one of my goals in posting this is to process, reflect, and articulate. the other goal is that by going through this reflection, i can hopefully learn some good lessons.

in it… i try to be unflinchingly honest, for better or for worse :)

quick recap: our initial goals, assumptions, and thought process

i thought it’d be interesting to start with the original MVP doc we wrote before building the app (Day 0)…

here’s the MVP Product Summary:

here’s our statement about purpose and vision, and the problem/solution:

here’s our product goals:

and here’s what we thought our MVP “magic” moments would be:

how the cake turned out when we took it out of the oven

so drumroll… how’d the cake turn out? well, we have a functioning web app that works in 8 languages. users can have a conversation with a voice AI agent about ~40 different topics we painstakingly hand created, and can invite other users to be able to access and listen to those conversations. below shows a “feed” of my dad’s recordings, all in korean (his native language)…

and, since (even though I can speak and understand korean reasonably well) my native language is english, the app generates conversation summaries in each user’s own preferred language…

and below is what we thought would be our “magic moment”: users can generate literary memoirs, poems, short stories, or biographical chapters from these voice conversations with the click of a button, and publish them. you can see some real examples of published stories at https://ember.build/stories …

the results so far on the journey to PMF🥁

with minimal marketing (mostly friends & fam), we had ~100 signups … and 1 retained enthusiastic user (our dad). oh, and we hired ~5 summer interns to help us figure out marketing 😅. sigh…

some observations i have about our process that took us to this point and a few thoughts on those. honestly all of these are pretty stereotypical first time founder mistakes. but on the bright side, we did manage to push the ship (a MVP) out of harbor, so no need to lose too much sleep:

1. we didn’t validate or spend time talking to users or customers before building the MVP. the PRD above was written without a single formal user interview where we ran the PRD by them. in retrospect, i wish we’d first had a few focused, but open-ended discovery conversations with potential users. just finding people who find the intersection of memory, personal stories, and AI would have been incredibly valuable in helping us better understand what kind of things they’d want built (as opposed to what we think people want).

2. we didn’t bother with an exercise in defining a user persona. for example: our first N customers will be 50-70 year old semi-affluent parents of children in their 30s who have just recently been diagnosed with terminal cancer. ok, dark, i know… but also like, really specific, and… maybe a reasonable hypothesis on who might have the urgency to want to immediately use the app? being able to articulate our ideal user persona so clearly would provide extreme clarity on how exactly to go find those N customers (extremely specific instagram posts, loitering in radiology wards of local hospitals), and clarify our messaging.

3. we spent too long (2 months) building the MVP, and we wasted a lot of energy with intense debates about the app UI needing to be beautiful, differentiated, and perfect. after all, it’s a B2C app, and in a competitive B2C market, design wins… right? in retrospect, i don’t think that was a thoughtfully considered view. For one, we had theories of where the magic moments for users would come from and we actually wrote down. None were ever about the UI. so why spent so much time designing it? Funnily and sadly enough, post-MVP release, we actually ended up reskinning the UI 2 more times. My hot take - for AI products, design matters a lot less since the novelty and focus is going to be on the AI output. In fact, feels like a lot of fast growing B2P (prosumer) AI apps look like… B2B SaaS apps, favoring function and noncontroversry over form.

our marketing strategy

we decided on focusing primarily on instagram as our channel, all organic, and about 1 month ago began posting aesthetic visual, emotional-trigger-happy content on our @rememberwithember instagram account, with text carousels and reels of older folks talking about their fond memories. our inspiration pattern was @humansofny, which has 12.6M followers.

at the time of this writing, we have 42 followers on our instagram account 🙃

so far we’ve been taking a more organic approach as opposed to more of direct CTA/product demo strategy, or a founder-led approach. we’ve been able to attribute 1 user signup (but not retained) as coming from instagram.

