You Are Right About the Prizes. You Are Wrong About What Gets Missed.

Two things share the one word, and only one of them is graded. Mechanics first, before the argument:

• Nobody is important enough to leave a public legacy, and private legacy has no entry requirement. The horizon is shorter than you think. Jan Assmann drew a line between communicative memory, which living people can personally recount, and cultural memory, which institutions, texts and rituals maintain on purpose. Communicative memory, on his account, runs about eighty to one hundred years, or three to four overlapping generations, and then stops. Almost nothing converts.

• The audience is about four people. They already know the achievements, because they were present for most of them. What they do not have is what you were thinking at the time.

• Specific outranks interesting. In the Autobiographical Interview (Levine and colleagues, Psychology and Aging, 2002), episodic details are tied to a particular time and place, and semantic details are true of your life but attached to no single moment. Only the episodic ones are inheritable.

• You undervalue ordinary material. Zhang and colleagues (2014) found people underestimate how much they will later want to revisit records of their own present, and that the error is larger for ordinary experiences than extraordinary ones.

• Going blank is a retrieval failure, not an empty life. Koppel and Rubin (2016) concluded that retrieval, rather than encoding and retention, drives the reminiscence bump, and that what people produce changes with the cue.

• The fix costs nothing. Ask four people what they would want to ask you in twenty years. Answer three of those into your phone. Put the files where two others can find them, and tell those two where.

Now the concession, because the premise is correct. You will almost certainly not be historically significant. Neither will your neighbours, your colleagues, your doctor, or whoever wrote the last legacy article you read. Historical significance is not a bar you fell short of. It is an accident that misses essentially everyone.

Name three winners of the Nobel Prize in Chemistry. Take as long as you like. Now name your eight great-grandparents and the towns they were born in. Most people can do neither, and the second failure is the one worth your attention, because those eight were not obscure. Each was somebody's whole world, inside living memory.

They did not fail to be impressive. Plenty of them were the most important person in somebody's life for forty years. Transmission failed. Nobody wrote anything down, or what got written down was a name and two dates.

So if the standard is whether strangers in 2160 know your name, the answer for you, for the chemists, and for almost everybody who has lived, is no. Drop that standard. Nobody was applying it to you.

Look at how the word gets used. Legacy giving. Legacy planning. Legacy of impact. Leaving a legacy. Every one describes a transfer to an institution or to strangers, judged by size. Of course you fail that test. It is a test about assets and reach, and you have normal amounts of both.

The thing families actually lose has no marketing budget behind it, so it has no word, so it never appears in the results you searched. Absence from a search result is not evidence of unimportance.

Somebody did try to measure the thing. In 2008 Marshall Duke, Amber Lazarus and Robyn Fivush at Emory University published a brief report on the "Do You Know" scale, twenty questions about family history put to adolescents. Knowledge of family history correlated with higher self-esteem, a more internal locus of control, better family functioning and lower anxiety, strongly enough that the authors proposed the scale as a clinical index.

That finding gets stretched, so take it at its real size. Brief report, modest sample, correlational. Families that tell a lot of stories differ from families that do not in a hundred other ways, and the study cannot separate those. It does not show that telling stories causes anything.

Fivush has since been clear about what the scale detects. Writing in Psychology Today in November 2025, she called the twenty questions "a rough index of whether this is a storytelling family". A child can only know these answers if somebody said them within earshot, which moves the requirement from having to telling.

Look at what the questions ask for: where your mother grew up, where your father grew up, how your parents met, where some of your grandparents grew up and where they met, the source of your name, illnesses and injuries your parents had when they were younger, the jobs your parents had when they were young, the names of the schools your mother went to. Nobody asked those children whether their grandparents had won anything. The name of your mother's school is not an anecdote. It is an ordinary fact repeated often enough to survive, and surviving is the whole bar.

Most people who say they are not important enough mean something narrower: that nothing worth telling has ever happened to them. Two words are fighting there.

Interesting is judged by a stranger. It needs a plot and a payoff, and it has to earn attention from somebody with no investment in you. Almost nobody's life clears that bar, which is why the relatives who tell great stories tell the same four.

Specific has a different judge, and that judge already cares. All it needs is a detail that is true, located and yours. No punchline. Nobody was ever there for the punchline.

"I was a nervous teenager" is semantic. "I was sick in the toilets before my first shift at the garden centre in the summer of 1979" is episodic. Both are true. Only the second is inheritable. A grandchild can do nothing with a nervous teenager. The garden centre puts them in the room.

