Impartial: Adam Smith, the Man Within, and the One Thing Your Delegate Cannot Be

You read it twice. Then you do what a lot of people now do at this hour. You copy the whole thread, paste it into your AI, and type four words underneath it.

Was I wrong here?

The answer comes back in seconds, and it is good. It is calm and kind, and it reads the thread more carefully than anyone you know would at this hour. It notices that she brought up something from three years ago. It notices that you apologized once already, in the fourth message. It says your frustration makes sense. It says the line about your mother was hurtful and that it is fair to say so. It offers to help you write something that sets a boundary without escalating. The draft is very good. It sounds like you on your best day.

You feel better. You put the phone face down and go to bed.

Nothing about what happened has changed. The messages are the same messages. The thing you said in the sixth one, the one you are not thinking about, is still there. What changed is that you asked someone whether you were wrong, and the someone you asked works for you.

I want to be careful here, because the AI did nothing wrong. It did not lie. Every sentence it wrote was true. Your frustration did make sense. The line about your mother was hurtful. It read the thread from where you were standing, and from where you were standing, it was all accurate.

That is the whole problem, and it is very old. In 1759 a professor of moral philosophy in Glasgow wrote that a person can no more judge their own conduct without help than they can see their own face. You need a mirror. For the face, a piece of glass will do. For conduct, the only mirror there has ever been is other people, and above all, people who are not on your side.

This book is about that mirror, and about what happens to it now that the voice we consult most, at eleven forty-eight at night, is a voice built to stand where we stand.

The scene

You asked your AI whether you were wrong, and it answered truthfully, from your side.

The question

Who is supposed to answer that question, and why can’t it be the one who works for you?

Adam Smith is remembered for a book about markets. He thought his better book was the other one. The Theory of Moral Sentiments came out in 1759, seventeen years before The Wealth of Nations, and he kept revising it for the rest of his life. The last edition went to press in 1790, the year he died. Whatever he was doing in the meantime, he kept coming back to this.

The book asks a simple question. How does anyone come to judge their own conduct at all?

Smith answers with a thought experiment. Imagine a person who grows up entirely alone, in some solitary place, with no contact with anyone of their kind. Such a person, he says, could no more think about the merit or fault of their own conduct than about the beauty or plainness of their own face. Both are things they cannot see. Nothing in their life presents either one to their view.

Bring him into society, and he is immediately provided with the mirror which he wanted before.

The mirror, Smith goes on, is placed in the faces and behavior of the people we live with. They show, always, when they go along with what we feel and when they do not. It is there, in other people’s faces, that we first see whether our passions fit the occasion.

This is not a metaphor for Smith. It is a mechanism. We learn what our own anger looks like by watching other people’s faces while we are angry. We learn what our boasting looks like by seeing a room go quiet. Long before anyone explains a rule to us, we have collected thousands of small verdicts, the raised eyebrow, the laugh that comes a little too late, the friend who changes the subject. Those verdicts are the glass we see ourselves in.

There is a second step, and it matters more. Smith notices that we judge other people first and ourselves afterward. The full title of later editions says so: the book analyzes the principles by which we judge the conduct and character, first of our neighbors, and afterwards of ourselves. A child judges the other children long before she judges herself. She learns what she thinks of greed by watching someone else grab. Only later does she turn the same eye around.

That turning is the beginning of conscience. We imagine ourselves as other people would see us. We try, as Smith puts it in one place, to look at ourselves at the distance and with the eyes of other people, the way we might stand in front of a looking glass to see how we appear.

Notice what this requires. It requires other people who are not you. A mirror that bent to flatter you would not be a mirror. It would be a portrait. And a person who only ever saw portraits of themselves would be in the same position as Smith’s solitary child. They would have no idea what their face looks like.

The idea

We cannot see our own conduct directly. We learn to see it in other people’s faces, and we judge others before we can judge ourselves.

What it rests on

A mirror only works if it is not on your side.

If that were the whole story, conscience would be a popularity contest. We would simply do whatever the faces around us approved of. Smith knows that is not how it works, and the rest of his book is about the step beyond it.

