone they matter to you is to spend something on them that you cannot get back.

Third, a person in a room is continuous. They are the same person who was here last month and will be here next month. They can be recognized by the people around them. They have neighbors, a car in the lot, a name on the door. A face in a room is attached to a whole life that can be found.

The thing all the other rungs stood for

Here is something that becomes obvious once you see it. Every rung below the top was always a substitute for the top. A signature stood in for you being there to agree. A phone call stood in for you being there to speak. A video stood in for you being there to be seen. We built the whole ladder because being there was expensive, and we wanted cheaper ways to approximate it.

For a long time the substitutes were good enough, because they were hard to fake. Now they are easy to fake, so they are no longer good substitutes. And when a substitute fails, people go back to the original.

That is the movement I expect over the next decade. Not a collapse of trust, but a return of trust to its source. The substitutes will still be used for small matters, where being fooled costs little. For anything that matters, people will climb back up to the top rung.

What this does not mean

It does not mean everyone will stop using the phone or the internet. It does not mean every purchase will need a meeting. Most of life is small, and small matters will run on cheap signals, increasingly handled by AI on both ends.

It means the line between small and large is about to become very visible. Below the line, machines talk to machines and nobody minds. Above the line, people want a face. Learning where that line sits, for your work, your customers and your family, is one of the most useful skills of the coming years.

Chapter 4

The Shadow of the Future

In 1980 the political scientist Robert Axelrod invited experts to send him computer programs that would play a simple game against each other, over and over. In each round, two players could cooperate or betray. Betraying paid more in a single round. Cooperating paid more if both did it. The question was which strategy would win over many rounds.

The winner was the simplest program entered. It was called tit for tat, submitted by the psychologist Anatol Rapoport. It cooperated on the first move and then did whatever the other player had done last. It was nice, it retaliated, it forgave, and it was easy to read. Axelrod wrote it all up in a book, The Evolution of Cooperation, and gave the key condition a lovely name: the shadow of the future.

Why cooperation needs tomorrow

Cooperation, Axelrod showed, only pays when the players expect to meet again. If this is the last round, betrayal wins. If there are many rounds ahead, a reputation for being trustworthy is worth more than anything you could gain by cheating today. The future casts a shadow backward over the present, and that shadow is what makes people behave.

This is why trust in a small town feels different from trust in a big city. It is why you are more careful with the mechanic you will see for twenty years than with a stranger at a highway exit. It is why the corner store can run a tab and the airport kiosk cannot. The shadow of the future is long in one place and short in the other.

Trust is not a feeling. It is what people do when they expect to meet again.

A face is how you know it is the same player

Here is the part that matters for us. The shadow of the future only works if you can recognize the other player. Tit for tat needs to know who betrayed it last time. Reputation needs a name that sticks.

For most of human history, the thing that let you recognize someone was their face. You saw it again, you knew it, you remembered what happened the last time. A face was an identity that you could not easily swap out. It was the anchor that made reputation possible.

Digital life weakened that anchor long before AI. An account can be deleted and replaced. A seller with bad reviews can reappear under a new name. A caller can use any number. Every new identity resets the game to round one, where betrayal pays.

AI makes new identities nearly free and infinitely persuasive. A new voice, a new face, a new writing style, a new history, all generated in seconds. In a world like that, the shadow of the future disappears from every interaction that can be entered with a fresh identity. Every one of those interactions becomes a last round.

What people will do about it

People are not stupid about this. They will do what tit for tat would do if it were a person: they will look for the players they can recognize, and they will save their cooperation for them.

In practice that means a strong preference for people who are physically findable. The contractor with a yard and a truck. The adviser with an office. The shop owner who is behind the counter every day. The colleague you have had lunch with. These people cannot reset the game. Their face is their account, and they cannot open a new one.

It also means a new value on continuity. The longer someone has been findable in the same place, under the same face, the longer the shadow of their future, and the more they can be trusted. Seniority in a place will matter again. So will roots.

