I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously.
The firt citation for marriage data is listed as "Hajnal (RH31)", but links to a seemingly unrelated WorldCat search. The second citation is to "Kaplan (RH03)", which I was able to track down on sci-hub. It is a broad theory of human evolution that doesn't appear to mention the Yangtze Basin or China.
Another citation is to "[RH109] Model-supplied gap-fill (Claude Fable 5 and Claude Opus 5, 2026-08-05). Bounds and items written to close gaps no dataset we hold covers. Not a published work: see each row's `basis` for its stated reason. (2026)" -- this is an interesting way to describe having an LLM hallucinate something for you.
Please stop making vibe-coded websites -- it is not useful to provide incorrect facts. You are polluting the commons with garbage.
James was born on the British Isles, near Boulogne-sur-Mer, in 2009 CE.
James's table is mostly oat and barley bread, pork and dairy, and cabbage and peas, with apples in season, and coffee and sugar as well. He is literate.
Odd for a gen-z person in Britain. I doubt he’d eat a lot of cabbage, and I have no idea when the “season” for apples is as they are always in the supermarket and there aren’t any orchards near where I live.
him eating apples when in season seemed to me at first like an anachronism, but actually, i do that. years ago i got fed up with apples that were whatever, so i decided to only have apples when i could find mcintosh (my favorite apple), and eat loads of them.
Welcome to the future of widespread information, where the uneducated masses don't understand what context, or the definition of words for that matter, even is.
Youtube is full of this, history slop. Some techbro hyperfixated on this and LLMs gives it life before he's bored and moved on to the next thing to hyperfixate on.
It's techbro ejaculate. 10 minutes of euphoria for slop to come out in the end to spread to unsuspecting victims.
The first part likely is just an artifact of an unwritten common assumption. That is, it's probably 96% of women, who lived to adulthood, who were ever married.
Maybe, but I doubt it. The vast majority of the sources, including the two listed for this claim, are listed as "attributed", meaning "A model (Opus or Fable) claims that this is what the cited source says. In my spot checks, I've found these claims to be accurate; still, since it's just a model claiming that author x says y, there's a nonzero chance that it's a hallucination."[1]
I couldn't track down the first source, but the second source goes over something completely different. Maybe Claude accurately remembered the first source when it was stolen during the training process, but forgive me if I'm skeptical. Feel free to track it down yourself and let me know. Otherwise, I will assume it is an LLM hallucination.
(Also, given my background knowledge of the Tang dynasty, I would be surprised if the average marriage age was as high as 18!)
I don't follow what your point is here. If you're not aware, Anthropic pirated many books via libgen and similar tools for use during the training process[1]. Maybe you are not aware of this, or do not consider it theft?
Key distinction though: Anthropic can be said to have "stolen" or pirated those books when they downloaded them illegally, but not when they trained Claude, because that same lawsuit found that training is fair use.
Anthropomorphizing makes for poor analogies, but it's kinda the same way your acquisition of background knowledge of the Tang dynasty from whatever source material you used is not considered stealing or pirating either.
Did the original owners of the books no longer have them in their possession afterwards?
Theft involves removing from the source, Copyright violation is distributing copies without a legal right to do so. Training is fair use.
The judge found that it constituted copyright violation. Purchasing the books would cover their responsibilities with regards to training. The claim that they are destroying books to hold copyright is incorrect. The judge cited precedent that said instances of not destroying the originals were also fair use, his claim about destruction was to say even if this precedent were not considered fair use, the destruction that was already occurring (because it is easier to scan loose pages), invalidated one of the arguments against it.
The ruling was on training took the form
If A is ok then B is ok, but even if A were not ok, B might still be ok.
Excellent. I will thus take a copy of the latest Disney movie and train myself on it while claiming fair use as a defence to the accusation of copyright infringement.
You can do that legally as long as you do not confuse people that what you create might be an actual Disney product, or that you try to represent it as actually created by Disney.
Both of those are protected by law. Learning from a Disney movie and creating your own thing is not, and should not be, disallowed.
Even if this wasn't vibe coded it would be impossible to fact check billions of possible lives. I think thought provoking toys like this are a great use of AI, assuming no one takes them seriously.
you can't fact check real individual lives but you can draw up a probabilistic representation
that is if your probabilities are actually correct, that the anthropological or archaeological or sociological field for that specific region and demographic is robust (and there is good primary evidence for supporting the claims), and you're not some vibe-coded AI tool that spits out a convincing lie constructed on not even educated guesses, casting shadows on the wall of the Platonic cave where the people who trust slop uncritically are chained to their ignorance for perpetuity
I am genuinely puzzled as to why this kind of vibe-coded site are so popular on HN.
I consider the crowd here to be much smarter than I am, which is why I enjoyed the links that were shared and the discussions that followed.
The website is pretty, but the statistics and stories are the main point. Yet we are being fed hallucinated false claims, and somehow, that doesn't seem to be a problem...
I don't understand why it gets so many upvotes. We know Claude can create websites like this brilliantly now. The novel and interesting part is the transformation of the data. But if that part is sloppy, what's the point?
Until I read the comments here, I did not notice the stats were bad. I was just skimming the facts I thought were interesting, not looking for contradictions.
The site itself is a really cool concept! I just wish the information was accurate. Now that I know it's not, I feel cheated.
I suspect a lot of people don't read the comments.
Well for one example, if you select a person born on the Italian peninsula in the 1400s and they die of bubonic plague, their year of death is not correlated to years when there were large bubonic plague outbreaks on the Italian peninsula.
This is not an error I would expect a human to make if they were interested enough in demography to make a website about it. For example, the wikipedia page on this topic[1] does a reasonable job of distinguishing that plagues did not occur every year and that they varied in lethality.
This is such a weird nitpick, it's clear the website deals in averages and assumptions. It's meant to be an interesting window into possible lives intermixed with historical happenings at, or around, those times.
I get that some people just hate AI, but of all things, this is a pretty fun & educational use of it.
I think our core disagreement is that I do not consider teaching falsehoods to be educational. My antipathy towards LLM outputs that haven't been fact-checked is mostly downstream of things like that.
This website doesn't "teach" anything. It's a fun way to explore and get exposed to history. Kind of like Civilization 6 or something. Do you think CIV6 "teaches falsehoods" because you can build the Pyramids as the French?
I haven't played Civilization. Generally I would not consider something educational if it contains mostly false information. Perhaps clearly demarcated parts of it with true information could be considered educational -- for example, looking at screenshots I don't see anything immediately objectionable in the civiliopedia "historical context" sections.
I think most players would understand that in-game choices will not reflect the choices people actually made historically. Having not played Civilization, I'm not sure whether I would consider it to be educational -- I think it would depend on whether playing the game provides insight into real historical events.
For example, I would consider a sufficiently detailed LARP of a historical event to be educational[1] -- the actual choices and outcomes do not match what happened historically, but they can teach participants a great deal about the motives and limitations of real historical actors.
This website, on the other hand, seems to consist of 90% AI hallucinations and 10% data that was actually extracted from a real source[2]. Since these types of data are thoroughly mixed together, I think this website is about as educational as just making up something to believe about the past.
I don't think a hand-crafted app necessarily estimates things more accurately, but a sensible pre-LLM developer would probably not list so many dubious citations to trick people into believing it's accurate.
> I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously.
Yeah, I got: You have drawn a girl, born in the Yangtze Basin in 1295 CE.
Typical age at marriage: 18
Ever married (of adults): 96%
Children per woman: 6.5
Life expectancy at birth: 28 yrs
Died before age 5: 39%
Died before age 15: 44%
Very similar to your "random" draw, and the facts are suspect. There was a link to the source, but it's behind a paywall.
There are plenty of problems with the site but that some stats like marriage do not count dead newborns as unwed instead of invalid for the stat is not really one of them.
What age of death discounts someone as counting towards the marriage percentage? If a human made the website, there would be an answer to this question. That answer might not be published and might vary between datasets, but it would exist.
In this case there is no answer, because the percentage is just made up by the LLM. There is no methodology. Your answer is inappropriately anthropomorphizing the LLM by imputing a methodology onto its "data collection" process.
No anthropomorphization was required, I just looked at one of the sources on that section of the page and saw the statistic did not involve the number of newborns and so figured myself those facts could indeed be simultaneously true.
How do I know the methods described in the source were correctly used here? I don't, but I also wouldn't know that based on whether or not a human wrote the webapp. I'd know that by checking the method applied on the backend matches that of the source. If it does, it doesn't matter if it was generated by using every 17th letter of the dictionary - it'd either align or not.
Preston, S. H., Heuveline, P., & Guillot, M. (2001). Demography: Measuring and modeling population processes. Blackwell Publishers.
In GP's case I believe they are correct the Hajnal source is not relevant to the Yangtze Basin but reading the first page should have made it clear the facts could be simultaneously true regardless where the percentages came from or if they were accurate.
Again, there are plenty of problems with the site. That it was clear these facts can't all be true simultaneously was just not one of them.
This appears to be a general introductory demography textbook. I am surprised that you looked at it and it contains specific demographic tables about marriage rates at specific time periods? Perhaps I am misunderstanding what you mean by 'I looked at one of the sources and saw it did not involve newborn rates.'
I think it might be a simple misunderstanding, but one can never be certain :).
I believe we agree we haven't seen a valid citation for "the source in which 'x' came from in 'x%'" for this stat on the site. I hope the two of us also agree this is not the point I was disagreeing with.
Where I'm currently lost is why that should mandate a demography textbook, which was a different citation in the same section, must be unable to explain dead newborns need not be counted in such a statistic just because the textbook was not also the source of the specific percentages displayed? That'd be like saying a statistics textbook can't explain how the median can be less than the mean on the basis it wasn't also the source for the numbers displayed.
"I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously." Umm yeah they can be true. Dead people are not counted. The straightforward interpretation is 96% of women living to 18 years or greater marry at some point in life. However the app has bigger problems, namely it does not seem to select from the distribution correctly, pre 1500 lives seem to be vastly over represented.
Yes, if a human told me this collection of facts I would assume they failed to properly explain their methodology. Given that the LLM is just making up numbers (see the "sources" page, where these are listed as "attributed",) there isn't really a methodology to explain in this case.
This is neat, but I don't think it's actually drawing the year from the correct probability distribution - "a random birth is far more likely to fall near the present" is correct, but in 5 "random" choices, only one was from anywhere near modern day.
I did a statistical analysis and you are correct. The issue seems to be that crude birth rates from around 500CE and earlier are overestimated significantly. The bad source seems to be Kaneda & Haub, "How Many People Have Ever Lived on Earth?", where they assume a flat crude birth rate of 80 for all BCE dates. The probability distribution is just a function of pop and CBR, so it results in significantly decreasing modern birth probabilities.
There are more statistical issues. I drew a person from 1984, living in Germany and recalculated their life several times. In almost all cases, their mother died while they were in their thirties, often both parents. This is simply far from reality.
I think it's working as intended. The year is simply a randomly chosen year not related to the density of births in that year. It would be a boring tool if it actually chose "random year based on random birth" because almost every choice would be from 1900 onward...
This is a case where I think they should change the wording as opposed to the functionality. I like that it's a random year regardless of population density.
But as other have noted it's almost certainly not just a random year.
I do think the cleanest and most fascinating interpretation is what the title says "One life, drawn at random from all who have ever lived" which clearly is about random people and has nothing to do with the year.