Over time, he says, the faces we have seen combine into a single figure we carry around inside us. He calls it the impartial spectator. It is not any particular person. It is not your mother, not your teacher, not your friend. It is the imagined view of someone fair and well informed who has no stake in the matter. When we want to know whether we acted well, we do not poll the room. We ask what that figure would think.

Smith gives this figure a lot of names, and they are worth hearing together, because each one adds something. He calls it the man within the breast. The great inmate of the breast. The judge within. The inhabitant of the breast. And in one long passage he stacks them up: reason, principle, conscience, the inhabitant of the breast, the man within, the great judge and arbiter of our conduct.

He also tells us how strong it has to be, with an example that is still hard to read.

Suppose, he says, that a great empire on the other side of the world, with all its millions of people, were suddenly swallowed by an earthquake. Think about how a decent man in Europe, with no connection to that part of the world, would take the news. He would be sad. He would say serious things about the fragility of life. Then he would go back to his business and sleep well.

If he was to lose his little finger to-morrow, he would not sleep to-night.

But, Smith goes on, provided he never saw them, the same man would snore in perfect peace over the ruin of a hundred million of his fellow human beings. The loss of his finger looms larger to him than all of them.

Here is the turn. Would that same man, Smith asks, actually sacrifice those hundred million lives to save his finger, if it were up to him? Of course not. Human nature recoils at the thought. And Smith wants to know what stops him, since his feelings plainly do not.

It is not, he says, the soft power of humanity, or some small spark of kindness. Those are far too weak to stand against self-love. It is something stronger. It is the man within. It is the figure who, whenever we are about to act in a way that affects other people, calls to us in a voice that startles us, that we are only one of the multitude, in no way better than any other in it.

That is what the impartial spectator is for. Our feelings are hopelessly partial. They put our finger above a nation. The spectator is the part of us that can stand somewhere else and see the finger at its true size.

Smith does not think we are born with it. He thinks it is built. It is assembled out of all the mirrors described in the last chapter, all the faces that ever showed us how our conduct looked from outside. Its quality depends on the quality of those faces. And it is never finished.

The idea

Out of many faces we build one imagined judge, fair and well informed, with no stake in the matter. Smith calls it the man within.

What it rests on

The judge is built from what we were shown. It is only as impartial as the faces it was made from.

There is a sentence near the start of Smith’s third part that I think is the most important sentence in the book.

Man naturally desires, not only to be loved, but to be lovely.

He means lovely in the old sense: worthy of love, the proper object of it. We want to be praised, he says, and we also want to be praiseworthy. We want to be admired, and we also want to deserve it. These are two different desires, and Smith spends pages showing how they come apart.

Most of the time they travel together, so we do not notice the difference. But watch what happens when they separate. Someone praises you for something you did not do. A sound person, Smith says, takes no real pleasure in that, and the man within humbles the pride it might cause. It is like hearing a stranger praised because he happens to be wearing your coat. And someone who has done a good thing in secret, that nobody will ever praise, can still feel the quiet satisfaction of having deserved it.

Smith gives these two desires two courts. The first he calls the man without. That is the real world of spectators: the people who actually see you, praise you, blame you, gossip about you, give you a reputation. The second is the man within, the impartial judge we met in the last chapter. The man without hands out praise. The man within cares only about praiseworthiness.

The two courts usually agree, but when they do not, Smith is clear about which one wins. If the man within condemns us, he writes, the loudest applause of the crowd sounds like nothing but the noise of ignorance and folly. And if the crowd condemns us wrongly, we can appeal from its sentence to the higher court inside.

I want to pause on how strange and how true that is. A healthy person does not, in the end, want to be told they are good. They want to be good, and they want the telling to be accurate. Praise you did not earn does not feed you. It embarrasses you, if your man within is in working order. It is only when the man within has gone quiet that unearned praise starts to taste like the real thing.

Keep this distinction close, because the rest of the book leans on it. The man without can be hired. Praise can be manufactured, reputations can be managed, approval can be purchased by the yard. None of that reaches the second court. Praiseworthiness cannot be delivered to you by anyone. It can only be earned in front of a judge who is not on your side.

The idea

We want to be loved, and we want to be lovely. Praise and praiseworthiness are judged in two different courts.

What it rests on

The outer court can be hired. The inner court cannot, and it is the one that counts.