The quiet advantage of staying put

There is an irony here. For a generation the advice was to be mobile, to work from anywhere, to treat place as an inconvenience. Some of that will remain true for the work machines do. But for the work that depends on trust, staying put becomes an asset. A face that has been in the same town, at the same counter, answering for the same work for years, carries a shadow of the future that no new account can match.

You cannot fake having been there. That is exactly why it will be worth so much.

Chapter 5

Someone to Hold

Nearly four thousand years ago, the code of Hammurabi set out a rule for builders. If a builder built a house and it collapsed and killed the owner, the builder was to be put to death. If it killed the owner’s son, the builder’s son would die.

We are rightly horrified by the punishment. But look at the principle, because it has never gone away. The person who makes the thing must carry the risk of the thing. The builder had to live under the roof of his own promise. Nassim Taleb built a whole book around this idea and called it skin in the game.

Responsibility needs someone who can lose

Every system of trust eventually comes down to this question: if it goes wrong, who pays? Not in the abstract. Who, specifically, loses money, reputation, freedom or sleep?

A promise from someone who cannot lose anything is not really a promise. It is a statement. The reason we trust a surgeon is not only that she is skilled. It is that her name, her license and her livelihood are attached to the outcome. The reason we trust a pilot is partly that he is sitting in the front of the same plane.

Now ask the question about an AI system. If your AI makes a mistake in your taxes, your contract or your medication schedule, who pays? The model cannot be fined in any way it would feel. It cannot lose its license. It cannot be ashamed. It does not lie awake. It cannot be sued in the sense that matters, because there is nothing it cares about that a court can take away.

A promise from someone who cannot lose anything is not a promise. It is a statement.

The machine cannot hold the risk

This is not a complaint about AI. It is a fact about what AI is. In my own thinking, AI is an entity you manage, not a tool you use. It is a delegate. Delegation runs downward, from the person who holds responsibility to the one who carries out the work. The responsibility does not travel down with the task. It stays with the person who delegated.

That has a direct consequence. However capable AI becomes, someone has to stand behind what it does. Some person must be the one who answers the phone when it goes wrong, who signs the contract, who carries the insurance, who can be found. The more work AI does, the more important that person becomes, because more and more of the world’s output will need a human to answer for it.

The architect’s stamp

We already have a model for this, and it is old. An architect’s drawings are not legally valid until a licensed architect stamps them. An audit is not complete until a certified accountant signs it. A prescription is not filled without a pharmacist’s check. A ship has a master who answers for everything aboard. A building permit names a responsible person.

In each case, much of the work may be done by others: junior staff, drafters, software. But one named human puts their face on it and says, I answer for this. The stamp is not a technical step. It is the attachment of a person who can lose something to a piece of work that cannot.

That arrangement is about to spread to everything. When AI drafts the contract, prepares the return, designs the campaign, writes the code, someone will need to stamp it. The stamp will not be a rubber stamp. It will be a face.

What it means to be holdable

To be holdable is to be the kind of person who can be found when things go wrong and who has something to lose when they do. It is a combination of three things: identity, location and stake.

• Identity: a name and a face people recognize, that you cannot cheaply replace.
• Location: a place where you can be found, today and next year.
• Stake: something you would lose if you broke your word, money or reputation or license or standing in a community.

Machines will be able to do more and more of the work. They will not be able to be holdable. That job is reserved for people, and in a world full of machine work, it is the job that will be scarce.

Chapter 6

Reading the Room

Within hours of being born, a baby will turn its head to follow a pattern that looks like a face, more than a pattern with the same parts scrambled. Researchers found this in the early 1990s, and it has held up. Before we know anything else, we are looking for faces.

That makes sense. For our whole history, faces were how we found out whether to trust someone. We read intention, fear, warmth, deception and embarrassment in the small movements around the eyes and mouth. We are built for this. Much of it happens below the level of conscious thought.

The richest channel there is

A conversation in a room is a strange and wonderful thing when you look at it closely. Two people are not taking turns sending messages. They are adjusting to each other continuously. One starts to frown, and the other softens a sentence before it is finished. One leans in, and the other knows to go on. Misunderstandings are caught and repaired in a second, often without either person noticing.