But I think experientially, the preferred feeling for a user is a random deep dive into the circumstantiality of different eras and different geographies than our own, rather than a true-to-life re-rolling that would constantly turn up 20th and 21st century outcomes. So the spirit of it is probably closer to the random year idea, but it's literally not random year, but maybe randomly something else, like eras however designated.
If the implementation was correct it wouldn’t be hard to let the user choose either option.
Yes, many births would be within the last 200 years but it’s still very different to be a man born in the year 1950 in New York vs a woman in 1950 in rural China.
That would certainly be easier at this point, but I like the “random human life” idea more.
I’ve reflected in the past about how lucky I am to have been born where I was, when I was, and be given the abilities that I have. Sure there are plenty of people that were luckier, but also a lot who were less lucky, most of whom are dead.
But "is far more likely to fall near the present" = "is far more likely to fall near the present". The author makes the claim about probability distribution themselves, then does the opposite.
You're right, yet I agree with the website's approach. The goal of this site is to educate users about how people lived, not educate them about stats. IMO. I would have chosen the same algo.
It's clear to me that an even bigger HC crowd would howl "Oh, how boring, look at him choosing pretty much only still alive people" if he had done the 'uniformly random" selection. C'mon!
The site covers a 300,000 year span, so a randomly chosen year would average ~150,000 BCE, which definitely isn't happening.
The skew towards the present isn't as stark as that. World population has exploded in the past couple of centuries, but it hasn't been high for very long. 1900 to present only covers ~11% of all births.
1900 to 2026 is 127 years (inclusive), so 0.04% of the year range available, yet it only has 11% of all births. That's certainly a skew, although I suppose subjective whether or not that's stark. At a >3x order of magnitude, it's certainly significant though.
> "a random birth is far more likely to fall near the present" is correct,
If the selection is from the population at random (as the UI indicates), but it specifically says it's based on a random year (during which humans were present) first as a filter. It's much more likely to select far from the population explosion.
The x-axis being log scale by default visually exadurates the effect if you are thinking about it in terms of area under the curve. This confused me also.
The reason there are two Gs is that the first one is assimilated from a D. The etymological structure of the word is ex-ad-gero,¹ "from-toward-carry".
This happens for everything. Compare accelerate [ad-celero] / accommodate [ad-con-modus] / impossible [in-pot-s-ibilis] / intelligent [inter-leg-ents].
(Note that a doubled consonant in Latin is pronounced differently from a single one, as is still true today in Italian. English draws no distinction, which makes the etymological spellings likely to confuse English speakers.)
¹ This was already the form of the word in Latin. It looks malformed to me; I can't explain why exaggero is a first conjugation verb (with passive participle exaggerata) rather than simply being a prefixed form of its root (ag)gero, in which case the participle would have been exaggesta. Gero and aggero are both third conjugation verbs...
I guess you are right, but it makes it more interesting if the picks are not almost all from the recent past. On the other hand, I would also guess that spreading the picks evenly across the time range might result in many being from times with little data to differentiate them.
It makes the premise of the entire site flawed, though. And it defeats the purpose of the convenient button provided to redraw a year. That is what should allow you to find a pick that might be more interesting, rather than messing up the randomness promised by the site and the HN title.
How near is near? With ~100 billion people having ever lived, the odds of a randomly selected life falling within the past couple of centuries are not all that high.
Could you be a little more specific? https://www.prb.org/news/how-many-people-have-ever-lived-on-... puts the halfway point at over 2,000 years ago, so you'd expect less than 50% probability of getting a random year that's not BCE. "A random birth is far more likely to fall near the present" is true in relative terms; the timeline on the web site goes back 300,000 years and you're very likely to get something in the past few thousand years, which is pretty near to the present when compared to a third of a million years. The random years it gives me fit into that pretty well, so I guessed that the person I replied to might be using a narrower definition of "anywhere near modern day" than that. Hence why I asked. If I missed something, could you kindly point out what I missed, instead of whatever this is?
The stat I keep in my back pocket is that you can roughly divide the total time lived by humans (the integral of population over time) into three tranches, with one third of human experience being before the dawn of agriculture, and one third being after 1600. (Related fun fact: WWII was about 1% of human experience, based on the back of the envelope that ~50% of the world experienced "being at war" for nearly a decade.)
The distribution of human lives versus years lived by humans will skew earlier, because infant mortality was absurd for the first two tranches, with something like 50% of people dying before the age of five, so it will be more like 40-40-20 over those time periods.
I drew Vuk,a boy who lived in the 13th century Balkans. He died at 2. His dad was already gone. According to the statistics the life expectancy at birth for him was around 50%. I suspect a lot of our random people would have died early, if they were born before the 19th or 20th century...
Whole years! You'll are so lucky! The reincarnation roulette gave me this: 49 BCE, he lived 1 month, dying of dysentery. I guess I'll stop complaining about my life then...
I had a similar sensation. I often think about a friend of mine who mentioned, in passing, a sense of having won the lottery of life for being born in this day and age, in a first-world country
last night my wife and I were watching the first episode of a TV show Deadwood (set in the Dakotas in late 1800s), and my wife commented "I could never have lived then; I would have just killed myself immediately". The irony of course is that she would have been fine had she been born in that era as it would have appeared as "normal" to her then as the world today appears to us (ok things aren't normal right now, but I mean comparatively speaking), and "better" than 100 years earlier.
This is at both times super cool as a project at what AI can do, while also being extremely dubious and misleading about how accurate any of this data is, hallucinated or not.
Her farthest journey was a day's walk, to a nearby market town.
I think this is what boggles my mind most about antiquity: the vast majority of people rarely ventured much past where they were born (unless they did a likely once in a lifetime religious pilgrimage which took the better part of a year).
Seeing as most other people didn't venture further either, you really got to know the people around you very well. Common life limitations simply precluded much travel. The only foreigners you saw were either invading hordes, or traders who were constantly on the road.
My grandmother told a story about meeting an elderly female relative in Kingston, Tasmania, soon after my grandparents had got their first car. After visiting, they said they were now going back to the city (Hobart), and the old lady said, "Oh, how lovely, I've never been!" My grandmother reported, "And I should have told her 'get in the car then!', because she died not too long after."
There were and have always been trading and other exchanges all over the world. The Egtved girl was traveling from north to central Germany, probably for marriage. Now, if we mean most people did not do this, then sure. But there's a persistent idea that ancient cultures were not in contact with each other, and this is largely disproven.
We might recognize that the economics of making such a fascinating and artful experiment like this have completely changed. No ads. Not asking for anything except your attention, and then delivering an outcome that is fascinating and honestly, far beyond what a person would be incentivized to build before AI tooling.
I loved it and feel inspired from it.
(It looks like it's getting a lot of traffic and might get hugged to death)
> No ads. Not asking for anything except your attention, and then delivering an outcome that is fascinating and honestly, far beyond what a person would be incentivized to build before AI tooling.
I don't think I agree. There were plenty of interesting side-projects on the internet before AI.
I'm not sure why AI means that people don't need to support their creations with ads. Sure, it probably took fewer human-hours typing at a keyboard to create this, but for things like this, a labor of love, that's typically not the "expensive" part.
Yes indeed. What do we know about how this was created and how long it took? Before I look into it, I’m going to guess:
One experienced developer with extensive anthropology experience, probably a PhD, working with Claude Code and Claude Cowork, full time over a two month period. Or somebody with limited dev experience over four months.
Anyway, sad that all the first comments were complaints… what a wonderful project!
I think it's just that in the past, you would need X free time, Y motivation level, and Z skill level (partly tradeable for X and Y) to create something like this, and now all of those values are smaller. So it's 'economics' in the straightforward sense, because not everyone could afford to spend X hours on something they weren't going to monetize, and also in the sense of the incentives shaping people's productive/creative activity.
(As for whether it's a good thing, I'm conflicted. It's cool to see more stuff like this, but if it crowds out the more intense passion projects, we lose something too.)
I believe that part serves to emphasize the main point.
To elucidate for this particular example, before AI agents existed, projects with this level of data and depth might've needed to be monetized in order to justify the resources to bring it to life.
Now that the barrier is much lower, the creator didn't mind spinning up and providing it with no monetization (e.g. no ads).
> before AI agents existed, projects with this level of data and depth might've needed to be monetized in order to justify the resources to bring it to life.
The site was registered a month ago. If it ever actually gets and stays popular, see if it has no ads then. The economics of personal interest sites didn't change in the sense that they're any harder to make now compared to the mid-90s. The economics that changed is the Internet now has billions of users and attention being monetized can make you a lot richer now than then. It's very hard for people to see an opportunity to become a multi-millionaire by clicking a button but then not do it because they'd rather maintain a slightly better experience for their audience.
It's, of course, not even loading right now, which gets at the other problem with this "economics" framing. The biggest expense in running an actually popular website isn't creating and publishing the first version. It's paying for the infrastructure to serve a meaningful proportion of the entire world. This is clearly not a static site. It's presumably doing some computation on the server side to draw from all these hundreds of different probability distributions that works well enough for a few dozen simultaneous requests but not thousands. There might even be some kind of perverse anti-economics going on here where thought devoted upfront to efficiency and architecture might have made this far more scalable, but the "publish the first thing that passes a basic sanity check" vibe-slop ethos doesn't usually get you that.
Vibe slop? The structure of the site is quite solid, the UX decisions are spot on, the data looks good to me that knows little about this stuff, and it is entertaining and informative. This was not a one shot, a human out a lot of effort into this.
Yeah it is imperfect and was maybe not ready for the HN front page hug. But I say, bravo to the person or team!
I also think this shows what the changed economics do: it allows anyone to fill the web with convincing looking nonsense.
The person who built this did not have to put in months to build a cool web app. But they also did not have to put in months of research, gathering representative pictures. And they didn't. Instead they let a software create a well presented nonsense story about a boy from Anatolia in 200 BCE called Mehmet wearing a parka.
> We might recognize that the economics of making such a fascinating and artful experiment like this have completely changed.
And this is not limited to software development. In just a few years, the economics of delivering healthcare, education, entertainment and legal services are going to change completely. Transportation is changing as we speak. In a decade Physical AI will completely change the economics of manufacturing, food services, elder care, and so on.
The post-scarcity world will be here sooner than we think!
In 10 years? I predict that healthcare, education, legal services and transporation won't change in a meaningful way at all. People will go to a doctor, kids will attend school and aim to graduate, legal services will still be done by lawyers and decided in court. Transportation will still depend largely on what exists right now or is being built right now. There won't be self-driving cars in every city, only in a few, probably more in the US and China, fewer in Europe, probably close to zero in the rest of the world.
And how do you think will the post-scarcity affect the daily lives of most people? My prediction is that most people even in 30 years will still work for a wage, and probably still around 30-40 hours per week, because these gains of post-scarcity won't be distributed evenly. It will mostly make a few people very very rich.
> In 10 years? I predict that healthcare, education, legal services and transportation won't change in a meaningful way at all.
"We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten." -- Bill Gates, in The Road Ahead.
I’d guess that education will hardly be recognizable in 10 years. Bad news for number of teachers employed, but AI is going to result in a huge increase in the number of 15-year-olds with a PhD’s level of knowledge and abilities in their heads.
We are not in a scarcity world. Or it's deliberated scarcity. We have way enough means to educate, heal, feed and transport everyone. Yet most of the world work full time to produce more and more of meaningless products, and to make the economic curves go up. Also, a huge part of the population is still poor, while the rich people get richer. And on us dooms the perspective of an extreme climate crisis.
I don't know how you can think that AI or "Physical AI" will solve any of that without a really huge shift in our locals and worldwide politics. It will probably cause a massive economical crisis before anything else of importance.