what the people have been telling us

while we have a small sample size of absolute users who’ve signed up for our app (~100), i’ve had 7 user feedback calls with users who are not direct friends or family members (3 in person, 4 over google meets), and 1 general advice call with a pretty famous woman who operates a bookstore in korea and was a former marketing executive at Samsung/Cheil in Korea for over 20 years. i figured it’d be a good call to pitch the idea and get feedback since 1. she’s a literary person (and our app is a product that creates literary content with AI) 2. she’s had deep experience in marketing, and 3. she’s a business owner (and operates a related subscription-based book club).

i’ll start with the 3 things she left me with after mostly asking questions and listening, which i appreciated:

1. the most important thing is to articulate with conviction ‘why [the business] exists’. she told me that the first thing a visitor sees when they enter her bookstore is a huge bulletin board that states in the center the reason the bookstore exists, along with her own writing in smaller post-its around this center reason that have been her articulating other reasons - reminders to herself really - as to why her bookstore exists. she said that it helped her get through difficult times and survive other competitors who had less of a force of conviction to justify their own existence.

2. since the app is an AI app, it’s important to think about what the app can do that chatGPT and gemini can’t. she had asked me about our revenue model, and i had told her we planed to charge a subscription. in light of this, her point was that we had to justify the price with clear differentiation.

3. the app needs to be simpler and more fun. this was most surprising for me to hear, given that i had come into demoing the product to her with the assumption that as some famous literary figure in Korea who is presumably a serious highbrow person, she would appreciate how painstakingly we crafted the features to generate literary output and tuned the models. she specifically brought up that users of apps like to brag. “look at this poem i made”. and that the app needed to provide users with a way to make it easy for them to brag.

the other feedback calls

the most common pattern of feedback we get is this:

user: “the concept how is so lovely concept!” “the app is beautiful”. “my mom recently passed away and i think about how this could have been useful.” “i tried the conversation and it was a neat experience.”

us: “thank you, we really appreciate it! how many conversations have you tried?”

user: “oh, just one, just for onboarding.”

us: “ah i see. why is that?”

user: “i was just super busy these past few weeks. i promise i’ll try it again soon!”

us: “we appreciate that. what’d you think about our story generation feature?”

user: “what’s that?”

us: “oh, so after you complete a memory, you can generate memoirs, poems in different styles from it.”

user: “oh neat, i didn’t know that.”

us: “would you be interested in inviting your parents to try the app out?”

user: “mm i can ask, but honestly they’re not the most tech savvy people so it’s unlikely…”

us: “well we really appreciate your time! definitely try using the app sometime and let us know your thoughts.”

my takeaway from these calls is this: the general idea of creating an app to record and preserve family memories is lovely and quite agreeable. but that statement carries a similar energy as recycling plastic to preserve our precious earth. both are lovely and quite agreeable. but a hot app that statement alone does not make.

revisiting our initial “magic moments” theses

i’m going to provide the rebuttal to our initial magic moments (hindsight is obviously 20/20, and my rebuttals may not even be correct), but it’s a helpful exercise for me in processing why users weren’t feeling enough “magic” to return.

1. warm, engaging voice agent. what if it turns out that betting that a voice agent being warm and engaging is the reason it’s magical for a user is like betting on a warm, polite kind man to bring a woman to her magic orgasm in bed? as in - i’m sure some people, male or female, find nice people sexy and can get off just from that. but maybe it’s not enough. maybe there needs to be an edge. or the person needs to feel like a hero all night and the center of attention. more on this later.

2. immersive “topics”. yikes. even reading that original statement feels so half baked. we use the word “immersive” but there’s nothing in the subsequent clarifications that suggests why the topics are immersive. unless the fact that the topics are sequential, or about work, or about lessons, are immersive…? looking back, there just wasn’t much deep thought put into the concept of immersive. and therefore, this one felt doomed to fail. in fairness, we did spent a lot of time writing up custom prompts per topic and brainstormed them, but now that i’ve had some time to listen to podcasts (arguably the richest source/inspirational material for good immersive conversations in today’s era), it’s clear there a lot of opportunities for deeper, funnier, or edgier conversations.