Levine's team found something else that should bother you. Older adults drift toward the semantic and away from the episodic, and the pattern held even after extra structured probing for contextual detail. One polite question moves nobody off the general.

A long relationship is not a document. A long relationship is thousands of small consultations, most of them not worth repeating a week later. What do you make of this. Is this normal. Should I ring him back. Does this look infected. Am I overreacting. Would you have signed.

None of that is a request for a monument. His wife was flagging the loss of a service she uses constantly and has never once thought about, like the tap until the water stops.

The demand does not stop when you do. In 1996 Dennis Klass, Phyllis Silverman and Steven Nickman published Continuing Bonds, which gathered the evidence that bereaved people do not detach from the dead and were never meant to. They keep relating. They talk to the person, ask them things, imagine the reply, argue with the imagined reply. That revised a clinical model which had treated letting go as the goal.

So the asking already happens. Widows already ask their husbands things. Adult children already ask their dead mothers whether they are about to make a mistake. The only variable is whether anything comes back, or whether the person supplies both halves and wonders whether they got you right.

Guessing degrades in one direction. It flattens. The dead get reduced to two or three traits and a couple of catchphrases, kinder and duller than the person was, which is what being remembered wrong looks like from the inside.

You are also picturing the wrong room. Public legacy assumes strangers, and with strangers the measure is scale, so four people is a rounding error. Private legacy has an audience you can name out loud right now, and the speed at which the first three names arrive tells you the list was already there.

Those four know the impressive parts. They were present for most of them and have heard the rest at dinners. What they do not have is the interior: what you were thinking while it happened, what you decided against, what frightened you and did not show. They also ask differently once they are grown, which is what your child will want to ask you at thirty.

A small audience removes the quality bar too. A recording made for four named people has to be yours and findable, and both are jobs you can do badly and still succeed at. Nobody in that group is grading your delivery. They are listening for whether that is really how you talk.

You do not need to be important. You need to be specific, and you need somebody to ask.

Four people asking you things for the rest of their lives is a heavier load of use than a million strangers reading a paragraph about you once. Public legacy is measured in reach. What you have is frequency.

Start building your legacy Free build, 50 memories, no card required.

This part has been tested, and you lose.

In 2014 Ting Zhang, Tami Kim, Alison Wood Brooks, Francesca Gino and Michael Norton published four studies in Psychological Science, the first using a time-capsule design. Participants recorded details of their present lives, predicted how interested they would be in reading those details back later, and were asked again months afterwards. They consistently underestimated how much they would want to revisit their own records, the error was larger for ordinary experiences than extraordinary ones, and the authors put it down to misplaced confidence in memory for everyday events.

That mechanism is your objection, precisely. You judge future value by present value, and ordinary things have almost no present value, which is what ordinary means. The extraordinary you already know to keep. The one category you reliably throw away is the category you are most wrong about.

Age makes the mistake worse. In "Happiness from Ordinary and Extraordinary Experiences", published in the Journal of Consumer Research in 2014, Amit Bhattacharjee and Cassie Mogilner ran eight studies on the happiness people take from the two kinds of experience. Extraordinary experiences delivered at every age. Ordinary ones delivered more as people got older, and self-definition was the mechanism: ordinary experiences become increasingly self-defining as the sense of time remaining shortens.

At forty you score your material on how unusual it was, and by that scoring it comes out worthless. At seventy-five you will score the same material on how much of your life it actually was. Nothing about the material changed. The scoring did. And the forty-year-old is the one deciding whether any of it gets recorded.

So the ordinary material is the content, not the scenery. How you talk yourself down from a panic, in the actual sentences, in the order you say them. What went through your head the first time you held them, including the part that was not tender.

There is a quieter version of this objection: that recording yourself would look like self-importance. The standard is never applied evenly, though. Nobody who says "I am not important enough" goes on to say "and neither was my mother", and nobody has ever apologised for being in the family photographs.

Ask most people for something worth recording and they go blank, and the blankness feels like proof. It is evidence about the question rather than about your life.

Jonathan Koppel and David Rubin reviewed the reminiscence bump in Current Directions in Psychological Science in 2016. The bump is the documented excess of memories from youth and early adulthood found in adults over forty, and their review draws out that its shape changes with the cue used. Important memories produce one pattern. A plain word cue produces a smaller bump, earlier in life. An odour cue produces one earlier still. Their conclusion is that retrieval, rather than encoding and retention, drives the effect.

So memory is not a shelf you walk along. A closer analogy is a search that returns whatever the query reaches. "Tell me something interesting" is close to the worst query available, because it searches on importance, and importance throws away almost everything you own.