Both of these things can be true. It's a really good idea that immediately led me to another idea to play with. But it hallucinates so its accuracy can't be trusted.
This is brilliant. I think it could stand to be a little more imaginative in their stories. It seems like it should be possible to do that while still conveying an accurate picture. "He made his living as a professional" is simply a wasted opportunity. There shouldn't be anything so vague. If you we're listening to a story, you would feel cheated. Embrace your idea.
Shameless plug, but related, is the Watsi platform[1] I discovered a few months ago through another HN user. You can read the story of less fortunate but alive people and how much need there is in the world for one another. The founder's talk[2], Chase Adam, which made me a supporter, is a must-see. I guess it's time to rewatch it..
I am really happy to live in the modern age. It is quite amazing to think that people are facing cardiovascular disease and cancer instead of random parasites nowadays. Modern medicine is such a miracle.
I was thinking along the same lines as I was going person to person in the app. It made me more disgusted at MAGA and MAHA. These guys are living in modern times with access to things that people back then couldn't even possibly dream of. The thing that really drove it home is the life expectancy of the people that came up in the draw. Most died by their 30s and some even before their 30s, most never knew their parents as they died when people were single digits of age, and they were forced to live very simple life with simple foods and conveniences.
So much of what we have has been brought to us thanks to ubiquitous fuel, ubiquitous technology, and institutions that took centuries to form and solidify. All of this built on the backs of people who sacrifiec and sometimes did horrendus things to allow us to be in the places we are. There's just so much hubris and misunderstanding of what they have been given in their lives that they don't appreciate. Pissing all this away is fine to them. Yes, I get it...they didn't get the same quality of life as their parents did(who came of age in an abberation).
What they don't realize is that it could get much worse. Many of them may not even live to see how bad it could get. This is so frustrating, and it just increases my disrespect for them.
How many random parasites has modern medicine removed from you? Likely you never even had one. It's not just medicine, we also have abundant, safe food, clean water and warm housing
Yes! I think it’s really easy for us to take for granted the long list of incredible things we take for granted. Every time I hear someone lament how horrible the world is today, I wonder when they’d rather have been born. Especially as a random human like this.
"You have died of Dysentery [on the Korean Peninsula in 3064 BCE at 2 months old]"
I was actually a bit surprised since it also said I was exclusively breastfed which seems to be fairly protective against diarrhea. I have a feeling this didn't account for all the joint probability distributions...
Yeah, it doesn't give totally absurd combinations (like a male dying of birth complications at age 10), but the probabilities are still likely rather rough approximations (eg. sorting cause-of-death distributions to a few bins like "infant", "child", "adult", "elder").
That is surprisingly similar given how few people seem to have commented so far. Makes me wonder if some part of this is starting from a constant seed or something...
I do wonder if that life expectancy is dragged significantly down by all the ways you can die early on in life, so that it's not so much that 28 years is an old age for the time as it is a function of all things that prevent you from getting to old age.
And so if you went to a town square or market in that time (if it's a time that had such a thing), perhaps it might still be perfectly normal to see people in their 30s, 40s, 50s.
The life expectancy of a person who lives to 18 increases dramatically across the board.
You even see this today. If you live to 70, you can expect to live 14 to 17 more years. So to about 84 on the short side and 87 on the long side.
However, make it to 75, your life expectancy is 11 to 14 years. Which is 86 on the short side and 89 on the long side.
Basically, the longer you live, the longer you should live, but eventually the range collapses. And that's because people come in under the average all the time. Some people die at 71. And those people stop getting counted when determining life expectancy at 75.
Right, that 42% dying before age 5 means the survivors typically lived well beyond 28. There weren’t a huge number of people living into their 80s, but 28 wouldn’t have been “old man” territory.
I think the model has an unusual preference for the Korean Peninsula before Christ. A number of commenters, including myself, draw from this area and period.
The life template kinda breaks down if you land in a modern era which makes me think it's probably equally as silly sounding to apply it to a hypothetical life from 10k years ago.
But I guess it's inevitable when you try to paint individual lives through pure statistics.
Seojun -
Seojun is not a real person. But his life story is created from the statistical reality of the Korean Peninsula in 1999 CE.
In 1999 CE, Seojun was born in the countryside near Pongch’ŏn-ŭp. His village is home to about 250 people. Among four children, he was the fourth. He eats largely rice, soybean paste and tofu, dried fish and shellfish, and pickled cabbage and radish, with persimmon and jujube in season, and tea as well. He knows how to read and write. Seojun stands about 167 cm tall (5 ft 6 in).
---
if it's not obvious why this is problematic and extremely offensive I don't know what I can say to you. You're talking about GenZ who has access to tiktok and food delivery apps as if you're talking about someone from Joseon dynasty from hundreds of years ago. No one's favorite food is rice and soybean paste. He knows how to read and write and sky is blue.
Now this is fun vibe coding. A fascinating way of reckoning with our cosmic luck. On "only the first roll counts" rules, I rolled a girl who died at 8 months from a respiratory infection in the Indus valley, in 7785 BCE (9,810 years ago).
I rolled again, of course, but at that point it feels like fishing for a best case scenario. But I at least got someone from the North China Plain in 75 BCE who lived to 74 and had 6 grand children. Married at 19 and stayed married to age 71. Hard to hope for better than that, against these odds at least.
Edit: Obviously progress isn't inevitable, and at any point in time you can be unlucky. But it does seem like, for most points in history, you would rather be you, in your version of the present, than a randomly rolled person from any prior point of history leading up to your moment.
So most of the time, anyone playing this would feel more lucky to be themselves than any of the options they would have to roll for. Maybe if a version of this was offered to people 10,000 years from now they would say "crap, I rolled 2026".
I was so lucky! One lived to 35 in the Paris basin a few thousand years ago, the other lived to 25 in China, died in 1945. Both were married and left living children!
I think it would be more interesting if you entered your own real data and then the LLM gave you statistics based on that info on where you are on your own life path based on consensus data.
"He died of dysentery in the first days of life. His body was laid in a stone cist or a great urn beneath a ring of boulders, with iron tools and black-and-red pots."
The data will be much clearer in the future, once all of our data has been mined. We create so much personal media and collected data that some point in the future when one entity has loaded that, a crazy real autobiography could be created.
I think graph being log makes it a little hard to understand that the median birth year was around 2000 years ago. On the graph that is in the flat part but the flat part was a really, really long time.
I will say this with all honesty and sincerity to everyone who interacted with this site: If you didn't realize this was just LLM-generated nonsense within the first 45-60 seconds of using it, it's completely and unequivocally over for you. Your lack of discernment will lead to you being jerked around by AI-generated slop for the rest of your life.
After having my first child (a girl who is now nearly 3 years old) these stats impact me so much more. My first roll was a girl in the Iraniam plateau in the 18 century who had a 33% chance of dying before they reached 5. She managed to live to 70 but experience so much death in her life.
I can't imagine my daughter dying before she reaches 5, it feels like it would be the worst feeling I could possibly imagine and then some.
Visit any older cemetery and look for the tiny headstones and the matching very brief lifespans, or read an older obituary for someone from your great-grandparent's generation or further back, or watch a couple of seasons of _Call the Midwife_.
This reminded me about the US nutters advocating return to the "way things were": they would do well to consider life expectancy. My draw from 1881 Mississippi would max out at 36.
I think this would be cool if it actually worked as intended. This simulation doesn’t even draw birth years correctly.
The right way would be to bucket people by something like 1 million. Everyone in the bucket has the same sex, birth year range, and location range. Then draw a bucket uniformly and infer the details.
It's funny to pick a birth year from our era and think of how history will represent us:
Martha was born in the countryside near Short Pump, in 1993 CE. Her village holds about 300 people. She makes her home in a timber-framed detached house. Martha has neither brothers nor sisters. She eats mostly wheat bread and maize, pork and dairy, and potatoes and garden vegetables, with apples in season, and coffee, soft drinks and snack foods too. She knows her letters. Her height is about 171 cm (5 ft 7 in).
An “Eastern Woodlands” person born 1989 is recorded to have experienced a “volcanic winter” and “dimmed sky” at age 2 as a result of the eruption of Mount Pinatubo.
This is an… interesting contextualization.
I am wondering how many historical situations are sexed up to be something significant, yet were nothing to write home about in the moment. All it took was a historian with an active imagination to bring it to the forefront.
Given the huge infant mortality rates until very recently in human history, this is often pretty misleading. Life expectancy at say, 2yo, is far greater.
I was wondering if the generated people “fill up a spot” and temporarily change the stats for generating new lives. Eg, if the app already generated 10M stories, is it still using the same initial distribution for the new ones, or is generation conditioned on the stats of the already generated?
I drew "girl, born in the Balkans in 3208 BCE", which I find super remarkable since this fits the place and time for a relatively small population of people who spoke Proto-Indo-European as their native language.
Before clicking, I said to myself "Indian peasant".
> You have drawn a girl, born in the Ganges Plain in 1628 CE.
> She was born, married at 16, lost her mother at 19, was orphaned at 30, had 8 children, and died at 39 of pneumonic plague, with 6 surviving children.
Nailed it. Though in retrospect, I drew really well. It would have been safer to guess "toddler death".
In 429 CE, Devadatta was born on the Deccan Plateau, near present-day Hyderabad. His hamlet was home to about 100 people. Among six children, he was the second. Devadatta died of neonatal tetanus or sepsis within days of birth.
Since this website has information about how many children born per woman, and how many survive on average, I would like to see a tool that, given a particular person, can estimate how many people today are their genetic descendant.
I noticed the traffic and am dropping by. A few quick answers after scanning through comments:
Q: Vibecoded?
A: Mostly, not completely. I've been programming for 30+ years, and like a lot of people have gotten to the point where I seldom need to look at code anymore. Even so, this took around 4-6 weeks (part time) to get mostly right. A lot of the effort was in assembling the data in a mostly non-hallucinogenic way. More on that below.
Q: What's up with the inaccuracies?
A: There are definitely inaccuracies. One of this big issues is that the data we have (mostly academic historical sources) only exist at a certain level of granularity, and attempting to "stretch" statistics is fundamentally error-prone. Different data sets overlap and disagree, and there are genuine unknowns. There's also a fundamental issue with using independent probabilities that should be conditional. For instance, x% of people died before 15, and y% in that same population died in an accident. Combining these statistics to get an estimate of children who died by accident involves coming up with a reasonable distribution of accidents by age, which is guesswork for a given historical time/place. In all, I'm shooting for mostly accurate given patchwork statistics, and a lot of filling in blanks in a reasonable way.
BTW, in order to test the accuracy of stories, I generated a few hundred random stories and created nit-picking "History teacher" agents to evaluate each for accuracy, then aggregated and addressed issues by frequency. It would have been hard to come anywhere close to that level of scrutiny without AI on tap. It's also why I'm so sure that there are still inaccuracies to be found. I was also able to point Claude at comments here, aggregate issues and fix them (with my ok), which means that the issues reported here should be mostly fixed.
Q: Do you cite sources?
A: I pushed pretty hard on trying to cite all sources, being as open as possible about where data is coming from. Each element of the story links to its source, and the sources page lists all the sources used, included the lines in each table that come from that source.
For me, this was one of the interesting bits of fallout from the project. How should we source well in the age of AI? I wound up making a distinction between actual downloaded data (from a real publisher), vs what AI "claims" is accurate, citing a source, vs what AI claims to be true, and seems reasonable, but Claude can't cite a source, making it smell funny. In the age of AI-flavored truthiness, it's important to push hard on being able to trace all claims down to their primary sources.