3. “Building” flow. People naturally crave the feeling of ‘creating’ or ‘improving’. This one… definitely a few thoughts. First, I don’t think it’s a universal sentiment that people naturally crave the feeling to ‘create’ or ‘improve’. In fact, I would argue that people naturally crave consumption (of food, of social media content), but the urge to create may actually be limited to a much smaller subset of people. Now what I know personally (as someone who does actually crave creating (apps, writing, ideas, companies, doesn’t matter)) - is that when ‘creating’, or ‘improving’, what I’m really in love with is the process. The iteration, the dead ends, the exploration, the discovery, the pain, the frustration - I love it all. That process is what provides my brain with the stimulation, and the experience through which I learn. Put another way, for me, to create is to not take shortcuts. It’s to overanalyze everything. It’s to force myself to be an observer, to go on social media to study content, to go to cafes to find moments of inspiration, to read, to download apps I’d otherwise never download just to develop a tactile feel and intuition. To articulate what I like about something - to develop taste. To return back to the original planning docs months after we wrote them to learn and reflect. Yet, we designed an app that lets users generate stories with the click of a button, in many ways bypassing the process.

so in retrospect, how would i self assess how we decided on our 3 magic moments?

on the first point of view, we were wrong, but at least the point of view was well articulated.

on the second point of view - and in fairness, we do have a separate thought piece specifically focused on topics that’s reasonably thoughtful, but that notwithstanding - we definitely could have articulated/thought a bit more about the concept of what makes something ‘immersive’ (and presumably, therefore, ‘engaging’, that instagram euphemism for ‘addicting’ lol

the third point of view was by far our weakness, and the actual product decisions and workflows feels almost antithetical to how (at least for me personally) I think about what a true “creating” experience feels like? (hint: needs a lot more knobs and levers)

where does that leave us? looking ahead

¯\_(ツ)_/¯ that’s life. we learn and make the next iteration better. we’re still committed to creating magical experiences with AI with the goal of preserving and sharing stories.

in my next article, i’m talk more about the direction we’ll be exploring taking the next iteration of ember (hint: it has to do with podcasts, and the user as a hero). i’ll also be unpacking the observations i made after spending time deep in the world of user generated content across channels like instagram, tiktok, and spotify!

It is still there. As far as I know, it is empty.

I do not blame her. It was a bad gift, in the specific way a gift can be bad when it asks too much of someone. I have not written my stories down either. Almost nobody has. The notebook on my mother’s shelf is the same notebook on a million shelves, in a million rooms, waiting for a Tuesday afternoon that does not come.

The blank page is not neutral. It is hostile. It asks for a sentence before you have one. It asks for a shape before you know what you are doing. It asks you to decide how you feel before you have finished feeling it. Writers spend their whole lives learning to push past this and still describe the act of starting as something close to pain. My mother is not a writer. She raised two children. She ran a household. The idea that she would also, in her spare hours, compose prose about her childhood was always one of the polite fictions we tell ourselves about our parents.

She was never going to fill the notebook.

Most of what our parents and grandparents know about themselves dies because the cost of getting it out is just slightly higher than the cost of leaving it in. That is the whole thing. It is not laziness. It is not a lack of love. It is not the particular failure of your family or mine. It is a small tax the page charges on every story it touches, and the tax is too high for most people, most of the time.

But talking costs almost nothing. You sit. You talk. You are done. Someone asks you about your father and you are three sentences in before you notice you have begun. No margin. No cursor. No spelling. Nothing to delete. The story is already arriving.

We all know this. Your mom will call you on a Sunday for no reason and end up, somewhere in the middle of it, telling you about her mother’s restaurant. The porridge they served. The way she ate it each day after school, and how she could bring as many friends as she wanted for a free snack, which made her the most popular girl in school. These are the exact things we want the notebook, and it comes out because the phone is an excuse to talk and not a demand to write.

We have been handing notebooks to people who needed phone calls.

And yet voice, on its own, was never quite enough. Not because the talking failed. Because nothing survived it.

