Act One · The helmet
A graduate student lowers the head-mounted display over your eyes and the room goes out like a light. When it comes back, it is the same room — the same dimensions, the same white walls — except that it is made of triangles, and two metres behind you there is a mirror that was not there before.
The display is an nVisor SX: two panels, 1,280 by 1,024 pixels each, refreshing sixty times a second. Cameras on the ceiling track where your body is in the room; a sensor in the helmet tracks the pitch, yaw and roll of your head. All of that machinery exists for one purpose this morning, which is to make a reflection behave like a reflection.
You turn your head. The face in the mirror turns its head. You lean left; it leans left. It is unmistakably obeying you, which is how mirrors establish that they are about you and not about somebody else.
And it is about seventy years old.
The face was built from two photographs of you — one from the front, one from the side — using a program called FaceGen Modeller, which finds the landmarks of a face, wraps a three-dimensional model around them, and then runs an ageing algorithm over the shape and the skin. An artist greyed the hair afterwards in Photoshop. The whole thing took a technician an afternoon. It takes you about four seconds.
“At the extreme, with a total lack of psychological connectedness, a person’s future self might seem like a different person altogether.”Hershfield, Goldstein, Sharpe, Fox, Yeykelis, Carstensen & Bailenson, Journal of Marketing Research, 2011
You are given about a minute in there — sixty to seventy-five seconds of looking, plus a scripted conversation with a researcher who asks you innocuous questions while you talk in that voice out of that face. Then the helmet comes off, the real room comes back, and somebody hands you a sheet of paper about money.
Before we get to the money: the mirror. Build one. It will not be a photograph of you — nothing on this page uploads, scans or keeps a picture of anybody. It is a face drawn in code from a handful of dials, and you can push it toward yourself or away from yourself as much as you like.
Set the dials until the face is roughly plausible as you — or as nobody in particular, if you would rather. Then turn the years.
Skip any of this you like. Nothing here is scored, and there are no right answers — the only graded thing on this page is the quiz at the end.
Act Two · The scanner
Two years before the mirror, in 2009, three researchers at Stanford — Hal Ersner-Hershfield, Elliott Wimmer and Brian Knutson — slid eighteen undergraduates into an fMRI scanner and asked them to do something very boring. Words appeared: brave, lazy, curious, honest. The task was to say whether each word described a person.
Which person changed by block. Sometimes it was you, now. Sometimes you, in ten years. Sometimes it was a famous stranger — Matt Damon for the men, Natalie Portman for the women — now, or in ten years.
A patch of cortex called the rostral anterior cingulate did something legible. It fired one way for me and another way for a movie star — no surprise there. The surprising part was where me in ten years landed. For some people it sat close to me. For others it drifted toward the movie star.
A week later, the same eighteen people came back and answered questions about money: would you take a smaller amount now, or a larger amount later? The size of that drift in the scanner predicted their answers — r(16) = .47. The people whose brains treated future-them more like a stranger were the people who did not want to wait for them.
Hold on to that number for a moment, because it is small, and Act Six is going to be rude about it.
Meanwhile, the same group needed a way to ask this question without a two-million-dollar magnet. What they landed on is two circles — borrowed from a 1992 scale that measured how close you feel to another person, and pointed instead at your own future. Seven pairs, from not touching at all to almost on top of each other. Pick the pair that feels right.
The real instrument asked about ten years. We are going to ask four times, because the interesting thing is not the number — it is where the number falls off.
Your own overlap ratings, 1 (no overlap) to 7 (almost the same circle).
Fill in all four and the line will draw itself.
Skip any of this you like. Nothing here is scored, and there are no right answers — the only graded thing on this page is the quiz at the end.
In late 2016 the Consumer Financial Protection Bureau put a version of this question to a nationally representative sample of over six thousand Americans, scored 0–100. The average was 68. The median was 80. The single most common answer was 100 — a great many people say they are exactly the same person they will be later. Roughly a third said something much lower. Continuity rose steadily with age, which is the opposite of what most people guess, and it rose with income too. And it had no relationship at all with whether someone described themselves as a planner. The researchers called that “curious”, which is academic for “we did not expect that”.