Every narrower channel loses some of this. A video call drops the timing and most of the body. A phone call drops the face. Text drops the voice. Each step down removes information, which is why hard conversations go better in person and why so many relationships fray over email.

For a long time we accepted those losses because the narrower channels were cheap and convenient. That trade still makes sense for most small matters. But there is a second loss now, and it is larger. The narrow channels have become fakeable. The face on the screen may not be a face. The voice may not be a voice.

The room is the only channel that is both the richest and the hardest to fake.

Face as something you can lose

The sociologist Erving Goffman wrote an essay in 1955 called On Face-Work. He used the word face in the old sense we still keep in the phrase losing face. Your face, in that sense, is the public self you claim in front of others. You can keep it, you can save it, and you can lose it.

This gives the word a second meaning that matters here. A face is not only what lets people recognize you. It is something you can lose in front of them. When you sit across a table from someone and make a promise, you are putting your face on the table. If you break it, you lose face with that person, and with everyone they talk to. That possible loss is part of what makes the promise worth something.

A generated face cannot lose face. It has no public self to protect, no history with the people watching, no tomorrow in which it must look them in the eye. It is all expression and no exposure.

What the room does that nothing else does

Being physically present does a few things that will matter more and more:

• It proves you are real. Not just that a face exists, but that this specific person spent this specific hour here.
• It exposes you. You can be read, questioned, pressed and seen to hesitate.
• It binds you. You have put your face on the table in front of a witness who will remember.
• It honors the other person. You spent something you cannot get back on them, and they know it.

The last one is easy to overlook. Showing up is a gift. It says you matter enough that I gave you a piece of my life that I could have given to anyone else. In an age when everything else can be sent, the visit becomes the most personal thing one human can give another.

A skill worth keeping

There is a risk in all of this. If people spend their days talking to screens and machines, they may lose some of their skill at reading rooms. That would be a strange loss at exactly the moment the room becomes most valuable.

The people who will do best in the coming decade are those who can sit across from someone and read them well, and who can be read honestly in return. That is not a soft skill. It is the skill on which trust, and therefore business, increasingly depends.

Chapter 7

The Law of the Scarce Complement

Some goods come in pairs. Cars and gasoline. Printers and ink. Coffee and cups. Economists call these complements: things that are used together, so that the value of one depends on the other.

Complements obey a simple and powerful rule. When one side of the pair becomes cheap and plentiful, the value of the pair moves to the other side. If gasoline were free, people would pay more for cars. If cars were free, people would pay more for gas, parking and roads. The money goes to whatever is still scarce.

Clayton Christensen called a version of this the law of conservation of attractive profits. When one layer of an industry becomes a commodity, the profits move to an adjacent layer that is not. The software writer Joel Spolsky put the business version more bluntly in 2002: smart companies try to make their complements cheap, because it makes their own product more valuable.

Work and the person who answers for it

Now look at the pair that AI is pulling apart. For all of economic history, work and responsibility came bundled. The person who did the work was also, mostly, the person who answered for it. The accountant prepared your return and signed it. The lawyer drafted your contract and stood behind it. The designer made your logo and fixed it if you were unhappy.

AI splits that bundle. It can do a growing share of the work: the drafting, the calculating, the designing, the coding, the researching. It cannot do the answering. It cannot be held, as we saw. So the work is becoming abundant while the answering stays scarce.

Apply the rule. When the work becomes cheap, the value moves to its complement. The complement of AI work is an accountable human who stands behind it.

When the work becomes cheap, the value moves to the person who answers for it.

Figure 2

AI unbundles the work from the answering

The work

• Drafting, calculating, designing
• Coding, researching, scheduling

Becoming abundant. AI does more of it every month.

Value moves

The answering

• Taking the call when it fails
• Carrying the liability and the judgment

Staying scarce. It needs a face, a place and a stake.

Why this surprises people

Most predictions about AI and jobs focus on the work. Which tasks can a machine do? Which jobs will vanish? Those are fair questions, but they look at only one side of the pair.

The other side is where the surprise is. As the cost of the work falls toward nothing, the total amount of work done will rise enormously. Every business will produce more contracts, more designs, more code, more analysis, more messages. And every one of those outputs will need someone to answer for it, to the extent it matters. The demand for answering grows with the volume of work.