One more note: the biggest source of inaccuracy is the AI-generated images. Comments run about 2:1 in favor of ditching them, but a lot of people really like the idea of adding humanity to a story by supplying a human face, even if AI-image gen insists on dressing the guy from 1375BCE in a snappy crew-neck tee.
"A girl was born in 1276 BCE, on the Deccan Plateau, near present-day Hyderabad... She lived on a labourer's wage"
Wage labor didn't exist yet and the cited source is for labor history between 1500CE and the present.
The AI illustration also doesn't make any sense for the period.
This is the genuinely bad kind of slop because it's pretending to be semi-factual.
The more details this gives, the less likely it is that the story is true. Since the output is a highly detailed story, the statistical reality of this site is that's complete bullshit.
The website is quite clear that the stories are made up based on statistical likelihood, so that part didn't bother me at all. It's a cool way to explore the statistics.
What bothers me is that the statistics themselves are bad.
The basic premise of the site, that these lives are “taken at random from real data” is a lie. They are fictional extrapolations from a dataset of questionable provenance twisted into a narrative format by an LLM. These are not humans. They’re not even data points. It’s vibey bullshit.
Yes, but I definitely think the probabilities are off - the explanation, which I agree with, states "a random birth is far more likely to fall near the present."
But it appears to be drawing much more uniformly from the years than you'd expect given the probability curve (population curve) pictured.
Bear in mind that the population curve defaults to a log scale on the year axis. If you switch it over to linear, that big fat area at the end shrinks to a needle-like spike..
I am almost certain this is miscoded. I drew about 20 lives and I think 17 of them were before 1500. Meanwhile the vast majority of the integral of population over time is right of 1500.
an effective way to glean a little slice of understanding of the human experience throughout the ages, though 'every human never' might be more apt a name as its all statistics from the demographics to the generative portrait
> You have drawn a girl, born in the Levant in 1042 CE.
> She was born, married at 18, had 4 children, lost her mother at 28, was orphaned at 41, was widowed at 70, and died at 71 of heart failure (degenerative illness), with 3 surviving children and 9 grandchildren.
First try. Honestly I'm going to call this a W under the circumstances.
Interesting. It rolls a few dice and presents a snapshot of someone who might plausibly have lived at a random place and time. Mine was a girl who lived on the Deccan Plateau in India, and died at six of tuberculosis.
This site is a fantastic demonstration of what crap AI is when operated by nit wits. Notably the drawing is not in line with the actual population distribution over time. Pre 1500 years are vastly over represented. If the point is draw a random year without regard to population distribution then the site is pretty pathetic. I would be embarrassed to post this on HN.
Herodotus used the term Palaistine in his work The Histories in the 5th century BCE. The earliest written record of the word "Judea" is from extra-biblical writings from 733 BCE.
The most "official" designation was arguably when the Roman's designated Syria Palaestina in 135 CE
I wrote a short story based on a similar scenario. (Caveat, I did it with AI help, but my ideas and many iterations and me doing final editing to get it the way I want).
Without wishing to be unkind, this reads a lot more as 'one-shot with AI' than 'many iterations'; if you truly did that much editing, you didn't put it where it showed.
You've got all the AI favourite names and rhythms at once in the second paragraph, for example.
Yeah, I think the AI probably shows. But in fact it is the result of many iterations.
But what I was trying to get across are the ideas, not the language. I think it you were to read the whole thing, you'd probably find it interesting. But if not, feel free to let me know.
But my hypothesis is that there are people with skills for ideas that have the potential to be interesting stories, but not necessarily great prose skills; I don't think those necessarily line up in the same person. Up until now, to write a worthwhile short story you had to be among the best at both.
Maybe AI can help change that, and maybe (I hypothesize), that would be for the better. I understand that such a hypothesis would be controversial, but to me, it feels worth exploring.
The reason I did it the way I did is that I believe that Fable writes better prose than I do. I could write it so that people would be less likely to think it's written by AI, but I wouldn't think it's as good.
So, I'm happy with it as-is; I think the ideas are good and deserve to be out there. I think it just needs a context where people will be more forgiving about AI prose because they are interested enough in the ideas that are conveyed.
But of course, the challenge with creating such a context is that one person with an AI could spew out 1,000 crap stories a day and submit them all...
Not bad!
Did you ever read Stephen King’s short autobiographical story on how he sold his first few stories? He had to be persistent as could be, just years of writing and submitting to publishers and getting rejected. It is a good read.
It involves one life, drawn at random, from human history. I can't imagine how you could miss that if you read it.
But I am realizing that AI writing is so offensive to many people that if they get that vibe, they just won't read it, and therefore won't have any idea what is being attempted. Maybe there could be a context where that wouldn't be the case, but HN is probably not it.
I read it all, and the central theme seemed to be more about collecting random thoughts/memories rather than "lives", with a final twist about the apocalypse or something like that.
You really need to stretch quite a bit to find similarities between this post and this short story.
And since you incorrectly assumed that I didn't read it, I'll give you more unsolicited feedback. Yes the writing is horrendous, but the idea itself is not novel or thought provoking. The general idea has been done many times in the classics like Dick, Asimov, Chiang and others. And the twist at the end ("Oh my god! They couldn't see things in the future because there is no future!!!!!") is immediately predictable when they can't get samples in the future.
That being said, keep writing and enjoying yourself but 1) try to dial down on the AI assist, it really makes for a poor experience for the reader, 2) don't get on your high horse when someone points it out.
Nearly 500 upvotes for something that is so very obviously just LLM slop. I think Hackernews is just totally cooked at this point. Whatever discernment the hackernews crowd once had is gone.
I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously.
The firt citation for marriage data is listed as "Hajnal (RH31)", but links to a seemingly unrelated WorldCat search. The second citation is to "Kaplan (RH03)", which I was able to track down on sci-hub. It is a broad theory of human evolution that doesn't appear to mention the Yangtze Basin or China.
Another citation is to "[RH109] Model-supplied gap-fill (Claude Fable 5 and Claude Opus 5, 2026-08-05). Bounds and items written to close gaps no dataset we hold covers. Not a published work: see each row's `basis` for its stated reason. (2026)" -- this is an interesting way to describe having an LLM hallucinate something for you.
Please stop making vibe-coded websites -- it is not useful to provide incorrect facts. You are polluting the commons with garbage.
James was born on the British Isles, near Boulogne-sur-Mer, in 2009 CE.
James's table is mostly oat and barley bread, pork and dairy, and cabbage and peas, with apples in season, and coffee and sugar as well. He is literate.
Odd for a gen-z person in Britain. I doubt he’d eat a lot of cabbage, and I have no idea when the “season” for apples is as they are always in the supermarket and there aren’t any orchards near where I live.
> James died in 2027 as he jumped to his death from his room in a five-story hotel —his belly full of barley juice.
apple season is fall, if you're wondering.
him eating apples when in season seemed to me at first like an anachronism, but actually, i do that. years ago i got fed up with apples that were whatever, so i decided to only have apples when i could find mcintosh (my favorite apple), and eat loads of them.
I draw a man in China in 1819, and it said he "died of the war in Oman" without any explanation of how that could have happened. What a load of crap
That’s basically what I’m thinking.
Not only is this pointless, it’s outright mis information. Who gains anything from having an llm create an imaginary person ?
I guess for pure entertainment, but this looks like it’s a historical website
Welcome to the future of widespread information, where the uneducated masses don't understand what context, or the definition of words for that matter, even is.
Youtube is full of this, history slop. Some techbro hyperfixated on this and LLMs gives it life before he's bored and moved on to the next thing to hyperfixate on.
It's techbro ejaculate. 10 minutes of euphoria for slop to come out in the end to spread to unsuspecting victims.
The first part likely is just an artifact of an unwritten common assumption. That is, it's probably 96% of women, who lived to adulthood, who were ever married.
Maybe, but I doubt it. The vast majority of the sources, including the two listed for this claim, are listed as "attributed", meaning "A model (Opus or Fable) claims that this is what the cited source says. In my spot checks, I've found these claims to be accurate; still, since it's just a model claiming that author x says y, there's a nonzero chance that it's a hallucination."[1]
I couldn't track down the first source, but the second source goes over something completely different. Maybe Claude accurately remembered the first source when it was stolen during the training process, but forgive me if I'm skeptical. Feel free to track it down yourself and let me know. Otherwise, I will assume it is an LLM hallucination.
(Also, given my background knowledge of the Tang dynasty, I would be surprised if the average marriage age was as high as 18!)
[1]https://anyhumanever.com/sources
“ Claude accurately remembered the first source when it was stolen during the training process”
I see…
I don't follow what your point is here. If you're not aware, Anthropic pirated many books via libgen and similar tools for use during the training process[1]. Maybe you are not aware of this, or do not consider it theft?
[1] https://www.reuters.com/world/us-judge-approves-anthropics-1...
Key distinction though: Anthropic can be said to have "stolen" or pirated those books when they downloaded them illegally, but not when they trained Claude, because that same lawsuit found that training is fair use.
Anthropomorphizing makes for poor analogies, but it's kinda the same way your acquisition of background knowledge of the Tang dynasty from whatever source material you used is not considered stealing or pirating either.
Fair enough; what I meant was more like 'Perhaps Claude accurately recalled the stolen documents it was provided during training'
Did the original owners of the books no longer have them in their possession afterwards?
Theft involves removing from the source, Copyright violation is distributing copies without a legal right to do so. Training is fair use.
The judge found that it constituted copyright violation. Purchasing the books would cover their responsibilities with regards to training. The claim that they are destroying books to hold copyright is incorrect. The judge cited precedent that said instances of not destroying the originals were also fair use, his claim about destruction was to say even if this precedent were not considered fair use, the destruction that was already occurring (because it is easier to scan loose pages), invalidated one of the arguments against it.
The ruling was on training took the form
If A is ok then B is ok, but even if A were not ok, B might still be ok.
Excellent. I will thus take a copy of the latest Disney movie and train myself on it while claiming fair use as a defence to the accusation of copyright infringement.
You can do that legally as long as you do not confuse people that what you create might be an actual Disney product, or that you try to represent it as actually created by Disney.
Both of those are protected by law. Learning from a Disney movie and creating your own thing is not, and should not be, disallowed.
This is the true reason for watermarking by LLMs. So they don't get mad LLM disease by digesting their own product.
Even if this wasn't vibe coded it would be impossible to fact check billions of possible lives. I think thought provoking toys like this are a great use of AI, assuming no one takes them seriously.
I am not sure why it is thought-provoking for a website to lie to you? There is plenty of real history and demography to read about.
you can't fact check real individual lives but you can draw up a probabilistic representation
that is if your probabilities are actually correct, that the anthropological or archaeological or sociological field for that specific region and demographic is robust (and there is good primary evidence for supporting the claims), and you're not some vibe-coded AI tool that spits out a convincing lie constructed on not even educated guesses, casting shadows on the wall of the Platonic cave where the people who trust slop uncritically are chained to their ignorance for perpetuity
This is exactly the kind of reckless inaccuracy I've come to expect from AI.
I am genuinely puzzled as to why this kind of vibe-coded site are so popular on HN.
I consider the crowd here to be much smarter than I am, which is why I enjoyed the links that were shared and the discussions that followed.
The website is pretty, but the statistics and stories are the main point. Yet we are being fed hallucinated false claims, and somehow, that doesn't seem to be a problem...
I don't understand why it gets so many upvotes. We know Claude can create websites like this brilliantly now. The novel and interesting part is the transformation of the data. But if that part is sloppy, what's the point?