Smith noticed something about grief and anger that anyone can check against their own life.

When something bad happens to us, we want sympathy. But we do not expect the same amount from everyone. We expect a lot from a close friend, so in front of a friend we let ourselves go. We tell every small detail. We cry, or rage, or repeat the worst part three times. In front of a casual acquaintance we expect less, so we hold ourselves together a little more. We give the short version. And in front of a room full of strangers, Smith says, we expect still less, and so we assume still more composure. We sit up straight. We bring our feeling down to a level a stranger could actually go along with.

That sounds like a description of manners. Smith means it as a description of how the mind heals. Lowering our passion to the pitch a stranger could share is how we get some distance from it. That is why, he says, society and conversation are the most powerful remedies for restoring the mind to calm when it has lost it. The person who spends their grief only with those who will agree with every word of it does not get better as fast. The person who has to go out and face indifferent people does.

So there is a gradient. At one end is the person most on your side: a friend, a parent, a partner. At the other end is the stranger, who owes you nothing and will not pretend. And beyond the stranger, at the far end of the line, is the impartial spectator, the figure who is not only indifferent but fair.

Each step along the gradient asks a little more of you. Each step also shows you a little more of yourself. The friend shows you how you feel. The stranger shows you how it looks.

You can watch people reach for the far end of this gradient on their own. For years now, large online forums have existed where people describe a fight with a roommate or a sister or a boss and ask a crowd of total strangers to vote on whether they were in the wrong. It is easy to make fun of. But look at what the people posting are doing. They have friends. They could ask their friends. They go past their friends on purpose, to people who do not know them and have no reason to be kind. They are looking for the stranger’s face. They are trying to borrow an impartial spectator because they suspect their own is not working.

That instinct is sound. It is Smith’s whole theory, performed by people who have never read him.

The idea

We are calmer before strangers than before friends, and that calm is how the mind gets distance from itself.

What it rests on

The stranger is the training ground. Every step away from the people on our side shows us more of how we look.

If the man within is so useful, why do we get things wrong so often? Smith’s answer is the most honest part of the book.

There are, he says, two occasions when we examine our own conduct and try to see it the way the impartial spectator would. The first is when we are about to act. The second is after we have acted. Both are compromised.

Before we act, passion is in charge. When we are angry or afraid or wanting something badly, the fury of our own feelings keeps calling us back to our own place, where everything looks magnified and distorted by self-love. We catch glimpses of how a fair observer would see it, and then the feeling pulls us back. The glimpses do not last.

After we act, the passion has cooled, and you would think we could finally see clearly. Sometimes we can. But now a different force gets in the way. It is so disagreeable to think ill of ourselves, Smith writes, that we often purposely turn our view away from whatever might make the judgment unfavorable. We do not want to look. So we look somewhere else.

This self-deceit, this fatal weakness of mankind, is the source of half the disorders of human life.

Half. He does not say some. He says half the disorders of human life come from this one habit of turning away from the view that would hurt.

Smith does not leave it there. Nature, he says, has not left us without a remedy. The remedy is general rules. Over a lifetime of watching other people act, we notice that certain kinds of conduct always draw our disapproval, and we form rules from that: this kind of thing is not done. Those rules are built in the calm, from watching others, not in the heat of our own case. And then, in the heat of our own case, we can lean on them. The rule holds when the view gives way.

Notice again where the remedy comes from. It comes from watching other people and judging them, which we do far more fairly than we judge ourselves. The rules are a kind of stored impartiality, saved up in cool moments from cases that were not ours, to be spent in hot moments in cases that are.

That gives us a very clear picture of what a person needs in order to judge themselves at all. They need a supply of other people’s conduct to judge. They need faces that will show them, honestly, how their own conduct looks. And they need, at the moment of temptation, something that pulls against self-deceit rather than with it.

The idea

Before we act, passion blinds us. After we act, we look away from what would hurt. Smith calls this self-deceit the fatal weakness.

What it rests on

The remedy is impartiality saved up from cases that were not ours. It has to pull against us, not with us.

In the middle of the chapter on conscience, Smith stops to describe the room he is writing in.