So the scarce job is not the doing. It is the standing behind. And standing behind requires exactly what a machine lacks: a face, a place and a stake.

The price moves, it does not vanish

People sometimes say AI will make professional services nearly free. Parts of them will be. The labor inside a service will get very cheap. But the price of a service was never only the labor. It was also the promise, and the person making it.

I expect the bill to change shape. The line for hours spent doing the work shrinks. The line for responsibility, judgment and presence grows. Clients will pay less for someone to write the document and more for someone to look them in the eye and say, this is right, and if it is not, I will fix it.

That is a better deal for the client and a better job for the professional. It removes the drudgery and keeps the part of the work that was always the point.

The complement has a face

Here is the conclusion of the argument so far. Trust lives in what is costly to fake. The costliest thing to fake is a person in a room. Trust needs someone you will meet again and someone you can hold. And when work becomes cheap, value moves to whoever answers for it.

All four lines point at the same place. The scarce complement to abundant machine work is a human face, attached to a findable person with something to lose. In the next two chapters I want to show what that looks like in practice.

Chapter 8

The Face of Record

Every ship has a master. Every building permit names a responsible party. Every patient has an attending physician. Every company files with a registered agent. Each is a way of answering a question that complex systems always eventually face: when something goes wrong, whose door do we knock on?

I want to give the answer a name for the AI era. Call it the face of record. The face of record is the named human who carries trust, liability and responsibility for a piece of work, whoever or whatever actually did it.

Every digital product will have one

My prediction is that every digital product and service will come to have a human face of record. Not a mascot. Not a support line. A specific person whose name is on it, who can be reached, and who loses something if it fails.

That sounds backward. For decades, the direction of software was to remove people from the loop. The ideal product needed no humans to sell it, support it or stand behind it. It just worked, at scale, everywhere at once.

That ideal made sense when the hard part was the work. If software could do the work, removing people saved money. But once AI makes the work cheap for everyone, a product with no face behind it is no longer special. Anyone can make one. What is rare is a product with a person who will answer for it.

A product with no face behind it is something anyone can make. A product with a face behind it is something people will trust.

What the face of record does

The face of record is not the person who does all the work. Most of the work may be done by AI. The face of record does four things a machine cannot:

• Takes the call when it goes wrong, as a person, not a script.
• Carries the liability, through insurance, contract, license or reputation.
• Makes the judgment calls that cannot be reduced to rules, and owns them.
• Shows up, in person when it matters, so the customer knows who they are dealing with.

Notice that none of these require the face of record to be an expert in every task. The architect does not draw every line. The master does not trim every sail. What they bring is accountability, and the judgment to know when something is wrong.

The small firm with a big fleet

I think this produces a new shape of business. A small number of trusted human faces at the front, with a large fleet of AI delegates behind them doing the work. The faces carry the relationships, the liability and the judgment. The fleet carries the volume.

This is roughly how my own business, Charleston AI, is built. We deliver finished AI work for local businesses, and a person stands behind each delivery. The AI does much of the producing. The relationship, the responsibility and the face are human. That is not a compromise between old and new. It is the arrangement the logic points to.

The same shape will appear in law, accounting, design, medicine, real estate, insurance and trades. One trusted face, many delegates. The face is the scarce part. The delegates are abundant. And because the face is scarce, it is where the value collects.

The face of record in the contract

I also expect the face of record to show up in contracts. Today a software agreement names a company. In the coming years I think buyers will increasingly insist on naming a person: who, specifically, answers for this system in our business? Who do we call, and who comes to our office if it fails?

Companies that cannot answer that question will be at a disadvantage. Companies that can will charge for it. The face of record will become a line item, then a selling point, then a standard expectation, the way a warranty did.

I wrote about a version of this in The Warranty Layer: software earns its place by the promise it can keep. The face of record is the person who keeps the promise.

Chapter 9

Zoom Is the New Email

There was a time when a phone call was a serious thing. People wrote letters for ordinary matters and called for urgent ones. Then calls became cheap, and the serious thing became the visit. Then email arrived, and the phone call became the serious thing again, for a while. Every time a new channel gets cheap, the old ones shuffle in rank.