Until I read the comments here, I did not notice the stats were bad. I was just skimming the facts I thought were interesting, not looking for contradictions.
The site itself is a really cool concept! I just wish the information was accurate. Now that I know it's not, I feel cheated.
I suspect a lot of people don't read the comments.
Can you offer any evidence that it is any less accurate than a page hand crafted by a human?
I thought it provided an interesting tool for people to put things into perspective.
Well for one example, if you select a person born on the Italian peninsula in the 1400s and they die of bubonic plague, their year of death is not correlated to years when there were large bubonic plague outbreaks on the Italian peninsula.
This is not an error I would expect a human to make if they were interested enough in demography to make a website about it. For example, the wikipedia page on this topic[1] does a reasonable job of distinguishing that plagues did not occur every year and that they varied in lethality.
https://en.wikipedia.org/wiki/Second_plague_pandemic#Italian...
This is such a weird nitpick, it's clear the website deals in averages and assumptions. It's meant to be an interesting window into possible lives intermixed with historical happenings at, or around, those times.
I get that some people just hate AI, but of all things, this is a pretty fun & educational use of it.
I think our core disagreement is that I do not consider teaching falsehoods to be educational. My antipathy towards LLM outputs that haven't been fact-checked is mostly downstream of things like that.
This website doesn't "teach" anything. It's a fun way to explore and get exposed to history. Kind of like Civilization 6 or something. Do you think CIV6 "teaches falsehoods" because you can build the Pyramids as the French?
I haven't played Civilization. Generally I would not consider something educational if it contains mostly false information. Perhaps clearly demarcated parts of it with true information could be considered educational -- for example, looking at screenshots I don't see anything immediately objectionable in the civiliopedia "historical context" sections.
I think most players would understand that in-game choices will not reflect the choices people actually made historically. Having not played Civilization, I'm not sure whether I would consider it to be educational -- I think it would depend on whether playing the game provides insight into real historical events.
For example, I would consider a sufficiently detailed LARP of a historical event to be educational[1] -- the actual choices and outcomes do not match what happened historically, but they can teach participants a great deal about the motives and limitations of real historical actors.
This website, on the other hand, seems to consist of 90% AI hallucinations and 10% data that was actually extracted from a real source[2]. Since these types of data are thoroughly mixed together, I think this website is about as educational as just making up something to believe about the past.
[1] https://college.uchicago.edu/news/academic-stories/history-c...
[2] https://anyhumanever.com/sources
I don't think a hand-crafted app necessarily estimates things more accurately, but a sensible pre-LLM developer would probably not list so many dubious citations to trick people into believing it's accurate.
> I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously.
Yeah, I got: You have drawn a girl, born in the Yangtze Basin in 1295 CE.
Typical age at marriage: 18 Ever married (of adults): 96% Children per woman: 6.5
Life expectancy at birth: 28 yrs Died before age 5: 39% Died before age 15: 44%
Very similar to your "random" draw, and the facts are suspect. There was a link to the source, but it's behind a paywall.
Worst part is that it's often not clear at the first glance was it something carefully crafted or steamrolled by LLM.
I got a man in India 1000 years ago who died at 92. Impressive yet rather unlikely.
The thing about stats is outliers exist. Whether those outliers themselves are interesting is another matter.
Unlikely, but not implausible.
There are plenty of problems with the site but that some stats like marriage do not count dead newborns as unwed instead of invalid for the stat is not really one of them.
What age of death discounts someone as counting towards the marriage percentage? If a human made the website, there would be an answer to this question. That answer might not be published and might vary between datasets, but it would exist.
In this case there is no answer, because the percentage is just made up by the LLM. There is no methodology. Your answer is inappropriately anthropomorphizing the LLM by imputing a methodology onto its "data collection" process.
No anthropomorphization was required, I just looked at one of the sources on that section of the page and saw the statistic did not involve the number of newborns and so figured myself those facts could indeed be simultaneously true.
How do I know the methods described in the source were correctly used here? I don't, but I also wouldn't know that based on whether or not a human wrote the webapp. I'd know that by checking the method applied on the backend matches that of the source. If it does, it doesn't matter if it was generated by using every 17th letter of the dictionary - it'd either align or not.
Which source was that?
In my case it was:
Preston, S. H., Heuveline, P., & Guillot, M. (2001). Demography: Measuring and modeling population processes. Blackwell Publishers.
In GP's case I believe they are correct the Hajnal source is not relevant to the Yangtze Basin but reading the first page should have made it clear the facts could be simultaneously true regardless where the percentages came from or if they were accurate.
Again, there are plenty of problems with the site. That it was clear these facts can't all be true simultaneously was just not one of them.
This appears to be a general introductory demography textbook. I am surprised that you looked at it and it contains specific demographic tables about marriage rates at specific time periods? Perhaps I am misunderstanding what you mean by 'I looked at one of the sources and saw it did not involve newborn rates.'
I think it might be a simple misunderstanding, but one can never be certain :).
I believe we agree we haven't seen a valid citation for "the source in which 'x' came from in 'x%'" for this stat on the site. I hope the two of us also agree this is not the point I was disagreeing with.
Where I'm currently lost is why that should mandate a demography textbook, which was a different citation in the same section, must be unable to explain dead newborns need not be counted in such a statistic just because the textbook was not also the source of the specific percentages displayed? That'd be like saying a statistics textbook can't explain how the median can be less than the mean on the basis it wasn't also the source for the numbers displayed.
"I drew a woman born in 715 CE in the Yangtze Basin. It claims that 96% of women were married, the average age of marriage was 18, and 44% of women died before age 15. Clearly these facts can't all be true simultaneously." Umm yeah they can be true. Dead people are not counted. The straightforward interpretation is 96% of women living to 18 years or greater marry at some point in life. However the app has bigger problems, namely it does not seem to select from the distribution correctly, pre 1500 lives seem to be vastly over represented.
Yes, if a human told me this collection of facts I would assume they failed to properly explain their methodology. Given that the LLM is just making up numbers (see the "sources" page, where these are listed as "attributed",) there isn't really a methodology to explain in this case.
This is neat, but I don't think it's actually drawing the year from the correct probability distribution - "a random birth is far more likely to fall near the present" is correct, but in 5 "random" choices, only one was from anywhere near modern day.
I did a statistical analysis and you are correct. The issue seems to be that crude birth rates from around 500CE and earlier are overestimated significantly. The bad source seems to be Kaneda & Haub, "How Many People Have Ever Lived on Earth?", where they assume a flat crude birth rate of 80 for all BCE dates. The probability distribution is just a function of pop and CBR, so it results in significantly decreasing modern birth probabilities.
In other words, the Flat Birth Theory.
We can call them flat birthers
Off with the Flat Birthers, I'm a Reverse Birther!
https://www.youtube.com/watch?v=OsAIhGoKVuo
Come on now. We all know the birth is not flat. It can be clearly seen from space.
https://www.youtube.com/watch?v=KKpILoUMMb8
There are more statistical issues. I drew a person from 1984, living in Germany and recalculated their life several times. In almost all cases, their mother died while they were in their thirties, often both parents. This is simply far from reality.
[dead]
I think it's working as intended. The year is simply a randomly chosen year not related to the density of births in that year. It would be a boring tool if it actually chose "random year based on random birth" because almost every choice would be from 1900 onward...
That’s not what it says on the tin though. “Drawn at random from everyone who has ever lived” means years will not be uniformly distributed.
This is a case where I think they should change the wording as opposed to the functionality. I like that it's a random year regardless of population density.
But as other have noted it's almost certainly not just a random year.
I do think the cleanest and most fascinating interpretation is what the title says "One life, drawn at random from all who have ever lived" which clearly is about random people and has nothing to do with the year.
But I think experientially, the preferred feeling for a user is a random deep dive into the circumstantiality of different eras and different geographies than our own, rather than a true-to-life re-rolling that would constantly turn up 20th and 21st century outcomes. So the spirit of it is probably closer to the random year idea, but it's literally not random year, but maybe randomly something else, like eras however designated.
If the implementation was correct it wouldn’t be hard to let the user choose either option.
Yes, many births would be within the last 200 years but it’s still very different to be a man born in the year 1950 in New York vs a woman in 1950 in rural China.
That would certainly be easier at this point, but I like the “random human life” idea more.
I’ve reflected in the past about how lucky I am to have been born where I was, when I was, and be given the abilities that I have. Sure there are plenty of people that were luckier, but also a lot who were less lucky, most of whom are dead.
Should we call this a cargo cult
"random" != "uniformly random", if you want to be pedantic.
But "is far more likely to fall near the present" = "is far more likely to fall near the present". The author makes the claim about probability distribution themselves, then does the opposite.
You're right, yet I agree with the website's approach. The goal of this site is to educate users about how people lived, not educate them about stats. IMO. I would have chosen the same algo.
It's clear to me that an even bigger HC crowd would howl "Oh, how boring, look at him choosing pretty much only still alive people" if he had done the 'uniformly random" selection. C'mon!
Notice the log scale on the year axis. There is far more probability mass during older years because of that.
Log scale should only affect how the years are displayed on the chart, not how they are sampled.
The site covers a 300,000 year span, so a randomly chosen year would average ~150,000 BCE, which definitely isn't happening.
The skew towards the present isn't as stark as that. World population has exploded in the past couple of centuries, but it hasn't been high for very long. 1900 to present only covers ~11% of all births.
1900 to 2026 is 127 years (inclusive), so 0.04% of the year range available, yet it only has 11% of all births. That's certainly a skew, although I suppose subjective whether or not that's stark. At a >3x order of magnitude, it's certainly significant though.
You could definitely call it stark, but the comment I replied to suggested that it would be well over 50%.
> "a random birth is far more likely to fall near the present" is correct,
If the selection is from the population at random (as the UI indicates), but it specifically says it's based on a random year (during which humans were present) first as a filter. It's much more likely to select far from the population explosion.
The x-axis being log scale by default visually exadurates the effect if you are thinking about it in terms of area under the curve. This confused me also.
>exadurates
this word is usually spelled exaggerate (yes, with two g's.)
> yes, with two g's
The reason there are two Gs is that the first one is assimilated from a D. The etymological structure of the word is ex-ad-gero,¹ "from-toward-carry".
This happens for everything. Compare accelerate [ad-celero] / accommodate [ad-con-modus] / impossible [in-pot-s-ibilis] / intelligent [inter-leg-ents].
(Note that a doubled consonant in Latin is pronounced differently from a single one, as is still true today in Italian. English draws no distinction, which makes the etymological spellings likely to confuse English speakers.)
¹ This was already the form of the word in Latin. It looks malformed to me; I can't explain why exaggero is a first conjugation verb (with passive participle exaggerata) rather than simply being a prefixed form of its root (ag)gero, in which case the participle would have been exaggesta. Gero and aggero are both third conjugation verbs...
They should have scaled the probability curve by the derivative of log(time).
Why?
This is a good point, it's not as far off as it seems from a quick glance.
I guess you are right, but it makes it more interesting if the picks are not almost all from the recent past. On the other hand, I would also guess that spreading the picks evenly across the time range might result in many being from times with little data to differentiate them.
It makes the premise of the entire site flawed, though. And it defeats the purpose of the convenient button provided to redraw a year. That is what should allow you to find a pick that might be more interesting, rather than messing up the randomness promised by the site and the HN title.
That's sort of boring though, we generally know what it's like to live near the modern day already.
I think it'd be cool to compare a modern day life in rural Ethiopia to what I'm used to.