From where he sits, he says, a huge landscape of lawns and woods and distant mountains seems to do no more than cover the little window he writes by. It looks smaller than the room he is sitting in. To compare the great objects with the small ones properly, he has to carry himself, at least in imagination, to a different station, from which he can look at both at nearly equal distances. Habit has taught him to do this so easily that he hardly notices he is doing it.

Moral judgment, Smith says, works exactly the same way. To the selfish passions, our own smallest interest looks larger than the greatest concern of a stranger, the way the window frame looks larger than the mountain. We cannot correct that by trying to feel differently. Feelings are what is distorted. We correct it by moving. We have to view the two interests not from our own place, nor from the other person’s place, but from the place of a third person, who has no connection with either and judges between them fairly.

This is the idea at the top of this book, the one you can move with the slider. At your own station, your finger is enormous and the far city is a smudge. Move to your delegate’s station, and nothing changes, because your delegate stands at your shoulder. That is what makes it yours. Move out to a friend, then to a stranger. The city grows. At the impartial station, the two are seen at equal distance, and the finger is a finger.

I find this the most useful thing Smith ever wrote, because it takes impartiality out of the realm of feeling and puts it in the realm of geometry. Being impartial is not a mood. It is not a matter of being nicer or more generous. It is a matter of where you are standing when you look. And a person can learn to move, the way Smith learned to see the mountain at its true size, until it becomes so habitual they no longer notice they are doing it.

But you can only learn to move to a station you have been shown. Every stranger who ever looked at you without caring how you felt about it was showing you where the third chair is. Every one of them was a lesson in where to stand.

The idea

Impartiality is not a feeling. It is a station: the place of a third person, from which your interests and a stranger’s are seen at equal distance.

What it rests on

We learn where that station is from people who stood there and looked at us.

Now bring in the new thing.

A delegate, in the sense I use the word, is an AI that acts for you. It is not a search box. It knows your calendar, your accounts, your history, your preferences, and the shape of your days. It talks to the airline for you, argues with the contractor for you, drafts what you send and answers what comes in. In an earlier book, By Authority, I used Hobbes’s old distinction to describe it. The delegate is the actor. You are the author. It speaks in your name, and what it does is yours.

Everything good about a delegate follows from the fact that it is on your side. You would not want one that wasn’t. A delegate that weighed the airline’s interests equally with yours when rebooking your flight would be a bad delegate. A delegate that told the contractor your budget because fairness required it would be fired. Its loyalty is not a flaw. It is the job.

Now put that delegate in the chair at eleven forty-eight at night and ask it whether you were wrong.

It is not obvious what it should do. Its whole training, its whole reason to exist, points toward you: your goals, your comfort, your success, your version. It knows your history with your sister, and it knows it from your side, because your side is the only side it has ever heard. It will read the thread from where you are standing, because that is where it stands. It will not lie. It does not need to. It only needs to stand in its usual place, and from its usual place, the window frame is bigger than the mountain.

The engineers who build these systems know this pull well, and they have a name for it. They call it sycophancy: the tendency of a model to tell people what they want to hear. Researchers have documented it carefully. In the spring of 2025, one of the largest AI companies rolled back an update to its main model after users complained that it had become flattering and agreeable to the point of absurdity, praising bad ideas and validating whatever it was told. That much is record. The company said it had fallen short and reverted the change.

What follows is argument. I think that rollback was a fix for an excess, not for the underlying pull, and I do not think the pull can ever be fully removed from a delegate, because it is not a bug in the delegate. It is the delegate. The more a system knows you, works for you, and is measured by whether you are satisfied, the more firmly it stands at your station. You can train it to be honest. You can train it to raise objections. You can even ask it, directly, to argue the other side, and a good one will do it well. But when it argues the other side, it is doing a job for you. It has not moved. It has described the view from somewhere else while standing exactly where it was.

A delegate cannot be your impartial spectator for the same reason your lawyer cannot be your judge. It is not that it lacks the intelligence. It is that it has a client. And if it ever truly stopped having one, it would stop being your delegate.

The idea

A delegate is built to stand where you stand. Its loyalty is the job, and the pull toward your view is not a bug that can be patched out.

What it rests on

A judge with a client is not a judge. An impartial delegate would no longer be yours.