AI is about to shuffle them all at once, and in the same direction. Email, text, phone and video are all becoming fakeable together. That puts them all on the same side of a new line.

The new ranking

Here is how I think people will come to rank channels, from least to most trusted, within a few years:

• Anything written: email, text, chat, documents. Assumed to be possibly generated.
• Anything voiced: calls and voice messages. Assumed to be possibly synthesized.
• Anything on a screen: video calls and recordings. Assumed to be possibly rendered.
• A person in the room. Assumed to be real.

The first three will collapse into one tier. Today we still feel a difference between an email and a video call. That difference is fading. Within a few years, a video call will carry roughly the trust an email carries now: fine for routine work, not enough for anything important.

Within a few years a video call will carry roughly the trust an email carries now.

What this does to business

This changes the economics of distance. For twenty years, remote work and remote selling got steadily cheaper and better. Distance became nearly free. A consultant in one city could serve clients anywhere through a screen.

For routine work, that will stay true and get truer, as AI handles more of the routine. But for trust-heavy work, distance is about to become a discount. If the only way I know you is through a screen, I will trust you less than someone I can meet, and I will pay you less, or not hire you at all.

So I expect a split. Remote becomes the cheap tier: efficient, scalable, and limited to matters where being fooled costs little. In-person becomes the premium tier: slower, scarcer, and reserved for matters that count. The same professional may offer both, and price them very differently.

Rituals come back

As the screen loses trust, people will rebuild ways to verify each other, and many will look old-fashioned. The handshake deal. The signing in a room. The first meeting over a meal. The office visit before a contract. The return to a family home for a big decision.

None of this is nostalgia. Each ritual is a costly signal. It cannot be faked from a distance, and it costs time that cannot be refunded. People will reach for these rituals for the same reason you called your daughter back on her known number: when the signal can be faked, go around it to something that cannot.

Video is not the enemy

I want to be careful here. Video calls are wonderful tools, and they will remain useful for the vast majority of meetings where trust is already established. Once I know you, a video call with you is fine. It is a continuation of a relationship that started in a room.

The issue is beginnings and big moments. The first meeting. The large payment. The hard conversation. The commitment. Those will move back into the room, because that is where the trust is born. The screen will carry the relationship. The room will create it.

Chapter 10

Be the Face Your Customers Trust

I teach students who are around twenty-four years old, just starting their careers. In First Customer, I wrote down the career strategy I give them: let your first employer be your first customer. Keep your needs low, sell your work, and build a relationship rather than a job.

There is a second piece of advice I give them, and this book is really an argument for it. Be the face that your customers trust.

What a customer is

Start with what a customer actually is. A customer is not a transaction or an account number. A customer is a human relationship whose basis is wealth. Money moves between you, but the relationship is human, and like every human relationship it runs on trust.

Most of that trust is implicit. Nobody writes it down. The customer simply expects that you will do what you said, at the quality you implied, and fix it if it goes wrong. Sometimes the trust becomes explicit, in an agreement or a contract. But even the contract is only as good as the trust behind it.

So when I tell students to be the face their customers trust, I am telling them to become the person on whom that implicit trust rests. Not the brand. Not the website. Not the AI that answers the questions. The person.

A customer is a human relationship whose basis is wealth. Like every human relationship, it runs on trust.

How you become that face

There is nothing mysterious about it. It is built from ordinary acts repeated in the same place over time.

• Show up in person, especially at the beginning and at the big moments.
• Be in the same place, so people know where to find you.
• Say your name and put it on the work.
• Answer for mistakes yourself, face to face, without blaming the tools.
• Let your AI do the work, and never let it stand in for you.

The last one matters most for young people. Your generation will be the first to have a capable AI from day one. It will be tempting to let it handle everything, including the customers. Do not. Let it handle the work. Keep the face for yourself. The work is abundant. Your face is not.

Why this is good news for the young

There is an old complaint that young people cannot compete with experienced ones because they lack skill. AI changes that. A capable young person with good AI can produce work at a level that used to take years to reach.