Act Three · The money
Back in the lab. Fifty students, helmet off, blinking. Half of them had spent their minute with a mirror showing their face as it is. Half had spent it with the seventy-year-old. Now everyone gets the same imaginary windfall and the same four envelopes: spend it, put it in a current account, treat someone, or lock it away for retirement.
The students who had just met the old face put an average of $172 in the retirement envelope. The students who had met their current face put in $80. Same room, same money, same questionnaire; the only difference was whose face had been in the mirror sixty seconds earlier.
Three more studies followed, without the helmet — photographs on a screen, a slider that made the face smile as you saved more and sag as you spent, and a version with a completely neutral face to prove the smiling was not doing the work. All of them pointed the same way. Within a year the effect had left the lab entirely: in 2012 a bank shipped a web tool that took a photograph through your webcam and aged you on the spot, with a savings calculator underneath.
Underneath all of this sits a much older, much weirder fact about wanting things.
One apple today, or two apples tomorrow? Now try again: one apple in a year, or two apples in a year and a day?Richard Thaler’s illustration of dynamic inconsistency, Economics Letters, 1981
Almost nobody takes the single apple in the second question. Plenty of people take it in the first. The wait is identical — one day — but in the first question the day stands between you and something you could have now, and “now” is a drug.
Thaler measured it. He asked people how big a future prize would have to be to feel as good as $15 today. Three months out, the median answer was $30. A year out, $60. Three years out, $100. Those are annual interest rates of 277%, 139% and 63% — the same person, charging themselves wildly different rates depending on how far away the money is. Not a straight line down. A cliff, then a slope.
Here is that cliff, run on you. Six choices, twice. The bench will draw your own curve.
Nothing real is at stake, which is exactly the problem with measuring this — see the wobbles. Pick something you actually want, and answer fast and honestly rather than sensibly.
Two questions. Answer the first before you read the second.
One apple today, or two apples tomorrow?
One apple in a year, or two apples in a year and a day?
Skip any of this you like. Nothing here is scored, and there are no right answers — the only graded thing on this page is the quiz at the end.
Act Four · The volunteer
In 2008, before any of the mirrors, three psychologists at Princeton — Emily Pronin, Christopher Olivola and Kathleen Kennedy — published a paper with a title that gives away the whole game: Doing Unto Future Selves As You Would Do Unto Others.
Their trick was to make the decision real. In one study, participants had to decide how much of a genuinely unpleasant liquid somebody would have to drink in the name of science. In another, they signed somebody up for hours of tutoring during exam week. Sometimes that somebody was themselves, right now. Sometimes it was themselves, next semester. Sometimes it was another student.
The pattern held across studies: what people signed future-them up for looked like what they signed other people up for — and both looked nothing like what they were willing to do this afternoon. The effect showed up strongly when the decision was real and faded when it was hypothetical, which is the opposite of how a mere reporting artefact usually behaves.
Future-you is generous, energetic, disciplined and available. Future-you would love to. Future-you is also, conveniently, not in the room.
So: sign somebody up.
Pick a job that is dull but genuinely useful. Then decide how many minutes of it each of these three people is signing up for.
Skip any of this you like. Nothing here is scored, and there are no right answers — the only graded thing on this page is the quiz at the end.
Act Five · The nobleman
None of the psychologists got here first. In 1984 the philosopher Derek Parfit published Reasons and Persons, a book that spends several hundred pages taking the idea of a single continuous self apart to see what is holding it together, and concludes: not much. What there is, Parfit said, is connectedness — a rope of overlapping memories, intentions, beliefs and habits. Ropes have no single thread running end to end. They hold anyway.
