Can’t recommend the book that coined this acronym enough: The WEIRDest People in the World
Book by Joseph Henrich.
It is such an eye-opening piece that explains so much of the world around us. He’s an anthropologist that goes into the psychology of it all. Touching on points like how religion plays a part in shaping the America of today and even how humans are worst at discerning faces today because we need to discern letters and words and dedicate brain power for that.
There are so many interest studies mentioned there, one that really stuck with me is how Protestant-raised Americans will work harder for the next day after having (reasearch-led) incestuous thoughts when compared to Catholics and Atheists.
He explains how monogamy is to blame for a lot of our western views today, and how Mormon towns in Utah were affected by not having monogamy as the basis of society (women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man).
One of the wildest claims in there is the one that the north of Italy is more developed today because it was part of the Holy Roman Empire while the south wasn’t. About a thousand years separate these and he finds effects still. Mostly in connection to the spread of read/write to the public being a core tenant of Protestantism.
Anyway, this is not a summary of the book but instead a few points from it that really stuck with me after reading it. Fascinating stuff
I have to ask (and I don't mean this combatively) - given the ongoing realization of the replication crisis, how likely is it that the book you mention reflects a summation of the "too pat" studies about human behavior that, en masse, always seem pithy in an interesting headline, but years later end up being completely bunk?
I've noticed over the years many chains of reasoning - made up of what I believe someone called "cocktail party" pithy takes - that only last as long as you don't dig into the nuts and bolts of them. Pleasant little takes on our psyche and behavior that makes for nice reaffirming thoughts of our views but break down under later analysis.
It feels like we have sometimes accreted an amalgam of these pithy takes based on very small, one off, studies (never replicated) that let us comfortably assemble an affirmation of our broader takes.
This is a rotten thing to say about your book recommendation, given I have never read it (I hope you'll forgive me), but based on the last few years of the replication crisis, do you think, in your heart of hearts, that what you are describing truly does stand up?
I do not have the time now to craft you a full answer as I don’t have the book on hand and have only been commenting from memory so far. But to give you a quick answer: I don’t think all of it is shallow, especially given the format: the book is mostly a prose re-writing of the author’s own peer-reviewed anthropology scientific papers. Most of the authors claims are backed by actual papers for reference on the footnotes. As for replication it seems that the author himself replicated some of his studies with different hunter gatherer societies in the world. It’s been a good while since I read it.
I can tell you from my personal experience that the info there has helped me understand the differences between how people think in Brazil (where I come from) and how people think here in the US. Could it be me pattern matching? Possibly
I wouldn’t expect all of it to be true, but I would be very surprised if most of the sources the author provide are false or lack theory and tests, since he explain control groups and experiments in details.
I’m not that married to the book either, as I find some claims rather bold (like the Italy divide)
The title does sound catchy tho
Edit; the author’s main point is how the papal rule on monogamy changed Europe and its colonies to this day, which I didn’t capture on my main comment. Lots to unpack there
> I don’t think all of it is shallow, especially given the format: the book is mostly a prose re-writing of the author’s own peer-reviewed anthropology scientific papers.
That's not the issue. The replication crisis is the phenomenon that many scientific results and conclusions which originate from serious, peer-reviewed research, couldn't be replicated by other researchers, and sometimes not even by the original scientist. This is especially concerning because many results with strong statements – unintuitive ones as well as bias-confirming ones – turned out to be non-existent [1]. It doesn't necessarily have anything to do with "shallowness" or "cocktail-party takes", although the strength of the purported effects, combined with pop-science simplifications and reductions, lend themselves well to such memetically spreading factoids.
[1] The "softer" sciences tend to be much more affected by this than the "harder" sciences.
I know we are not supposed to talk about karma (and therefore a downvote ironically may be appropriate on this comment), but I agree vote behaviour on HN lately shows a still-small but growing tendency toward suppressive downvoting.
This seems to happen generally for two reasons: Even a neutral comment is evaluated for what stance it most closely aligns with, and then downvoted to suppress the opposing view just in case. Or alternatively, a comment that appears low-key combative (but really isn't directly so) gets downvoted in an attempt to ensure harmony.
Both moves to me have "culture war vibes", and come from either adopting those habits or feeling very tired from strife.
I think it's increasingly easy to fall into either bracket, but let's not do that on HN! If a comment is generally polite enough, the only bar to meet is adding new information or new thought into a conversation. None of us come here to be pandered to, and getting challenged by viewpoints that force you to consider the corner cases of your own views is half the fun.
>No one thinks about moving the starting or ending point of the bridge midway through construction.
it is a common rhetorical device to phrase something as an absolute when the negation of it is only an edge case.
hence
>Hillel interviewed a civil engineer who said that they had to move a bridge! Of course, civil engineers don't move bridges as frequently as programmers deal with changes in software but,
I noticed years ago that majority Catholic and Orthodox countries are generally less prosperous than Protestant countries and have higher tolerance for corruption, etc. why is that - perhaps cultural due to difference in theology is a factor?
If only the protestant countries' secret services would stop arming rebels every time they democratically elect someone that wants to stop funneling riches to said protestant countries…
1. The reformation increased literacy/education in the populace to a greater extent in protestant areas, because you no longer needed clergy to talk to God, or understand the bible. Protestant countries have had better education for longer and it has a compounding effect.