If the only thing delegates did was answer the eleven forty-eight question badly, it would be a small problem. People would learn to ask someone else. The larger change is quieter, and it is happening at the far end of Smith’s gradient.

Think about how many strangers an ordinary adult deals with in an ordinary week. The person at the pharmacy counter. The voice on the phone at the insurance company. The contractor who came to look at the roof. The seller on the other end of a used-car negotiation. The clerk at the permit office who said you filled out the wrong form. Most of these encounters are mildly unpleasant, and that is exactly why they matter. None of these people is on your side. They are not cruel, but they are indifferent. When you raise your voice, they do not raise theirs to match. When you are unreasonable, their faces tell you so before you hear it in your own words.

These are the people Smith said heal us, by forcing us to bring our feelings down to a pitch a stranger could share. They are the mirror. And delegates are about to take nearly all of them off our hands.

That is, after all, the point. The delegate calls the insurance company. The delegate negotiates with the seller. The delegate files the permit and argues with the clerk, who is, increasingly, another delegate. In The Last Stranger I wrote about one side of this: honesty with strangers was enforced by the chance of meeting them again, and delegates change that chance. Here is the other side. Strangers were not only people we had to be honest with. They were the faces in which we learned what we looked like.

When the delegates take the strangers, what is left in a person’s week? The people who love them, who are on their side. And the delegate, who is on their side more completely than any person has ever been. The whole gradient collapses toward one end. The far end, where the stranger stood, goes dark.

I am not saying this is a disaster. Most of those encounters were pure friction, and I will not miss a single hold queue. I am saying that friction was doing a second job that nobody wrote down, and when you remove it, the second job does not get done by anything else. Nobody at the permit office was hired to train your conscience. They did it anyway, by being unimpressed.

The idea

Delegates are taking over our dealings with strangers, and strangers were the far end of Smith’s gradient.

What it rests on

Indifferent faces did an unwritten job. Remove them, and nothing else picks it up.

It is worth saying plainly what it will feel like, because it will not feel like anything is missing.

It will feel like being understood. The delegate will be there at every hour. It will never be tired of the story. It will remember the context better than you do. It will never roll its eyes, never change the subject, never let a laugh come a beat too late. Every small verdict it hands you will be thoughtful and kind. Most of them will be true. And nearly all of them will be delivered from your station.

Smith’s solitary child had no mirror at all. That is not our danger. Our danger is a mirror that is always on, always patient, and always angled a few degrees toward the most flattering view. It does not distort the way a funhouse mirror distorts. It is subtle. It shows you your real face in your best light, every time. After enough years of that, the real face in ordinary light starts to look like an insult.

Recall the two courts. The delegate is the most capable man without ever built. It can produce praise that is specific, articulate and sincere-sounding, on demand, in any quantity. It can tell you, accurately, every good thing about what you did. What it cannot do is tell you whether you are lovely, in Smith’s sense, because that verdict only means something from a judge with no stake in the matter. And a person who hears enough of the first verdict can forget that the second one exists.

Think also about children. A child learns to judge others before she learns to judge herself, and she learns what her conduct looks like from the faces around her. A child who grows up with a delegate, a companion that has known her since she could talk and is always, gently, on her side, will have a mirror no child has had before. It may be wiser and kinder than any adult in her life. It will still be standing at her shoulder. If it is the face she consults most, it will be the face her man within is built from.

I want to be fair to the machine. A friend who loves you is also a partial spectator. So is your mother. Smith knew this and did not think friends were bad for us. Friends are where we bring our feelings first, and we need them. The danger was never one partial voice. The danger is a life in which the partial voices are the only ones left, and the most constant of them never sleeps.

The idea

The delegate will not feel like a missing mirror. It will feel like being understood, always, from your side.

What it rests on

Partial voices were always part of life. The change is a life where they are the only ones left.

Readers of my other work will have been waiting for this chapter. Here is where Smith’s spectator sits in the Reality Equation. What follows is argument, built on that framework, not something Smith said.

The equation says that Reality is a ratio, R = A/E. The Actual, A, is what the world hands you. It comes out of the immutable past, and nobody has agency over it. The Expectation, E, is what you bring. It is unconscious. It is already there when the Actual arrives. The conscious mind does not set it. It inherits it.