What AI cannot give them is trust, because trust is built from presence over time. But here the young have an advantage too. They have time. If they start showing up now, in one place, for the same customers, and answering for their work in person, they will have something in ten years that no machine and no newcomer can copy: a long shadow of the future, attached to their own face.

Trust compounds

Trust behaves like interest. Each kept promise adds a little, and each addition makes the next one easier. A customer who trusted you once is more likely to trust you again, and to tell others. Over years, the effect is enormous.

Fakes cannot compound. A generated identity has no history to build on. A business that hides behind an interface starts every relationship at zero. Only a face that persists in a place can accumulate trust. That is why I tell my students to choose a place, show up, and be the face. It is the one investment AI makes more valuable rather than less.

Chapter 11

Delegate Down, Never Delegate a Person

There is a rule I live by and teach. Delegate the small matters. Keep for yourself only what is worthy of human attention.

I use the word delegate on purpose, rather than agent. To delegate is to send down. The Latin root carries that sense of sending someone on your behalf. Delegation only runs one way: from the one who holds responsibility to the one who carries out the task. That is why I prefer to call the human a manager, not a user. You do not use your AI. You delegate to it.

Rank and order

Delegation depends on rank and order. Some matters are small: scheduling, drafting, sorting, researching, following up, paying the routine bill. These are worthy of a delegate’s attention. Others are large: a decision that changes a life, a conversation that changes a relationship, a person who deserves you.

The mistake people are about to make, by the millions, is to forget the order. Because AI can handle almost anything, they will let it handle almost everything, including the people in their lives.

The one thing you never delegate

Never have your AI deal with a person who is worthy of your own attention. A customer who trusts you. A friend. A family member. A colleague in a hard moment.

When you send your AI to deal with such a person, you are delegating them down. You are telling them, by your action, that they rank below your attention, among the small matters. That dishonors them. It is worse than the old complaint that someone used me. To be used is at least to be engaged by a person. To be delegated is to be handed to something beneath them.

To be used is at least to be engaged by a person. To be delegated is to be handed down.

People will feel this, even when they cannot name it. They will notice when the reply is too fast, too smooth, too general. They will notice when the birthday message has no fingerprints on it. And they will draw the conclusion: I was handled.

The face is reserved for persons

This rule fits the argument of the whole book. If the face is the scarce thing, then spending your face is how you show what matters. Where you show up, in person, is a public statement of rank. It says: this person is worthy of me.

So the right arrangement looks like this. Your AI handles the volume. It drafts, prepares, reminds, researches and runs the errands of your working life. It frees your time. And then you spend that freed time on faces: the customer in the room, the friend at the table, the child in the kitchen.

AI should give you more face time, not less. If it is giving you less, you have the order upside down.

What good delegation looks like

Picture a small business owner a few years from now. Her AI has prepared the contract, the invoice, the briefing on the client’s history and the three options for the hard decision. It has done the work of a small staff.

She drives across town and sits down with the client. She looks at him and says, here is what I recommend, and I will answer for it. The AI did the work. She brought the face. The client does not care how much of the document a machine wrote. He cares that she came, and that she will be there if it goes wrong.

That is delegation done right. The machine went down to the small matters. The human went across to the person.

Chapter 12

The Room

In January 2026 I opened a three-thousand-square-foot space in Charleston. Its purpose is simple. Local people and organizations can walk in, sit down and get help understanding and using AI. There is a door, a sign and a room with chairs in it.

People sometimes ask why an AI business needs a physical room. The work is digital. The tools are online. Why pay rent?

This whole book is my answer. The room is the costly signal. A lease is a commitment that cannot be faked from a distance. A sign on a building says I will be here tomorrow. A door you can walk through says come and find me. In a world where every website can be generated in an afternoon, a room you can walk into is a promise.

A storefront is a signal

Think about what a physical place tells you. It tells you someone committed money to being here for years. It tells you there is an address where you can show up angry. It tells you the people inside have neighbors who know them. It tells you the business has a stake in this town.

None of that can be generated. A website can claim to be local. A room is local. A chatbot can claim to care about you. A person behind a counter, in the same place every week, has shown you they are invested.