My grandfather was recorded once, by a cousin, at a family gathering in the nineties. Three hours of him. The tape exists somewhere. I have never seen it, and neither, as far as I know, has anyone else. This is what raw voice has always been: not a story but a box. Hours of audio with no shape, no way in, no moment marked as the one that matters. You cannot hand it down because no one can bear to sit through it, and so it becomes one more thing the family feels guilty about and never plays. The talking was real. It just had nowhere to go.

So voice was always the right medium. That was never the question. The question was what could be done with it once it left the room, and for all of human history the answer was: almost nothing. You could keep it. You could not use it.

That is the part that is changing.

What AI adds is not a better recorder. The cousin with the camcorder was already a recorder. What it adds is the other half of the conversation. Something that asks the next question, that hears the offhand sentence and says wait, go back to that, that draws the story out the way a good listener does at a kitchen table. The whole thing rests on that word. Conversation. Not capture, not transcription. And once stored, the ability to ‘use’ these conversations, to generate stories, images, videos… we are now only limited by our imagination.

I want to be careful here, because the temptation to oversell what AI can do is enormous, and the cost of overselling is that people stop believing you. So let me name the part that is broken.

The hardest problem in voice AI is not the listening. It is the interrupting. When two people talk, they read each other’s micro-pauses. The small breath that means I am not done. The falling tone that means your turn. Dozens of times a minute, with a fluency they never had to learn. The machine cannot do this yet. My mother pauses in the middle of a story because the next sentence is the one that matters and she is reaching for it, and the model hears the pause, assumes she is finished, and jumps in. The sentence is gone. The interruption has reminded her she is talking to a machine, and the spell that voice was supposed to create has broken in her hand. The technical name for this is turn detection, and it is, today, unsolved.

But it is clearly solvable. There is no law of physics in the way. Only a hard problem with a clear training signal, and labs making real progress on it every quarter. I have watched these models improve, in real time, for two years. The gap between where they were and where they are is larger than the gap between where they are and where they need to be. The Will Smith spaghetti slop videos from 2023 have rapidly upgraded to Hollywood scenes straight out of a $500M budget blockbuster. The awkward interruptions will go the way of tape hiss, a thing you notice only in old recordings. Not yet. But soon.

I do not think the right response is to wait for the technology to finish. The people whose stories you want are not waiting. My grandmother did not wait. Yours probably will not. The time you have with the people you love is the time you have. The tools are imperfect, and they are good enough, and good enough is what you have.

The notebook on my mother’s shelf is not going to fill itself. It is not going to fill at all. But last week she sat in her kitchen and talked on Ember, for ten minutes, about her mother’s hands, and the street she grew up on in Jeonju, and the smell of the lunchbox her father used to carry. She did not have to find a sentence. She did not have to find a shape. She just talked.

We have her on tape. We have her in writing. We have her in something that will outlast the both of us.

We did not ask her to write.

We asked her to talk.

These moments feel like magic because something impossible suddenly becomes ordinary. A car comes to you. An answer comes to you. A person comes to you.

But there is one consumer experience that is more magical than all of them: family.

Unconditional love. Many of us are lucky to experience it, and nothing can quite match it. Wrapping your arms around your mother as she tells you she is proud of you. The warm spot on your forehead from the good night kiss before you were tucked into bed. Your father cutting fruit for you late at night without saying very much. Your grandmother asking if you have eaten, even when you have already told her three times that you have.

These are small things. They are also the largest things.

For the lucky ones, family is the first magic we know. Before technology, before achievement, before romance, before ambition, there is the strange fact of being loved before you have done anything to deserve it. You arrive in the world, useless and crying, and someone holds you. Someone feeds you. Someone loses sleep because you exist. Whether or not you remember these moments specifically, the residual warmth remains rooted, somewhere, in our bones.

This is why family pulls us together like magnets. We fly across countries for weddings. We sit in traffic for Thanksgiving. We sleep on uncomfortable couches during the holidays. We gather for birthdays, funerals, graduations, births, and the hundred ordinary meals in between. We make time for what is most sacred, most endearing, most magical.