Yes, this is exactly what I expected (showing the anthropic principle in action), so I am disappointed they are not sampling correctly.
How near is near? With ~100 billion people having ever lived, the odds of a randomly selected life falling within the past couple of centuries are not all that high.
Okay step away from the waving hands and read the source material.
Could you be a little more specific? https://www.prb.org/news/how-many-people-have-ever-lived-on-... puts the halfway point at over 2,000 years ago, so you'd expect less than 50% probability of getting a random year that's not BCE. "A random birth is far more likely to fall near the present" is true in relative terms; the timeline on the web site goes back 300,000 years and you're very likely to get something in the past few thousand years, which is pretty near to the present when compared to a third of a million years. The random years it gives me fit into that pretty well, so I guessed that the person I replied to might be using a narrower definition of "anywhere near modern day" than that. Hence why I asked. If I missed something, could you kindly point out what I missed, instead of whatever this is?
The stat I keep in my back pocket is that you can roughly divide the total time lived by humans (the integral of population over time) into three tranches, with one third of human experience being before the dawn of agriculture, and one third being after 1600. (Related fun fact: WWII was about 1% of human experience, based on the back of the envelope that ~50% of the world experienced "being at war" for nearly a decade.)
The distribution of human lives versus years lived by humans will skew earlier, because infant mortality was absurd for the first two tranches, with something like 50% of people dying before the age of five, so it will be more like 40-40-20 over those time periods.
This could be pretty great for a play through of Thousand Year Old Vampire, a pretty terrific journaling game: https://thousandyearoldvampire.com/
And I suppose other RPGs, too, especially ones that are more story-than-dice driven, and/or set in the "real" world.
This can be great - Tom from the Yogscast has run a few of these, and they can get pretty wild. It's a simple premise but works well.
My daughter is four.
I drew a boy who lived 11000 years ago in Mexico.
His longest travel was a half day walk carried by his family to a nearby foraging ground.
He died at 4 from a zoonotic fever. Both parents outlived him.
It’s random, it’s ai generated. Makes me appreciate life nevertheless.
I drew Vuk,a boy who lived in the 13th century Balkans. He died at 2. His dad was already gone. According to the statistics the life expectancy at birth for him was around 50%. I suspect a lot of our random people would have died early, if they were born before the 19th or 20th century...
Whole years! You'll are so lucky! The reincarnation roulette gave me this: 49 BCE, he lived 1 month, dying of dysentery. I guess I'll stop complaining about my life then...
He could hardly complain about his, as he didn't have one.
I had a similar sensation. I often think about a friend of mine who mentioned, in passing, a sense of having won the lottery of life for being born in this day and age, in a first-world country
yeah, and you don't have to go back that far
last night my wife and I were watching the first episode of a TV show Deadwood (set in the Dakotas in late 1800s), and my wife commented "I could never have lived then; I would have just killed myself immediately". The irony of course is that she would have been fine had she been born in that era as it would have appeared as "normal" to her then as the world today appears to us (ok things aren't normal right now, but I mean comparatively speaking), and "better" than 100 years earlier.
This is at both times super cool as a project at what AI can do, while also being extremely dubious and misleading about how accurate any of this data is, hallucinated or not.
I mean it's only cool if it is actually real data that we can draw interesting and meaningful insights from. This is all just made up nonsense.
Her farthest journey was a day's walk, to a nearby market town.
I think this is what boggles my mind most about antiquity: the vast majority of people rarely ventured much past where they were born (unless they did a likely once in a lifetime religious pilgrimage which took the better part of a year).
Seeing as most other people didn't venture further either, you really got to know the people around you very well. Common life limitations simply precluded much travel. The only foreigners you saw were either invading hordes, or traders who were constantly on the road.
My grandmother told a story about meeting an elderly female relative in Kingston, Tasmania, soon after my grandparents had got their first car. After visiting, they said they were now going back to the city (Hobart), and the old lady said, "Oh, how lovely, I've never been!" My grandmother reported, "And I should have told her 'get in the car then!', because she died not too long after."
Kingston is 13 kilometers from Hobart.
There were and have always been trading and other exchanges all over the world. The Egtved girl was traveling from north to central Germany, probably for marriage. Now, if we mean most people did not do this, then sure. But there's a persistent idea that ancient cultures were not in contact with each other, and this is largely disproven.
15,370 BCE feels like a lottery roll. I doubt I’ll live long.
Edit: oh it’s not weighed by population? I was expecting 99% of rolls to be in recent times.
We might recognize that the economics of making such a fascinating and artful experiment like this have completely changed. No ads. Not asking for anything except your attention, and then delivering an outcome that is fascinating and honestly, far beyond what a person would be incentivized to build before AI tooling.
I loved it and feel inspired from it.
(It looks like it's getting a lot of traffic and might get hugged to death)
> No ads. Not asking for anything except your attention, and then delivering an outcome that is fascinating and honestly, far beyond what a person would be incentivized to build before AI tooling.
I don't think I agree. There were plenty of interesting side-projects on the internet before AI.
I'm not sure why AI means that people don't need to support their creations with ads. Sure, it probably took fewer human-hours typing at a keyboard to create this, but for things like this, a labor of love, that's typically not the "expensive" part.
In fact, there was a time when the internet was mostly interesting side projects, if you go back far enough (the mid 1990s, say).
It’s true. It feels like that age may be revisiting us. Makes me happy to see stuff like this which isn’t upselling something.
Yes indeed. What do we know about how this was created and how long it took? Before I look into it, I’m going to guess:
One experienced developer with extensive anthropology experience, probably a PhD, working with Claude Code and Claude Cowork, full time over a two month period. Or somebody with limited dev experience over four months.
Anyway, sad that all the first comments were complaints… what a wonderful project!
> There were plenty of interesting side-projects on the internet before AI.
This is a strawman.
The original point was "the economics ... have changed"
I read this as, it's much easier to spin up a project of this scale even non-monetized, and more people are able and willing to do so.
Yes, it's easier - but I'm not sure what that has to do with ads or asking for things beyond attention.
I think it's just that in the past, you would need X free time, Y motivation level, and Z skill level (partly tradeable for X and Y) to create something like this, and now all of those values are smaller. So it's 'economics' in the straightforward sense, because not everyone could afford to spend X hours on something they weren't going to monetize, and also in the sense of the incentives shaping people's productive/creative activity.
(As for whether it's a good thing, I'm conflicted. It's cool to see more stuff like this, but if it crowds out the more intense passion projects, we lose something too.)
I believe that part serves to emphasize the main point.
To elucidate for this particular example, before AI agents existed, projects with this level of data and depth might've needed to be monetized in order to justify the resources to bring it to life.
Now that the barrier is much lower, the creator didn't mind spinning up and providing it with no monetization (e.g. no ads).
> before AI agents existed, projects with this level of data and depth might've needed to be monetized in order to justify the resources to bring it to life.
Yes, this was exactly what I meant in OP.
The site was registered a month ago. If it ever actually gets and stays popular, see if it has no ads then. The economics of personal interest sites didn't change in the sense that they're any harder to make now compared to the mid-90s. The economics that changed is the Internet now has billions of users and attention being monetized can make you a lot richer now than then. It's very hard for people to see an opportunity to become a multi-millionaire by clicking a button but then not do it because they'd rather maintain a slightly better experience for their audience.
It's, of course, not even loading right now, which gets at the other problem with this "economics" framing. The biggest expense in running an actually popular website isn't creating and publishing the first version. It's paying for the infrastructure to serve a meaningful proportion of the entire world. This is clearly not a static site. It's presumably doing some computation on the server side to draw from all these hundreds of different probability distributions that works well enough for a few dozen simultaneous requests but not thousands. There might even be some kind of perverse anti-economics going on here where thought devoted upfront to efficiency and architecture might have made this far more scalable, but the "publish the first thing that passes a basic sanity check" vibe-slop ethos doesn't usually get you that.
Vibe slop? The structure of the site is quite solid, the UX decisions are spot on, the data looks good to me that knows little about this stuff, and it is entertaining and informative. This was not a one shot, a human out a lot of effort into this.
Yeah it is imperfect and was maybe not ready for the HN front page hug. But I say, bravo to the person or team!
> the data looks good to me that knows little about this stuff
Do you see the problem here?
I also think this shows what the changed economics do: it allows anyone to fill the web with convincing looking nonsense.
The person who built this did not have to put in months to build a cool web app. But they also did not have to put in months of research, gathering representative pictures. And they didn't. Instead they let a software create a well presented nonsense story about a boy from Anatolia in 200 BCE called Mehmet wearing a parka.
> We might recognize that the economics of making such a fascinating and artful experiment like this have completely changed.
And this is not limited to software development. In just a few years, the economics of delivering healthcare, education, entertainment and legal services are going to change completely. Transportation is changing as we speak. In a decade Physical AI will completely change the economics of manufacturing, food services, elder care, and so on.
The post-scarcity world will be here sooner than we think!
In 10 years? I predict that healthcare, education, legal services and transporation won't change in a meaningful way at all. People will go to a doctor, kids will attend school and aim to graduate, legal services will still be done by lawyers and decided in court. Transportation will still depend largely on what exists right now or is being built right now. There won't be self-driving cars in every city, only in a few, probably more in the US and China, fewer in Europe, probably close to zero in the rest of the world.
And how do you think will the post-scarcity affect the daily lives of most people? My prediction is that most people even in 30 years will still work for a wage, and probably still around 30-40 hours per week, because these gains of post-scarcity won't be distributed evenly. It will mostly make a few people very very rich.
> In 10 years? I predict that healthcare, education, legal services and transportation won't change in a meaningful way at all.
"We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten." -- Bill Gates, in The Road Ahead.
I’d guess that education will hardly be recognizable in 10 years. Bad news for number of teachers employed, but AI is going to result in a huge increase in the number of 15-year-olds with a PhD’s level of knowledge and abilities in their heads.
We are not in a scarcity world. Or it's deliberated scarcity. We have way enough means to educate, heal, feed and transport everyone. Yet most of the world work full time to produce more and more of meaningless products, and to make the economic curves go up. Also, a huge part of the population is still poor, while the rich people get richer. And on us dooms the perspective of an extreme climate crisis.
I don't know how you can think that AI or "Physical AI" will solve any of that without a really huge shift in our locals and worldwide politics. It will probably cause a massive economical crisis before anything else of importance.
But its just AI slop though. None of the data is real. It's completely made up. There is nothing interesting about this.
Both of these things can be true. It's a really good idea that immediately led me to another idea to play with. But it hallucinates so its accuracy can't be trusted.
This is brilliant. I think it could stand to be a little more imaginative in their stories. It seems like it should be possible to do that while still conveying an accurate picture. "He made his living as a professional" is simply a wasted opportunity. There shouldn't be anything so vague. If you we're listening to a story, you would feel cheated. Embrace your idea.
Well, its just AI-generated nonsense, so the story is just going to be the same slop you could generate with ChatGPT or whatever.
Shameless plug, but related, is the Watsi platform[1] I discovered a few months ago through another HN user. You can read the story of less fortunate but alive people and how much need there is in the world for one another. The founder's talk[2], Chase Adam, which made me a supporter, is a must-see. I guess it's time to rewatch it..
[1] https://watsi.org/
[2] https://www.youtube.com/watch?v=WlT3UhC7NwQ
I am really happy to live in the modern age. It is quite amazing to think that people are facing cardiovascular disease and cancer instead of random parasites nowadays. Modern medicine is such a miracle.
> instead of random parasites nowadays
Maybe MAHA will bring this back, like they did measles.