Every website can be generated in an afternoon. A room you can walk into is a promise.

Why local comes back

For a generation, local businesses were told they had to compete with the whole internet. Often they lost. The online alternative was cheaper, faster and just as good, because the work was the point and the work could be done from anywhere.

AI makes the work cheap for everyone, near and far. That removes the advantage of distance. What is left is the advantage of nearness: being able to sit across from someone, be found, be held, and build a long shadow of the future in one place.

I expect hyper-local to become a strength again in exactly the industries the internet was supposed to flatten. Advice, design, finance, insurance, health, technology help, education. Anywhere the customer needs to trust the person behind the work, the person who can be visited will have an edge over the one who cannot.

The most powerful sentence in business

In the coming years I think one sentence will carry more weight than any ad, review or credential. It is this: come by.

Come by the office. Come by the shop. Come by and we will sit down. It is an invitation that only a real, findable, accountable person can make. A fake cannot say it, because a fake has nowhere for you to come. A distant firm cannot say it, because the trip is too far. Only the face in the room can say it and mean it.

When a business says come by, it is putting its face on the table. It is saying, I am real, I am here, and I will answer for this. That is the strongest trust signal left.

The room is for people

The room is not where the AI lives. The AI lives everywhere. The room is where people meet people. It is where a nervous business owner can watch someone work and ask the questions they would be embarrassed to type. It is where trust starts, and from there the relationship can continue on screens and through delegates.

I opened a room because I believe the future of AI is personal, and the personal begins face to face.

Chapter 13

Rituals of Verification

Every family will need a rule. Some already have one. It goes something like this: if anyone calls asking for money, no matter how real they sound, we hang up and call back on the number we know. If it is big, we see each other.

That rule is a ritual of verification. It is a way of going around a fakeable signal to one that is harder to fake. Families will adopt these rules the way they once adopted smoke detectors: slowly, then all at once, usually after someone they know gets burned.

Rituals at home

The family versions will be simple, and many will involve faces.

• A known number. Calls are returned on a line you have had for years, never on the one that called.
• A shared word or a shared memory. Something no outsider would know, asked in person or on the known line.
• A presence rule. Above a certain amount, no money moves until you have seen each other.
• A slow rule. Anything urgent and secret is treated as suspicious because of its urgency and secrecy.

What these rules have in common is that they spend something: time, effort or travel. They make verification costly on purpose, because cost is what makes signals honest. The family that has to drive across town before wiring money is a family that cannot be robbed by a voice.

The family that has to drive across town before wiring money is a family that cannot be robbed by a voice.

Rituals at work

Businesses will build the same rituals, at larger scale. In 2024, an employee at an engineering firm in Hong Kong transferred company money after a video call in which the senior colleagues he saw and heard were generated. Every face on the screen looked right. None of them were there.

Firms are learning from cases like that, and the lesson is not to buy better detectors. Detectors will always lag behind generators. The lesson is to change the rule. Large payments, changes to bank details, new vendors, executive instructions: these will increasingly require a human in a room, or a callback through a channel established in a room.

I expect the in-person approval to return as a formal step in companies of every size. Not for everything. For the moments where being fooled is expensive.

The return of ceremony

At a larger scale still, I expect a quiet return of ceremony in public life. Closings in person. Signings in person. Oaths in person. Votes, testimony and commitments that require a body in a place.

People will call this old-fashioned, and it is. It is old-fashioned for the same reason the wax seal was old-fashioned when the telegraph arrived. The old forms persisted where the new ones could be faked. Ceremony is simply the human practice of making an important moment costly, so that everyone present knows it was real.

Verification is a relationship

The deepest point about these rituals is that they are not technical. They are relational. A known number works because you have known it for years. A shared memory works because you lived it together. A presence rule works because you can find each other.

Verification, in the end, rests on relationships that existed before the moment of doubt. That means the best defense against a world of fakes is not a tool. It is a set of real relationships with real people whose faces you know, built before you need them.

So build them now. Know your neighbors. Know your banker by face. Meet your vendors. Have your clients to the office. The relationships you build in rooms today are the verification system you will rely on tomorrow.