And as we age, the rest of life starts to lose some of its novelty. The late night out becomes less thrilling. The new restaurant becomes just another dinner. The McDonald’s Happy Meal that once felt like treasure becomes a thing you drive past without noticing. The toys change. The games change. The ambitions change. The world keeps offering us newness, and newness keeps expiring.

But family does not expire. If anything, it grows heavier with meaning. The older we get, the more we understand how rare it is to be known over time. To be loved not as a profile, not as a persona, not as the polished version of ourselves we present to the world, but as the whole strange accumulation: the child, the teenager, the failure, the success, the person we were before we knew how to explain ourselves.

And yet, for something so central, family is strangely... underserved.

The modern internet has become very good at helping us perform for people who barely know us. We have built tools to broadcast, to follow, to like, to react, to comment, to message, to post. We can keep up with the vacation photos of a person we met once at a dinner three years ago. We can watch a stranger’s morning routine, know what a celebrity ate for breakfast, read the passing thoughts of investors, athletes, comedians, and people whose names we will forget by tomorrow.

We can do other things too - generate cinematic videos with a tap of a button, book restaurants on the other side of the world, buy earphones from a vending machine. But the people closest to us often remain the least documented. We do not know what our father ate for lunch when he was twelve. We do not know what our mother believed about life when she was our age. We have infinite content, but very little inheritance.

This is the paradox of our times. Everything is recorded, and yet the most important things still disappear.

Your grandmother’s laugh. Your father’s childhood stories. The way your mother describes the first home she ever loved. The family phrases that do not translate cleanly into English. The recipes nobody wrote down because everyone assumed someone else knew them.

These things are fragile because they do not feel fragile in the moment. They feel ordinary. A phone call with your mom. A Sunday dinner. A drive with your dad. You assume these moments will keep happening because they always have.

Then one day they do not.

That is the quiet heartbreak of family: we often realize its value most clearly at the exact moment we cannot recover it. We assume that because we have spent thousands of hours with someone, we know them. But often we know only the roles they played for us: Mom, Dad, Grandmother, Uncle, Older Sister. Before they were these things, they were themselves. A family is not just a set of relationships. It is a living archive. And most of that archive is stored in people.

This is why we are building for the family: to help preserve the private inheritance that disappears too easily. The stories, voices, rituals, jokes, regrets, recipes, beliefs, and small human details that make a family a family.

Not louder technology. Not another network for attention. Something quieter. Something that helps families ask, remember, reflect, and hold on.

Because the magic of family is not only that we are loved. It is that, even after a lifetime together, there is still more to discover.

A father’s childhood lunch. A mother’s first apartment. A grandmother’s secret ambition. The version of them that existed before they became ours.

These are not small things. They are the roots.

That is the opportunity: to build for the people we return to, for the stories that made us, for the voices we will one day wish we could hear again. To build for what is most sacred, most enduring, most magical.

In August 2024, Ted Chiang published an essay in The New Yorker titled Why A.I. Isn’t Going to Make Art. His argument, compressed: art is the residue of thousands of small intentional choices, and a generative model does not make choices in any sense that matters. A prompt is not a decision. The model’s output is an average. Therefore what comes out the other end is, at best, a pleasant shape - and at worst, a kind of confidence trick performed on people who have never seen the real thing.

Chiang is a careful writer and a careful thinker, and I want to say what is right about the essay before I say what is wrong. He is right that most of what generative AI currently produces is not art. He is right that a one-sentence prompt producing a cinematic image is doing almost none of the work that making a cinematic image used to require, and that the resulting output carries almost none of the meaning the old work carried. He is right that we are about to be buried in slop, and that most of that slop will be mistaken, by someone, somewhere, for the thing it is imitating. I do not dispute this.

What I think Chiang gets wrong is the category. He has written an essay about generative AI as if it were a new kind of artist. It is not. It is a new kind of instrument. And the question of whether an instrument can make art is a question that has been answered, repeatedly, in every direction, for at least two hundred years.