Try that on somebody else tonight, because you will believe your own eyes about a parent long before you believe an argument about yourself. Ask anyone with twenty years on you the general question. What was your childhood like. What comes back is a summary: it was happy, money was tight, there were a lot of us.

Then ask one small question. What did the kitchen smell like on a Sunday. What was the exact route you walked to school.

Often they look away, because they are looking at something. Then the small things arrive. Street names. A shortcut through somebody's yard. Whichever dog on the route everybody crossed the road to avoid. The exact words their mother used if they were late. The material was there the whole time.

Six to work from. Ask two or three, never all six.

1. What did the kitchen smell like on a Sunday when you were ten?

2. What was the exact route you walked to school, street by street, and who did you walk with?

3. How much was your first pay packet, and what did you spend it on?

4. What were you frightened of at nineteen?

5. What was the first car you owned, and what noise did it make?

6. What did your mother say, in her exact words, when she was angry?

Not one of those asks for a story. Every one asks for a detail, and the story arrives behind the detail. A longer bank sorted by subject is at a fuller set of life story questions, and how to interview your grandparents covers that conversation without turning it into an interrogation.

Then answer them about your own life. Most you can do right now, unprepared, with nothing remarkable ever having happened to you. That is material, produced by a test you were free to fail. What you lack is anybody asking properly, and the things only you know how to do points the exercise at yourself.

One more finding, because it is why we bother with voice. In 2010 Leslie Seltzer, Toni Ziegler and Seth Pollak published a study in Proceedings of the Royal Society B. They applied a social stressor to a group of girls, then assigned each at random to full contact with their mother including touch, their mother's voice only over the phone, or no contact. Full contact produced the highest oxytocin levels and the fastest return of salivary cortisol to baseline. The voice-only group showed what the authors called a strikingly similar hormonal profile. The no-contact group did not. Vocalisations, they concluded, may be as important as touch in regulating social bonding.

The limits matter, because that study gets stretched further than it goes. Children and their mothers, not adults. A live phone call, not a recording. It says nothing about recorded or synthesised speech, and anyone citing it as proof that a voice clone comforts the bereaved has walked past the evidence. Voice carries something on a channel separate from the words.

The walk to school, written down, reads like notes. Said out loud by the person who walked that route, the same words stop being notes. If the word clone is what stops you, what voice cloning actually involves sets out consent, authorisation and withdrawal, and how to keep your voice away from people who would misuse it covers the rest.

Four steps. Three of them have nothing to do with us.

Ask them what they would ask. Send this, more or less word for word, to the three or four people who would come looking. "Odd question. If you could ask me anything in twenty years and get a real answer, what would you ask?" Do not explain why. Almost none of the replies will be about your achievements, and that one step swaps your estimate of your worth for theirs.

Answer three of them badly. Voice Memos, one take, no script, three to five minutes each. Say the question out loud at the start so the recording makes sense to somebody who was not there. Do not tidy. Tidy is how you drift back into the summary, and nobody needs the summary. A plain method for recording your life story sets out the order and the file naming.

Put them where two other people can find them, and tell those two where. Shared drive, shared album, whatever they already use. A recording nobody can locate is not a recording. This is where these projects quietly die.

Do it once a month. In a year you have thirty-six answers in your own voice, more than almost anybody leaves behind, and you have spent nothing. If the reminder slides, a journal that asks you the questions is the prompted version, and the self-interview method is the long one.

Afterlife.ai™ is built on that argument: the valuable material is ordinary, it only comes out under a specific question, and nobody produces any of it by being asked to write about their life.

So we do not ask you to write. A guide called Idy interviews you in conversation, one question at a time, chasing specifics instead of summaries. Not what your father was like, but what he said the last time the two of you disagreed. Idy follows the answer you gave rather than the next item on a list, and returns to the same ground from another angle another day, because the second pass is where the specifics turn up.

What comes out is stored as memories, each one a discrete thing you said, kept with its context. What we build from those has a name, and we call it a "Persona". Personas answer from your memories rather than from a general model doing an impression of somebody like you, and where you never covered a subject they do not invent the rest. A concrete fact comes from your memories or gets cut, and one class of invention, a claim that you said something you never said, is checked in code and stripped mid-stream. Personas speak from what is there and leave the silence for the other person to fill. An AI trained only on you sets out the mechanics.