“My concern for my future may correspond to the degree of connectedness between me now and myself in the future… Since connectedness is nearly always weaker over long periods, I can rationally care less about my further future.”Derek Parfit, Reasons and Persons, Oxford University Press, 1984
Note the word rationally. Parfit is not scolding anybody. He is pointing out that if future-you really is only partly you, then caring less about them is not a bug in your character. It is arithmetic.
Then he tells a story that will not leave you alone. A young Russian nobleman is about to inherit vast estates. He is a socialist, and he intends to give the land to the peasants. But he knows what money does to people, and he suspects that in thirty years he will not want to give anything away. So he asks his wife for a promise: when the time comes, hold me to this — and refuse to release you from the promise no matter what I say later.
Thirty years pass. The land arrives. The older man, now conservative and comfortable, asks his wife to let him keep it.
Whom did she promise? Is she still bound? The older man is, by every legal standard, the same person who asked. By Parfit's standard he may be a successor — someone connected to the idealist by a rope that has thinned in the middle. She made a promise to a man who, in a real sense, is no longer here.
You do this to yourself constantly. Every alarm you set for the morning is a Russian nobleman's promise. So is every list you leave on the kitchen table, every "I'll deal with it Sunday", every glass of water beside the bed.
Which raises a question worth two minutes of your morning: what do these two actually want to say to each other?
Nobody reads these but you. They are not sent anywhere, and nothing on this page leaves this page unless you copy it out yourself.
Skip any of this you like. Nothing here is scored, and there are no right answers — the only graded thing on this page is the quiz at the end.
Act Six · Where this gets shaky
Everything above is real, published and checkable. That is not the same as settled. Here is where a careful reader should put their weight down slowly.
The headline result came from 50 undergraduates, split into two groups of 25. The standard deviation on that $172 was $214 — bigger than the mean. The test was one-tailed, p = .035, effect size d = 0.52. That is a genuine, medium-sized finding and also a very small study, and the money was imaginary.
In 2024 Kausel and colleagues ran the idea on 415 real customers of a real investment firm, with a real aged photograph. Intentions to save went up. The amount actually invested moved a little, briefly. A 2025 systematic review of 14 articles and 23 studies called the approach “promising”, with effect sizes ranging from negligible to very large — and several flat nulls.
The scanner study had 18 undergraduates and reported a correlation of .47 between one brain measure and one behaviour a week later. Correlations that size in samples that small are exactly the kind that later shrink. The finding is a clue, not a foundation — and “your brain files your future self under stranger” is a slogan, not what the paper says.
In the community sample, feeling connected to your future self correlated with having more assets (r = .34, and .23 once age was controlled). That is a correlation. Having savings may well be what makes the future feel like it belongs to you, rather than the reverse. And an earlier researcher, Shane Frederick, looked for this link in 2003 and found nothing at all.
The bench above fits a curve to six hypothetical choices you made in ninety seconds, about apples you will never eat. Real discounting studies take longer, pay actual money, and still find that the same person answers differently on a different day. Treat your curve as a snapshot of this morning, not a measurement of you.
The mirror, not the verdict
I could have all of this wrong. I am counting clicks and characters. You are the one who was there.
Eight questions on the people, dates and findings in this lesson. Your own answers above are not on the exam and never will be.
0/8
You need two people, two envelopes, two sheets of paper and a pen. Agree on a date — a month out is plenty; a year is better.
Each of you writes a letter to yourself on that date. Not goals. Not resolutions. Three things: what today actually smelled and sounded like, one thing you are quietly worried about, and one question you want them to answer. Seal it, write the date on the front, and swap — you hold theirs, they hold yours, because the surest way to lose a letter is to be its own postman.
Boss level. Before sealing, add a promise the way the Russian nobleman did: one instruction your future self is forbidden to talk you out of. Then, on the day, read it aloud and argue with it in front of the other person. Notice whether you are defending the letter or defending yourself against it.
And if the letter turns up something heavy — most of them do, sooner or later — that is a good moment to talk it through with someone you trust. Working things out loud with another person is not a failure of self-reliance; it is most of how anybody works anything out.