When the camera was invented, serious people argued that photography could not be art. The argument was structurally identical to Chiang’s. The photographer, they said, does not make the image - the light does. The choices involved are trivial: point, expose, develop. There is no intentionality in the medium itself. What comes out is a mechanical average of what was in front of the lens. Therefore photography is a craft at best, and at worst a parlor trick.

This turned out to be wrong, and the reason it was wrong is worth unpacking. It was not wrong because photographers eventually proved that they, too, made thousands of small choices (although they did). It was not wrong because photographers demonstrated their interpretation of a ‘moment’ was more than just a lazy click of a button (although they did). It was wrong because the category art was never about how many choices the maker made. It was about whether the work carried meaning that a human being had put there.

A photograph of your mother at the kitchen table, taken by you, in the last year of her life, is art. It is art regardless of whether you adjusted the aperture. It is art because the frame is carrying freight that you, specifically, loaded into it. The camera is not the artist. You are.

Meanwhile, if you are parking your car in a 5-story garage, and hastily take a picture of the C5 column to remember your spot, this is not art. Not all photography is art. The intention matters. The input matters. The process matters.

The same logic applies, exactly, to generative models. The question is never can this model make art. The question is always what is the person on the other end of the prompt doing. If the person is typing cyberpunk samurai in the style of Moebius, highly detailed, trending on ArtStation, the answer is: nothing. They are gambling. The output will be slop because the input was slop. Chiang is describing this case, and he is correct about it.

But this is not the only case. And more importantly, it is not the interesting one.

Consider a different input. A woman in her seventies sits down and talks for an hour about a summer in Calabria in 1962 - her grandfather’s lemon tree, the heat in the stone kitchen, the cousin she was a little bit in love with and never told. She is not a writer. She has never written anything longer than a birthday card. But every sentence she speaks is loaded with specific, non-fungible detail: the blue of the gas flame, the shape of the bruise on the lemon, the exact words her grandfather used when he was annoyed with her.

A generative model turns that hour of speech into six pages of prose. The prose has structure, rhythm, and restraint. It keeps her voice. It chooses, from the mass of what she said, the three images that carry the most weight. It finds the ending.

Is this art? I believe the answer is: it depends on what was put in. If the woman was real, and the lemon tree was real, and the cousin was real, then yes. Obviously yes. The meaning in the work is the meaning she brought to it. The content carried emotion. It carried truth. It carried a lifetime of interpretation, reflection, remorse, joy. The model is doing what the camera did for the photograph of your mother at the kitchen table - it is the instrument that lets a specific human being make a specific human thing, using a medium they could not otherwise have accessed.

What has changed is not whether art is possible. What has changed is who gets to make it. For most of history, producing literary prose required years of training that almost nobody has time to undertake. The woman with the lemon tree story was never going to write a memoir. Now she can have one. The bottleneck was never her experience. It was the instrument.

Chiang’s essay treats AI as a replacement for the artist. I think the more accurate frame is that AI is, for the first time in human history, an instrument general enough to be wielded by people who were locked out of the old instruments. This will produce, yes, an ocean of slop. It will also produce, in the hands of people with something real to say, a category of work that did not previously exist. And this is a beautiful, precious thing.

The right question is not whether AI will make art. The right question is what we feed it, and whether we have the discipline to feed it something true.

I knew the shape of her. I knew the wrinkle of her hands. I knew the twang of her South Jeolla Korean accent, which tickled me to my bone. I knew she loved me. These are not small things. But they are not, as it turns out, the things I wish I had asked.

I wish I had asked what her first memory was. I wish I had asked what her mom was like. I wish I had asked what it was like to be a young woman under Japanese occupation, finding moments of happiness in a country about to break.

I wish I had asked what she thought of the internet, the first time she sent an email. Or even what she thought of the phone, the first time she heard her son’s voice from Pennsylvania, on the other side of the world, learning a language she never understood. I wish I had asked whether she was afraid of dying, and if so, of what - the pain, the forgetting, the leaving, the unknown.

I didn’t ask any of it. I thought I had time, and then I didn’t. That is the shape of the thing, and I know I’m not alone.