Chapter 14

The Price of Presence

If the face is the scarce thing, it will be priced. That is what markets do with scarce things. The question is how.

Start with the basic constraint. A person can be in one room at a time. A day has a fixed number of hours. Travel takes time that cannot be compressed. So the supply of any one person’s presence is strictly limited, no matter how capable their AI becomes. AI can multiply your output. It cannot multiply you.

Two prices for the same professional

I expect many professionals to end up with two prices. One for work delivered at a distance: prepared largely by AI, reviewed by them, delivered through screens. This will get cheaper and cheaper, because the work inside it is getting cheaper.

The other for presence: they come, they sit with you, they look at your situation with their own eyes, and they put their face on the outcome. This will hold its price or rise, because it draws on the one input that does not scale.

We already see the first version of this in medicine, law and consulting, where a visit costs more than a message. AI will widen that gap. The message gets nearly free. The visit stays a visit.

AI can multiply your output. It cannot multiply you.

The scarce resource inside a firm

Inside companies, the same logic will change what counts as the scarce resource. For most of business history, the scarce resource was labor: enough skilled people to do the work. AI is loosening that constraint quickly.

The new constraint will be trusted faces: people whom customers will accept as the face of record, who can be in the room, who carry relationships and liability. A firm’s capacity will be limited not by how much work it can produce but by how many relationships its faces can hold.

That changes hiring. A firm will want fewer people who only do work, because AI does that. It will want more people who can be trusted in rooms. It changes compensation too. The person who can carry a client relationship in person will be paid more relative to the person who only produces output, because the first is scarce and the second is not.

Spending presence wisely

For each of us, presence becomes a budget. You have only so many hours of face to spend. Spending them well becomes a central skill, at work and at home.

The answer is the rank and order of the last part of this book. Spend presence on people worthy of your attention. Spend it on beginnings, where trust is born, and on big moments, where trust is tested. Let screens and delegates carry everything in between.

A good life in the coming decade may look like this: AI handles the volume of your working life, and you spend your freed hours in rooms with people. More presence, not less. Fewer messages, more meetings that matter. That is not a retreat from technology. It is what technology makes possible when you hold the order right.

A note on fairness

There is a worry here, and I want to name it. If presence becomes premium, will only the wealthy get it? Will everyone else be stuck with machines?

I think the opposite is more likely, for a simple reason. AI makes the work so cheap that a single trusted professional can serve far more people than before. The doctor, the lawyer, the adviser, freed from paperwork, can spend more of their hours face to face. The total supply of human presence available to ordinary people may actually rise, because so much of it was being wasted on work a machine can now do.

That outcome is not guaranteed. It depends on choices. But the possibility is real, and it is worth working toward.

Epilogue

Come By

She comes by. She sits at your kitchen table, and you see her face, and you know it is her before she says a word. Whatever she needs, you will help, because she is here, and because she drove across town to ask you, and because you can see in her eyes that it matters.

That is the world I expect. Not one where we trust nothing, but one where we have learned again where trust lives. The machines will do the work. They will draft, calculate, design, schedule and run the small matters of our lives better than we ever could. They will free enormous amounts of time.

And we will spend that time on faces. We will go to the office to meet the person who answers for our money. We will sit with the adviser before the big decision. We will walk into the local shop and ask the person behind the counter. We will drive to see the people we love.

The last signature

For most of history the signature was how you put yourself on a promise. It was a mark of your hand, hard to copy, attached to you. Machines can copy any hand now. The signature has become a picture of a signature.

What is left is the face. Not the image of it. The face itself, in a room, attached to a person who can be found tomorrow and who has something to lose. It costs a body, a place and an hour you cannot get back. That cost is not a bug. It is the guarantee.

The face is the last signature.

A copy costs nothing. That is why it is worth nothing as proof. A face costs everything a person has: their time, their presence, their name, their standing among the people who know them. That is why it is worth so much.

Be the face. Choose a place. Show up. Let your AI do the work, and keep the face for yourself.

And when it matters, say the two words only a real person can say.

Come by.