This will mostly self-correct. Nature really doesn’t give a fuck about ideology.
Just vaccinate yourself and your kids and listen to the advice of real doctors.
I was thinking along the same lines as I was going person to person in the app. It made me more disgusted at MAGA and MAHA. These guys are living in modern times with access to things that people back then couldn't even possibly dream of. The thing that really drove it home is the life expectancy of the people that came up in the draw. Most died by their 30s and some even before their 30s, most never knew their parents as they died when people were single digits of age, and they were forced to live very simple life with simple foods and conveniences.
So much of what we have has been brought to us thanks to ubiquitous fuel, ubiquitous technology, and institutions that took centuries to form and solidify. All of this built on the backs of people who sacrifiec and sometimes did horrendus things to allow us to be in the places we are. There's just so much hubris and misunderstanding of what they have been given in their lives that they don't appreciate. Pissing all this away is fine to them. Yes, I get it...they didn't get the same quality of life as their parents did(who came of age in an abberation).
What they don't realize is that it could get much worse. Many of them may not even live to see how bad it could get. This is so frustrating, and it just increases my disrespect for them.
How many random parasites has modern medicine removed from you? Likely you never even had one. It's not just medicine, we also have abundant, safe food, clean water and warm housing
>we also have abundant, safe food, clean water and warm housing
for now
Yes! I think it’s really easy for us to take for granted the long list of incredible things we take for granted. Every time I hear someone lament how horrible the world is today, I wonder when they’d rather have been born. Especially as a random human like this.
"You have died of Dysentery [on the Korean Peninsula in 3064 BCE at 2 months old]"
I was actually a bit surprised since it also said I was exclusively breastfed which seems to be fairly protective against diarrhea. I have a feeling this didn't account for all the joint probability distributions...
All babies were exclusively breastfed in 3064 BCE.
Yeah, it doesn't give totally absurd combinations (like a male dying of birth complications at age 10), but the probabilities are still likely rather rough approximations (eg. sorting cause-of-death distributions to a few bins like "infant", "child", "adult", "elder").
Strangely, I too, died in the Korean Peninsula at 8 months old, but circa 6000 BCE.
Hmm, I was at the Korean Penisula too. How representatively random is this thing?
I died at age 36 of a workplace accident in the Yangtze delta in the year 700
I died of malaria at age 1 in the West African Sahel!
I died at 1 in Mesopotamia in 3966 BCE of dysentery.
>You have drawn a boy, born in the Korean Peninsula in 3394 BCE
>Life expectancy at birth: 28 yrs
>Died before age 5: 42%
>Died before age 15: 50%
i knew life expectancy was obviously lower, but those stats were a bit eye opening.
unfortunately, died at 3 from an accident.
My first run got Korean Peninsula in 3064 BCE...
That is surprisingly similar given how few people seem to have commented so far. Makes me wonder if some part of this is starting from a constant seed or something...
"Any Korean Ever"
I do wonder if that life expectancy is dragged significantly down by all the ways you can die early on in life, so that it's not so much that 28 years is an old age for the time as it is a function of all things that prevent you from getting to old age.
And so if you went to a town square or market in that time (if it's a time that had such a thing), perhaps it might still be perfectly normal to see people in their 30s, 40s, 50s.
That's exactly the case. Overall average life expenctancy as adult didn't change much from older times - about 15 years more.
The website represents this well. My random person was born in the Neolithic and lived to 68. Not bad.
That's exactly what happens.
The life expectancy of a person who lives to 18 increases dramatically across the board.
You even see this today. If you live to 70, you can expect to live 14 to 17 more years. So to about 84 on the short side and 87 on the long side.
However, make it to 75, your life expectancy is 11 to 14 years. Which is 86 on the short side and 89 on the long side.
Basically, the longer you live, the longer you should live, but eventually the range collapses. And that's because people come in under the average all the time. Some people die at 71. And those people stop getting counted when determining life expectancy at 75.
Right, that 42% dying before age 5 means the survivors typically lived well beyond 28. There weren’t a huge number of people living into their 80s, but 28 wouldn’t have been “old man” territory.
I think the model has an unusual preference for the Korean Peninsula before Christ. A number of commenters, including myself, draw from this area and period.
The life template kinda breaks down if you land in a modern era which makes me think it's probably equally as silly sounding to apply it to a hypothetical life from 10k years ago.
But I guess it's inevitable when you try to paint individual lives through pure statistics.
This is what I got:
Seojun - Seojun is not a real person. But his life story is created from the statistical reality of the Korean Peninsula in 1999 CE.
In 1999 CE, Seojun was born in the countryside near Pongch’ŏn-ŭp. His village is home to about 250 people. Among four children, he was the fourth. He eats largely rice, soybean paste and tofu, dried fish and shellfish, and pickled cabbage and radish, with persimmon and jujube in season, and tea as well. He knows how to read and write. Seojun stands about 167 cm tall (5 ft 6 in).
--- if it's not obvious why this is problematic and extremely offensive I don't know what I can say to you. You're talking about GenZ who has access to tiktok and food delivery apps as if you're talking about someone from Joseon dynasty from hundreds of years ago. No one's favorite food is rice and soybean paste. He knows how to read and write and sky is blue.
It doesn’t say his favorite food is anything? Have you misread this and decided to take offense?
I think this app conflates ingredients with actual food. From the listed food the person probably eats 된장찌개 but the ingredients are listed one by one
Now this is fun vibe coding. A fascinating way of reckoning with our cosmic luck. On "only the first roll counts" rules, I rolled a girl who died at 8 months from a respiratory infection in the Indus valley, in 7785 BCE (9,810 years ago).
I rolled again, of course, but at that point it feels like fishing for a best case scenario. But I at least got someone from the North China Plain in 75 BCE who lived to 74 and had 6 grand children. Married at 19 and stayed married to age 71. Hard to hope for better than that, against these odds at least.
Edit: Obviously progress isn't inevitable, and at any point in time you can be unlucky. But it does seem like, for most points in history, you would rather be you, in your version of the present, than a randomly rolled person from any prior point of history leading up to your moment.
So most of the time, anyone playing this would feel more lucky to be themselves than any of the options they would have to roll for. Maybe if a version of this was offered to people 10,000 years from now they would say "crap, I rolled 2026".
Edit 2: I think the death hug might have got it.
I died within 2 days on my first roll. made it to 8 months on my second
I was so lucky! One lived to 35 in the Paris basin a few thousand years ago, the other lived to 25 in China, died in 1945. Both were married and left living children!
on my 6th only the first roll counts roll, i got a boy born near Nairobi in 36,112 BCE that lived to 70
probably like 5% of humanity has his genes now. that was fun vibe coding
I think it would be more interesting if you entered your own real data and then the LLM gave you statistics based on that info on where you are on your own life path based on consensus data.
Very nice idea!
I think some number are of by a magnitude (DIED BEFORE AGE 5: 4%, for Scandinavium in 1990 - according to UN it’s ~0.6%).
My first draw:
"He died of dysentery in the first days of life. His body was laid in a stone cist or a great urn beneath a ring of boulders, with iron tools and black-and-red pots."
The data will be much clearer in the future, once all of our data has been mined. We create so much personal media and collected data that some point in the future when one entity has loaded that, a crazy real autobiography could be created.
I think graph being log makes it a little hard to understand that the median birth year was around 2000 years ago. On the graph that is in the flat part but the flat part was a really, really long time.
I will say this with all honesty and sincerity to everyone who interacted with this site: If you didn't realize this was just LLM-generated nonsense within the first 45-60 seconds of using it, it's completely and unequivocally over for you. Your lack of discernment will lead to you being jerked around by AI-generated slop for the rest of your life.
This is really neat. It would be interesting to talk with the character generated.
[dead]
Thanks for making this! This was super fascinating and the type of thing I love seeing on HN!
I died of dysentery shortly after birth. So basically, Oregon Trail - Ancient Edition.
After having my first child (a girl who is now nearly 3 years old) these stats impact me so much more. My first roll was a girl in the Iraniam plateau in the 18 century who had a 33% chance of dying before they reached 5. She managed to live to 70 but experience so much death in her life.
I can't imagine my daughter dying before she reaches 5, it feels like it would be the worst feeling I could possibly imagine and then some.
Visit any older cemetery and look for the tiny headstones and the matching very brief lifespans, or read an older obituary for someone from your great-grandparent's generation or further back, or watch a couple of seasons of _Call the Midwife_.
This reminded me about the US nutters advocating return to the "way things were": they would do well to consider life expectancy. My draw from 1881 Mississippi would max out at 36.
It seems that this conflates the leading causes of death in $BIRTH_YEAR with the leading causes of death for someone born in $BIRTH_YEAR?
I think this would be cool if it actually worked as intended. This simulation doesn’t even draw birth years correctly.
The right way would be to bucket people by something like 1 million. Everyone in the bucket has the same sex, birth year range, and location range. Then draw a bucket uniformly and infer the details.
It's funny to pick a birth year from our era and think of how history will represent us:
Martha was born in the countryside near Short Pump, in 1993 CE. Her village holds about 300 people. She makes her home in a timber-framed detached house. Martha has neither brothers nor sisters. She eats mostly wheat bread and maize, pork and dairy, and potatoes and garden vegetables, with apples in season, and coffee, soft drinks and snack foods too. She knows her letters. Her height is about 171 cm (5 ft 7 in).
An “Eastern Woodlands” person born 1989 is recorded to have experienced a “volcanic winter” and “dimmed sky” at age 2 as a result of the eruption of Mount Pinatubo.
This is an… interesting contextualization.
I am wondering how many historical situations are sexed up to be something significant, yet were nothing to write home about in the moment. All it took was a historian with an active imagination to bring it to the forefront.
> LIFE EXPECTANCY AT BIRTH
Given the huge infant mortality rates until very recently in human history, this is often pretty misleading. Life expectancy at say, 2yo, is far greater.
It seems like the server might be overwhelmed
Very cool project
I was wondering if the generated people “fill up a spot” and temporarily change the stats for generating new lives. Eg, if the app already generated 10M stories, is it still using the same initial distribution for the new ones, or is generation conditioned on the stats of the already generated?
The random life I got: https://anyhumanever.com/life/1323665477-1682995380-16357802...
I drew one that died as a newborn, 960BCE, then another that lived a long long life! https://anyhumanever.com/life/2139137610-1449878292-21360978...
Ooh I like that it creates a URL that you can share that points to a particular generated human. Nice feature devs!
Hug of death I think.
Gave me a good and convincing first couple of tries. Good work.
Works but it's slow to load up.
I drew "girl, born in the Balkans in 3208 BCE", which I find super remarkable since this fits the place and time for a relatively small population of people who spoke Proto-Indo-European as their native language.
Before clicking, I said to myself "Indian peasant".
> You have drawn a girl, born in the Ganges Plain in 1628 CE.
> She was born, married at 16, lost her mother at 19, was orphaned at 30, had 8 children, and died at 39 of pneumonic plague, with 6 surviving children.
Nailed it. Though in retrospect, I drew really well. It would have been safer to guess "toddler death".
My first random selection lived to age 68! I felt very lucky, and then that put my modern life into interesting perspective.
I would like to see a life expectancy number that excludes deaths in childhood.
A frequency distribution of life span for those born in year X would be a helpful visual.
Really puts the marvels of modern life into perspective, most people did not live long or happy lives for much of human history.
"At 79, she died of a respiratory infection. She left 18 grandchildren and five great-grandchildren behind."