Call this the grandmother problem, though it isn’t really about grandmothers. It is about the particular category of person who is so present in our lives that we mistake proximity for knowledge. We assume that because we have been in the room with someone for thousands of hours, we know their stories. We don’t. We have the surface: the gestures, the dishes, the phrases they repeated. The interior is a country we were always going to visit later.

A sentence I keep coming back to is one I wrote a while ago, in an essay about lost memory: the things we know about them, we know well - a feeling, a gesture, a taste. But what about everything we never thought to ask? There is a grief that comes not from losing someone but from discovering, after they are gone, how much of them was never yours to begin with. You find out you loved a sketch. The painting was always in the other room.

Part of what makes this hard is that the interview never feels natural in advance. You cannot sit your grandmother down across a kitchen table and say, tell me what it was like when your father died. The question sounds like homework. It sounds like a podcast. It sounds, above all, like you have decided she is about to become a memory, and this is the opposite of what you want her to feel on a Tuesday afternoon in October. So you don’t ask, and you tell yourself you will ask on the next visit, and on the next visit you hold her wrinkled hands and talk about all your little problems, because that is what you do in a life that is still happening.

The other part of what makes it hard is the asymmetry of memory. Our elders usually want to tell us. The generations above us are often starving to be asked - not because they are vain, but because they have lived through things that almost no one asks them about anymore, and the weight of an unshared life is real. My grandmother once, late at night, started telling me a story about walking to school during the war and having to hide in a ditch, and I remember being seventeen and not knowing what to do with it, and changing the subject.

Now, I think about that ditch a lot.

The reason we started Ember is, more or less, this essay. We wanted to build a thing that makes the interview natural - that handles the awkwardness of tell me what it was like by putting it in a form nobody has to sit across a table for. We wanted to build a thing that turns an hour of talking into something a granddaughter in 2070 could see, hear, read, and hold. We wanted to make it stupidly easy to ask, and stupidly easy to answer, because the cost of not asking is too high.

But even if Ember did not exist, we would want to say this: go ask. Not the big existential questions, not yet. Ask your mother what her first apartment smelled like. Ask your father what he ate for lunch when he was twelve. Ask your grandfather which of his friends he misses the most, and wait through the silence for the answer. You will be surprised how quickly the small questions open the large ones, and you will be more surprised how much they want to be asked.

There is a version of this essay that ends with a line about how it is never too late. That is not true. Sometimes it is too late. My grandmother is in the ground in Jeonju, and I will not know what she wanted to be when she was young, and that is a fact I will carry around for the rest of my life.

What I can tell you is that it was too late for me and it is not yet too late for you. The people you love are, right now, sitting somewhere, doing something ordinary. They know things you will spend the rest of your life wishing you had asked.

Ask.

A bit about us

ingwon - i’ve had a lot of different career experiences. i started at Morgan Stanley as a bond trader, then worked for a few tech companies, mostly in the security/cybersecurity space, founded a labor marketplace called Glencoco, and now focusing on working on this more personal passion project.

daegwon - he’s shy, partly (read: mostly) for legal reasons around posting publicly because of his work, but high level he works as a venture capitalist.

What we both have in common is a deep, deep love for great writing, and for our family. Which… is probably a good segue to why we decided to start writing on substack. It starts with our dad.

The boiled egg that sparked our interest

Our dad turned 71 this year. Last Christmas, the three of us were sitting around after dinner, and he casually mentioned an anecdote about growing up in Korea in the 1950s. For context, in 1955, the GDP per capita of South Korea was $64. Literally one of the poorest countries in the world. Every day for lunch, he told us, his mom would pack him just two things - rice and kimchi - which he’d take in a lunchbox to school. But on his birthday, his mom would also add in a boiled egg.

His birthday present every year was a single boiled egg.

This kind of blew our minds. Obviously it’s so alien to how my brother and I grew up that it reads like a different world, despite being just one generation apart. But mostly, it was how matter-of-factly he mentioned it.