You record enough speech to clone your voice, so answers come back in your voice rather than a stock one. You nominate who gets access, and nobody else gets access, including anybody who would inherit your estate by default. Release runs through the Executor Lock™ verified release process: a nominated person reports that you have died, a seven-day clock starts, evidence is checked before anything opens, and if you are alive you stop the clock. The lock takes a snapshot with nothing pruned or down-weighted, and the personality cannot be changed after that by anyone. Before release everything stays private, and you can read back what was captured, edit it, delete individual memories, and close the account.

Now the limits, which you should hear from us rather than find out later. A build worth talking to comes together over weeks of short sittings, so expect a small habit. The interview is better on concrete ground than abstract: point Idy at a route, a job, an argument or a room and you will get somewhere, point Idy at what you believe about love and you will get sentences you have said before. Refusing to invent is a rule we write into the prompt and enforce in code where the claim is checkable, not a guarantee, because no language model holds a line perfectly every time. And nobody has run a controlled study on whether your four people will use what you make.

The first 50 memories are free, no card is asked for, and the free build never expires. That is deliberate. The worst version of this industry charges grieving families at the moment they are least able to refuse. None of the four steps above needs an account.

Nobody has ever come looking for a CV. They come with a question, at a bad time, years later, and either something answers in a voice they know or they supply both halves and hope they got you right. Answer three questions badly tonight and hear whether what comes back sounds like you.

Start building your legacy Free build, 50 memories, no card required.

Keep reading: what you would even say to an AI of yourself, being remembered wrong, the self-interview method, what the paid plans include.

Every claim above was read at source rather than taken from coverage, and where an author has publicly qualified their own finding, as Fivush has, the later account is used. The "Do You Know" items quoted here come from Fivush's own published list, and the quoted sentence about being asked questions comes from a real conversation. What is not settled is most of what matters practically: the family-history findings are correlational, the oxytocin study used live phone calls between children and their mothers, and nobody has run a controlled study on whether an AI of a dead person helps or harms the bereaved. Found something wrong or out of date? Send us the correction and we will fix the page at the next review.

• Zhang, T., Kim, T., Brooks, A. W., Gino, F., & Norton, M. I. (2014). "A 'present' for the future: the unexpected value of rediscovery." Psychological Science, 25(10), 1851-1860. DOI 10.1177/0956797614542274. PubMed record.

• Seltzer, L. J., Ziegler, T. E., & Pollak, S. D. (2010). "Social vocalizations can release oxytocin in humans." Proceedings of the Royal Society B, 277(1694), 2661-2666. DOI 10.1098/rspb.2010.0567. PubMed record.

• Duke, M. P., Lazarus, A., & Fivush, R. (2008). "Knowledge of family history as a clinically useful index of psychological well-being and prognosis: A brief report." Psychotherapy: Theory, Research, Practice, Training, 45(2), 268-272. DOI 10.1037/0033-3204.45.2.268. PubMed record.

• Klass, D., Silverman, P. R., & Nickman, S. L. (eds.) (1996). Continuing Bonds: New Understandings of Grief. Taylor & Francis. Routledge reissue DOI 10.4324/9781315800790.

• Assmann, J. (2008). "Communicative and Cultural Memory." In A. Erll & A. Nünning (eds.), Cultural Memory Studies: An International and Interdisciplinary Handbook, pp. 109-118. Walter de Gruyter. DOI 10.1515/9783110207262.2.109.

• Levine, B., Svoboda, E., Hay, J. F., Winocur, G., & Moscovitch, M. (2002). "Aging and autobiographical memory: dissociating episodic from semantic retrieval." Psychology and Aging, 17(4), 677-689. PubMed record.

• Koppel, J., & Rubin, D. C. (2016). "Recent Advances in Understanding the Reminiscence Bump: The Importance of Cues in Guiding Recall from Autobiographical Memory." Current Directions in Psychological Science, 25(2), 135-149. DOI 10.1177/0963721416631955. PubMed record.

• Bhattacharjee, A., & Mogilner, C. (2014). "Happiness from Ordinary and Extraordinary Experiences." Journal of Consumer Research, 41(1), 1-17. DOI 10.1086/674724. Publisher record.

• Fivush, R. (2025). "The 'Do You Know' Family Stories Game for Thanksgiving." Psychology Today, 24 November 2025. Article.

• Fivush, R. (2016). "The 'Do You Know?' 20 Questions About Family Stories." Psychology Today, November 2016. The published list of the twenty items, used for the item wording quoted on this page. Article.

Last reviewed: 30 July 2026. Next review due 28 October 2026. This page carries no legal or regulatory claims. The research citations and the product mechanics in the final section are on a 90-day recheck cycle.