What are the odds of 79. Must've been in the very top percentile
Oh nice! If you liked it, you might like:
yourlifeunderrome.com
I drew a man who lived to 80 in 47,000 BCE. Beat that!
In 429 CE, Devadatta was born on the Deccan Plateau, near present-day Hyderabad. His hamlet was home to about 100 people. Among six children, he was the second. Devadatta died of neonatal tetanus or sepsis within days of birth.
Hardcore
Since this website has information about how many children born per woman, and how many survive on average, I would like to see a tool that, given a particular person, can estimate how many people today are their genetic descendant.
Creator here (proof https://anyhumanever.com/verify).
I noticed the traffic and am dropping by. A few quick answers after scanning through comments:
Q: Vibecoded? A: Mostly, not completely. I've been programming for 30+ years, and like a lot of people have gotten to the point where I seldom need to look at code anymore. Even so, this took around 4-6 weeks (part time) to get mostly right. A lot of the effort was in assembling the data in a mostly non-hallucinogenic way. More on that below.
Q: What's up with the inaccuracies? A: There are definitely inaccuracies. One of this big issues is that the data we have (mostly academic historical sources) only exist at a certain level of granularity, and attempting to "stretch" statistics is fundamentally error-prone. Different data sets overlap and disagree, and there are genuine unknowns. There's also a fundamental issue with using independent probabilities that should be conditional. For instance, x% of people died before 15, and y% in that same population died in an accident. Combining these statistics to get an estimate of children who died by accident involves coming up with a reasonable distribution of accidents by age, which is guesswork for a given historical time/place. In all, I'm shooting for mostly accurate given patchwork statistics, and a lot of filling in blanks in a reasonable way.
BTW, in order to test the accuracy of stories, I generated a few hundred random stories and created nit-picking "History teacher" agents to evaluate each for accuracy, then aggregated and addressed issues by frequency. It would have been hard to come anywhere close to that level of scrutiny without AI on tap. It's also why I'm so sure that there are still inaccuracies to be found. I was also able to point Claude at comments here, aggregate issues and fix them (with my ok), which means that the issues reported here should be mostly fixed.
Q: Do you cite sources? A: I pushed pretty hard on trying to cite all sources, being as open as possible about where data is coming from. Each element of the story links to its source, and the sources page lists all the sources used, included the lines in each table that come from that source.
For me, this was one of the interesting bits of fallout from the project. How should we source well in the age of AI? I wound up making a distinction between actual downloaded data (from a real publisher), vs what AI "claims" is accurate, citing a source, vs what AI claims to be true, and seems reasonable, but Claude can't cite a source, making it smell funny. In the age of AI-flavored truthiness, it's important to push hard on being able to trace all claims down to their primary sources.
One more note: the biggest source of inaccuracy is the AI-generated images. Comments run about 2:1 in favor of ditching them, but a lot of people really like the idea of adding humanity to a story by supplying a human face, even if AI-image gen insists on dressing the guy from 1375BCE in a snappy crew-neck tee.
Roy: A Life from Ricky and Morty comes to mind.
Extra points if you can take your example life off grid.
https://youtube.com/shorts/CssbcJ44dos
Just my luck.
503CE, dead at six months from smallpox.
i love this. it really shows you someone else's perspective.
had you drawn a different lot in life things would be very different.
imo it is a good example of Rawl's veil of ignorance thought experiment.
How would you design the world if you will be randomly dropped into any position anywhere on earth.
well done.
"A girl was born in 1276 BCE, on the Deccan Plateau, near present-day Hyderabad... She lived on a labourer's wage"
Wage labor didn't exist yet and the cited source is for labor history between 1500CE and the present. The AI illustration also doesn't make any sense for the period.
This is the genuinely bad kind of slop because it's pretending to be semi-factual.
This is awesome. I love seeing how humans have evolved into what we are today.
> He was born, lost his father at 5, and died at 7 of malnutrition.
Yikes
I feel sorry for the person who draws their exact real-life stats, but dying this year.
Leading cause of death: dysentry
Had to laugh.
> You have drawn a boy, born in the Russian Plain in 3843 BCE.
He probably spoke something close to proto-Indo-European
“He was born and died at 1 of dysentery.“ - lucky me
I died of dysentery shortly after being born.
Makes Oregon Trail seem easy…
Spot checked a life from 1990s in the Iberian peninsula and it cites the population at 72 million people. Off by quite a bit. Pretty cool regardless!
I drew a woman from 500 BCE in the ganges... but lived to 75?
we are probably living inside a more advanced version of this game right now.
Somebody, somewhere, somewhen rolled the dice and here we are.
Died at 2 of dysentery. What is this, the Oregon trail?
The more details this gives, the less likely it is that the story is true. Since the output is a highly detailed story, the statistical reality of this site is that's complete bullshit.
The website is quite clear that the stories are made up based on statistical likelihood, so that part didn't bother me at all. It's a cool way to explore the statistics.
What bothers me is that the statistics themselves are bad.
Most likely died of smallpox...
Great, let's bring back smallpox and measles because the US elected officials are too dumb or genuinely want to destroy the world.
I had this idea for forever. Thank you for making it!
Absolute trash. No more worthwhile than the AI slop inaccurate copies of actually well made educational videos on YouTube. Why do we accept this crap?
The basic premise of the site, that these lives are “taken at random from real data” is a lie. They are fictional extrapolations from a dataset of questionable provenance twisted into a narrative format by an LLM. These are not humans. They’re not even data points. It’s vibey bullshit.
Has anyone managed to roll a life later than ~600CE?
Yes, but I definitely think the probabilities are off - the explanation, which I agree with, states "a random birth is far more likely to fall near the present."
But it appears to be drawing much more uniformly from the years than you'd expect given the probability curve (population curve) pictured.
Bear in mind that the population curve defaults to a log scale on the year axis. If you switch it over to linear, that big fat area at the end shrinks to a needle-like spike..
1956 CE: https://anyhumanever.com/life/1409269740-1599827252-31263759...
I got 1951: https://anyhumanever.com/life/982671145-1647234921-653406957...
1816 CE: https://anyhumanever.com/life/1323665477-1682995380-16357802...
Managed to roll a guy born in 1988 in the eastern US. He hasn’t married yet and he’s still alive.
1980CE, Nile Valley, died at 68 from cardiovascular disease.
You can chose.
Everyone keeps dying way too young :(
I am almost certain this is miscoded. I drew about 20 lives and I think 17 of them were before 1500. Meanwhile the vast majority of the integral of population over time is right of 1500.
> She was born and died at 2 of bubonic plague.
Yeah that does accurately sum up the 15th century.
Still slop coded bollocks though.
damn, well done
og/goat demo of synthetic data gen
only nit is 'draw' button often below the fold
I found it very unrealistic that my girl from 2000bc somewhere in china decided to divorce after 3 years of marriage when she was 23.
boy, born in the Guinea Coast in 211,966 BCE. Died from an accident at age 40.
Very cool. Reads like a Dwarf Fortress history. For those that don't play, every creature in the game has a life story.
rolled a random human and got... myself?? statistically wild. turns out i'm just a guy. weirdly humbling
an effective way to glean a little slice of understanding of the human experience throughout the ages, though 'every human never' might be more apt a name as its all statistics from the demographics to the generative portrait
I died at age 2 of Typhoid Fever.
Modern take on Oregon Trail vibes I like it
Very moving and a little hard.
This site is pure slop?
> You have drawn a girl, born in the Levant in 1042 CE.
> She was born, married at 18, had 4 children, lost her mother at 28, was orphaned at 41, was widowed at 70, and died at 71 of heart failure (degenerative illness), with 3 surviving children and 9 grandchildren.
First try. Honestly I'm going to call this a W under the circumstances.
Interesting. It rolls a few dice and presents a snapshot of someone who might plausibly have lived at a random place and time. Mine was a girl who lived on the Deccan Plateau in India, and died at six of tuberculosis.
This is super cool
"You have died of dysentery."
This site is a fantastic demonstration of what crap AI is when operated by nit wits. Notably the drawing is not in line with the actual population distribution over time. Pre 1500 years are vastly over represented. If the point is draw a random year without regard to population distribution then the site is pretty pathetic. I would be embarrassed to post this on HN.
[dead]
[dead]
[dead]
[flagged]
Herodotus used the term Palaistine in his work The Histories in the 5th century BCE. The earliest written record of the word "Judea" is from extra-biblical writings from 733 BCE.
The most "official" designation was arguably when the Roman's designated Syria Palaestina in 135 CE
https://en.wikipedia.org/wiki/History_of_Palestine
Love it!
"You have drawn a boy, born in the Paris Basin in 1970 CE."
Of all human history, it had to pick something off by 20 years and 200km.
I wrote a short story based on a similar scenario. (Caveat, I did it with AI help, but my ideas and many iterations and me doing final editing to get it the way I want).
https://garyrobinson.net/2026/08/18/everyone-who-will-ever-b...
Without wishing to be unkind, this reads a lot more as 'one-shot with AI' than 'many iterations'; if you truly did that much editing, you didn't put it where it showed.
You've got all the AI favourite names and rhythms at once in the second paragraph, for example.
Yeah, I think the AI probably shows. But in fact it is the result of many iterations.
But what I was trying to get across are the ideas, not the language. I think it you were to read the whole thing, you'd probably find it interesting. But if not, feel free to let me know.
But my hypothesis is that there are people with skills for ideas that have the potential to be interesting stories, but not necessarily great prose skills; I don't think those necessarily line up in the same person. Up until now, to write a worthwhile short story you had to be among the best at both.
Maybe AI can help change that, and maybe (I hypothesize), that would be for the better. I understand that such a hypothesis would be controversial, but to me, it feels worth exploring.
The reason I did it the way I did is that I believe that Fable writes better prose than I do. I could write it so that people would be less likely to think it's written by AI, but I wouldn't think it's as good.
So, I'm happy with it as-is; I think the ideas are good and deserve to be out there. I think it just needs a context where people will be more forgiving about AI prose because they are interested enough in the ideas that are conveyed.
But of course, the challenge with creating such a context is that one person with an AI could spew out 1,000 crap stories a day and submit them all...
Not bad! Did you ever read Stephen King’s short autobiographical story on how he sold his first few stories? He had to be persistent as could be, just years of writing and submitting to publishers and getting rejected. It is a good read.
What does the short story have to do with the scenario exactly? Seems completely unrelated.
And OMG, you need to edit the AI slop way way more. This is unreadable.
It involves one life, drawn at random, from human history. I can't imagine how you could miss that if you read it.
But I am realizing that AI writing is so offensive to many people that if they get that vibe, they just won't read it, and therefore won't have any idea what is being attempted. Maybe there could be a context where that wouldn't be the case, but HN is probably not it.
I read it all, and the central theme seemed to be more about collecting random thoughts/memories rather than "lives", with a final twist about the apocalypse or something like that.
You really need to stretch quite a bit to find similarities between this post and this short story.
And since you incorrectly assumed that I didn't read it, I'll give you more unsolicited feedback. Yes the writing is horrendous, but the idea itself is not novel or thought provoking. The general idea has been done many times in the classics like Dick, Asimov, Chiang and others. And the twist at the end ("Oh my god! They couldn't see things in the future because there is no future!!!!!") is immediately predictable when they can't get samples in the future.
That being said, keep writing and enjoying yourself but 1) try to dial down on the AI assist, it really makes for a poor experience for the reader, 2) don't get on your high horse when someone points it out.
Nearly 500 upvotes for something that is so very obviously just LLM slop. I think Hackernews is just totally cooked at this point. Whatever discernment the hackernews crowd once had is gone.