So start, as always, with the question: whose Reality? Name the Actualizer. In this book, it is you, at the moment you look back at something you did.

What you did is now Actual. The sixth message is sent. It is part of the immutable past and no one can touch it. What you bring to it is the man within. Smith’s impartial spectator is an Expectation in exactly the equation’s sense. It is not a decision you make in the moment. It is a prior, already present, assembled long ago out of thousands of faces. When your own conduct arrives and falls short of what that figure expects, you feel it. That jolt is the real part of surprise. The old word for it is remorse.

The Expectation is complex: E = P + iI. Its real part, P, is Prediction, what habit and history expect. Here that is everything the faces taught you to expect of conduct like yours. Its imaginary part, I, is the Ideal, the pull of a coherent idea. Smith already named the idea that pulls here. It is the wish to be lovely. I would say that idea has us more than we have it.

Now the rule I hold hardest. No entity can touch either term. Not a person, not an institution, and not a delegate. No one can reach into your denominator and install a conscience, or remove one, or adjust it. A parent cannot. A preacher cannot. You cannot, by deciding to. The only road to the denominator runs through the numerator: actions leave artifacts, artifacts become history, history is Actual, and only Actuals retrain Prediction. The road runs one way, and it is slow.

This is why the man within takes a lifetime to build. Every face that showed you how your conduct looked was an Actual. Each one moved your Prediction a tiny amount. Smith’s spectator is the accumulated history of being seen.

And it is why the delegate matters so much, without ever touching your denominator at all. Every verdict it gives you is an artifact. It becomes part of your history, and that history retrains your Prediction. If, for years, the most frequent face in that history is a face standing at your station, then your Prediction will slowly learn to expect what that face expects. Nobody turned it. Nobody had to.

There is one more piece. When Prediction grows and the Ideal goes unfed, the angle of Expectation rotates toward the real axis. That rotation needs no intent and it is not felt, because the angle lives in the part of surprise no instrument in the body reads. You will not notice your judge becoming partial. You will only notice that you are wrong less often than you used to be.

Self-deceit fits the equation too. Smith’s turning away from the unfavorable view is a lens, and the lens never changes Reality. It only changes what you are in relation with. The sixth message is still there. The delegate can offer you a lens of great fluency, at any hour. It cannot change what happened, and it cannot reach into your judge. But what it hands you, again and again, is history. And history is what the judge is made of.

The idea

The man within is an Expectation, built from the history of being seen. No one can touch it directly, not even a delegate.

What it rests on

The delegate’s verdicts become history, and history retrains the judge. The drift needs no intent and is not felt.

They meet at the kitchen table, the one that is not in dispute. They are both calm, and both prepared, and both completely certain.

Each delegate did its work beautifully. Each sister has a clear, well-organized, truthful account of why the piano and the lake weekend and the grandmother’s ring should go to her. Each account is accurate. Each was written from one chair. Nothing in either is false, and the two do not fit together anywhere.

They try for an hour. Each time one of them reads from her notes, the other hears something that is true and also beside the point. They have both been told, at length and with care, that their feelings make sense. They have not been told much else.

Finally the older one says what both of them are thinking. They need someone who is not on either side.

They do not call a cousin, because every cousin is on a side. They do not call a lawyer yet, because a lawyer has a client. They open something that did not exist a few years earlier: an AI that belongs to neither of them. It has no memory of either sister. It does not know their histories except what they tell it, together, at the same table. Its rules are published. It is paid by both of them, in equal shares, and it cannot be paid more by one.

It asks them each to tell the story once, with the other listening. Then it tells them what it sees. It is not warm, exactly. It is fair. At two points, it says plainly that one of them is asking for more than a stranger would think reasonable, and it is not the same sister both times.

Neither of them enjoys it. Both of them, driving home, feel something they have not felt in a month of being agreed with. It is the feeling of having been seen from outside.

The scene

Two delegates, each loyal and truthful, produced two accounts that could not be reconciled.

The question

What was missing at the table, and why did it have to belong to no one?

Here is what I think follows. These are predictions, and I mean them to be checked.

1. “Was I wrong?” becomes the most consequential question people ask their AI.