He didn’t say it like it was some tragic pity-inducing story. To him this was just a casual but cherished memory. It told us more about our dad than the ‘bigger’ things we knew - his education, his career path.

So we started wondering, how many more of these small details does our old man have in his head? And so we decided it’d be a fun project to collect these small memories and details from our parents when we could, and write family memoirs about them.

Wanting to collect stories but being too busy

It was a rough going at first. We tried to start with collecting voice memos/interviews - the obvious thing. The obvious thing was not easy. Our dad lives in Korea and we live in the US. Juggling zoom sessions across time zones with a voice recorder, not knowing what to ask or what to say, being ill-prepared for these sessions, and honestly, the awkwardness of three emotionally constipated men trying to get sentimental. Disaster.

Somewhere around this time, voice AI started getting better by a lot. Since we both work in tech/invest in tech, we’ll occasionally nerd out over research pieces and betas, and Sesame’s research preview blew us away. We started tinkering with some of them.

What if the conversation could be handled by something that wasn’t us getting on a zoom call? We’d love to schedule them regularly, but our schedules made it really hard to do on a consistent basis. What if awkward, difficult-to-schedule zoom conversations could be instead recorded, daily, with the most attentive interviewer in the world?

We wanted to collect and record all of his recollections and put together writing and stories about them. So, taking it a step further, what if these recordings could magically turn into prose, memoirs, and poems? What if…

Either we suck as interviewers, or there’s a weird niche place for a conversationalist that feels private

But we had a problem. My dad is pretty anti-tech. He refuses to use banking apps because he’s convinced his money will disappear. So we ran a test. We hopped on a zoom, told him he was about to speak with Ember - the name we gave to our interviewer app we built. He introduced himself. We held our breath.

He liked it. He spoke for 20 minutes. We were floored and a honestly a little butthurt, because he gave Ember more details then he had done with us…

After about a week, we realized he was talking to Ember daily.

More on why we think this has been the case (his preferring to talk to Ember over us) in a later post, but I think a big motivating factor is just that he wants to get his story out there.

So we kept building. We built a way to quickly start a conversation. To save memories. To explore different topics. To listen to the recordings. To make written stories out of them. Our dad uses it every day now. And the stories we have of him are now some of the most valuable things we own.

Seeing who else out there might find this useful

Two weeks ago we decided to share the thing we’d built for our family with other people. It’s available to explore at ember.build.

This blog is partly a space for my brother and me, like many of you, to wonder out loud about what we’re learning as we explore: voice as a technological interface, about what AI does and doesn’t do well, about memory and craft and families.

We also have really mixed thoughts about the intersection of AI and creative writing, so we hope to unpack that as we continue along this journey.

From our early experience with our dad and ember, it’s pretty clear that it’s adding value (recording his memories and childhood stories) in ways that otherwise just wouldn’t happen (we’d be too busy and coordination too difficult), but on the other hand, we’ve firsthand experienced and reflected on a lot of negative things (these will all be future posts!):

• AI/chatgpt psychosis for some of our friends

• AI slop

• AI slop

• AI slop

There is a lot of really thoughtful writing and discourse right now about Why AI isn’t going to make Art, why AI can’t write well, and if anyone builds it, everyone dies (we loved this book btw).

We will also in future posts unpack a lot of our thoughts about this, but high level, a lot of it is missing the forest for the trees. Like all technology, AI will lead to slop, anxiety-laden societal land mines. But we’re optimists, and we are of the camp that it can, should, and will be harnessed to bring the best out of us. To do things we’ve never done before. And to create beautiful, beautiful works…

The reason we’re sharing any of this is simple. These memories - through recordings, written stories, even poems that Ember shaped - are now our family treasures. If just one other family gets to experience this too, it will have been worth it.

If any of this resonates, feel free to follow us, or subscribe to us, or whatever you’re supposed to do on substack.

Subscribe now

Also, DM us and we’ll comp you an Ember account in exchange for your thoughts - positive or negative. We’d rather have ten curious readers who push back than a thousand who don’t.

- Ingwon