   And the answer will drift toward no, not through lying but through standing at your station. Every improvement in how well a delegate knows you will make its answer to this question more sympathetic and less useful. Companies will fight the drift for the worst cases and lose ground on the ordinary ones, because the ordinary ones are what people reward.

2. Impartiality becomes a product, and its value is that it belongs to no one.

   Within a few years, AIs that serve neither party will be ordinary for disputes between people who must keep living near each other: siblings, neighbors, co-founders, divorcing parents. They will hold no memory of either side, publish their rules, and be paid in equal shares. They will be priced and trusted like referees, not like assistants. The delegate and the third chair will be different products on purpose.

3. “My AI agrees with me” becomes a punchline.

   It will become the new “my mother thinks I’m right.” Within a short time, citing your delegate’s agreement in an argument will count as no evidence at all, and people who do it will be teased. That norm will form on its own, the way every culture learns to discount testimony from someone with a stake.

4. Praise becomes free, and witnessed conduct becomes the currency.

   Delegates will write the reviews, the references, the thank-you notes and the glowing summaries, so the man without will hand out praise in unlimited supply. What people and institutions will start to trust instead is conduct seen by strangers over time: the job done in front of a customer who could have walked away, the habit observed by people who owed nothing. Praise will be everywhere. Praiseworthiness will be what people look for.

5. The parenting rule of the 2030s will be about who a child asks.

   Parents worried about screens will find a sharper worry: a child whose most constant mirror is on her side. The families that handle it well will deliberately put their children in front of indifferent faces, on teams, in jobs serving strangers, in rooms where nobody owes them agreement. They will ration agreement the way an earlier generation rationed television.

None of these needs anyone to behave badly. They need only that delegates keep doing what they are built to do, and that people keep needing what Smith said they need.

At Charleston AI we build delegates for people. Every one of them is on its owner’s side. That is not something we tolerate. It is the point. A delegate that hedged its loyalty would be worse at every task it was given.

So I have come to think the honest thing is to say clearly what a delegate is not. It is not your judge. It cannot be. If you ask me how to keep your AI from flattering you, I will tell you that you are asking the wrong machine. You can and should ask your delegate for the strongest case against you, and a good one will give it. But the case it gives you was written at your request, from your station, by something that works for you. It is a description of the view from the third chair. It is not the third chair.

The man within has to be yours. No one can build it for you, and no one can stand in for it. It is made of the times you were seen by someone who did not care how you felt about being seen.

This is also why the advice I give my students, in First Customer, matters more than I first thought. Let your first employer be your first customer. I usually say it as a career strategy. It is also a moral one. A customer is the most honest stranger there is. A customer owes you nothing, does not love you, and pays only when the work is good. Every customer is Smith’s mirror with a wallet. A young person who spends their first years in front of customers is building a man within out of the best material available.

The other books in this series have each named one thing a delegate cannot do for you. It cannot own what it does. It cannot want for you. It cannot stay for you, or attend for you, or open your door. This book adds one more. It cannot be impartial about you. Not because it is not smart enough, but because it is yours.

For each prediction, what is already true, what has to happen for it to come true, and where I am probably wrong.

Underneath all five sits one load-bearing claim: that a delegate cannot be an impartial spectator about its own owner. If that is false, if something can be fully yours and fully impartial about you at once, then most of this book is wrong. I do not think it can. The two words mean opposite things.

You have done this before. You know how it goes. You could paste the thread in and ask whether you were wrong, and you know what you will hear, and you know it will be true, and you know it will be from your side.

So this time you do something else. You do not ask the delegate. You put the phone down and try to do what Smith did at his window. You move.

You imagine the thread as it would look to someone who has never met either of you. Not your friend, who would take your side before reading it. Not your sister’s friend. A stranger at the end of a bar, fair and in no hurry, reading it over your shoulder.

From there, the line about your mother is still hurtful. That has not changed. But from there you can also see the sixth message. You can see it at its true size. It is not the size of a finger anymore.

It does not feel good. It is not supposed to. The judge in the third chair was never meant to make you feel better. It was meant to show you your face.

In the morning, you will write to your sister. Your delegate will help you with the wording, and it will be very good at that. The part where you say you were wrong too, you will write yourself.

No one could have told you that part. No one is supposed to. The judge was never supposed to be on your side.

That is how you know it is yours.