2. The "Marriage and Family Program" (the "MFP")... protestant areas discouraged cousin marriage and levirate marriage much earlier than catholic countries, and it is still very common in the rest of the world. Consanguineous marriage is ludicrously prevalent in the middle east, it makes most of the rest of the world more tribal and you end up with compounding genetic defects. By making cousin marriage taboo, it encouraged children to move to a different town and made people less clannish.
I don't think that is a fruitful line of reasoning, given that the vast majority of the world is not Catholic/Orthodox/Protestant! You should consider digging deeper for underlying causes that go beyond localized religions.
Put another way - that would seem to be an effect, not a cause.
> Put another way - that would seem to be an effect, not a cause.
Protestantism does seem to become the preferred variant of Christianity in areas of bottom-up power systems, such as the UK (at least wrt Magna Carta), which does make perfect sense given the Vatican being the ultimate in top-down thinking.
> I noticed years ago that majority Catholic and Orthodox countries are generally less prosperous than Protestant countries
Based on what? I can't think of a single pairing of Protestant and Catholic/Orthodox countries that genuinely had a similar enough history, geography, geopolitical situation to be able to compare them easily. Maybe the closest would be Belgium and Netherlands, but even then Belgium spent a few hundred years more under a faraway empire milking them, and as a frontier for lots of fighting with the French, and occupied by the Germans. And Belgium is also smaller in land and population, and had a very different colonial way of working. It also had resources (coal) that allowed it to industrialise quickly, while the Netherlands didn't so focused mostly on trade and trade posts.
And... Belgium has a slightly worse economy by most classic metrics (GDP, GDP per capita, etc.).
Maybe the only other even remotely comparable countries with different religions are the Baltics (Estonia was Lutheran when they were religious, Latvia is very mixed, Lithuania is mostly Catholic). Estonia and Lithuania have pretty similar GDP per capita, with Latvia a bit behind.
But seeing this through a religious lens is missing the forrest for the trees at best.
The book examines this in multiple different ways, not just at the national level, but even within countries (provinces that are more catholic vs more protestant, and even within Germany, how far the city was from Wittenberg), as well as comparing third world countries that encountered catholic missionaries vs protestant missionaries.
> It feels like we have sometimes accreted an amalgam of these pithy takes based on very small, one off, studies (never replicated) that let us comfortably assemble an affirmation of our broader takes.
The patterns are there and are hard to deny. The reasoning and explanations of these types of books? Don't take them for granted, do your own research if anything is of particular interest, think for yourself, etc. The books can be of value without being 100% correct.
> humans are worst at discerning faces today because we need to discern letters and words and dedicate brain power for that.
I have absolutely nothing to back this up, but my gut tells me this risks being one of those bold claims that grows legs and runs for a while until we debunk it.
Yeah, has that Malcolm Gladwell knowledge porn vibe. A book that empowers its reader with secret knowledge of explanation that all fits together a little too neatly and loses nuance or is often just plain wrong.
I think the “secret explanation” can simply be things you don’t know (or that most people don’t know) because they’re not interesting per se, but when combined they make an interesting whole.
I enjoyed some books that don’t have anything unknown in its parts but that brought a lot of shift in perspectives for me, such as “Man’s Worldly Goods”[1] and “The Drunkard’s Walk”[2].
After finding out even Think Fast, Slow (a book from a very creditable researcher and nobel laureate) is full of replication crisis, I approach pop-sci as entertainment instead of self-education.
*nobel memorial laureate.
This is exactly why people get annoyed with the branding of the bank of Sweden’s economics prize. We have yet to see the prize for chemistry awarded for research that does not reproduce.
As one of the professors I had undergrad classes with liked to say "Economics is the only field where you can be awarded the Nobel prize for showing A and then next year someone gets a Nobel prize for showing not A".
Was it? I thought Millikan's measurement had a minor error from an incorrect viscosity of air, and several other researchers' subsequent measurements were fabricated to agree with Millikan's.
There are people who have face-blindness (inability to discern faces) or dyslexia (visual processing disorder that leads to severe difficulties with reading). The two aren't strongly correlated.
Dyslexia seems to be tied to some broader visual processing issues, which impair the ability to discriminate faces somewhat. But not the other way around.
If the two skills were strongly related, you'd expect a very strong and obvious link. Maybe in form of both performing poorly, if damage to the same pathways impairs both. Or as one performing poorly while another performs unusually well (super-recognizers? children who learn reading at 2?) - if the two skills compete for brain real estate and create a performance tradeoff, as claimed.
I'm struggling to think of any way to test the hypothesis which is (A) practical and (B) accurate.
For example, suppose you sampled a group today and found an inverse-correlation between "good at recognizing many faces" and "good at recognizing written text"... That still wouldn't show that one facility grew causing the other to shrink, because maybe people are just born (or early-development-ed) with a certain bias.
I'm quite willing to believe that human brains do Weird Stuff with respect to reusing circuits and development, but I still don't see how they concluded "the skill encourages the brain state" by disproving "the brain state encourages the acquisition of the skill."
It would be ethically difficult to randomly assigned children to groups (A) taught to read versus (B) forced to remain illiterate while ensuring both groups had the same number of people's faces in their social circles.
IIRC the author compares brain scans and recognition abilities of children of hunter gatherers that where sent to school vs same age relatives that were not. I’m bringing this up from memory now and I’m not so sure of this, but this claim stems from some studies of the author
It seems that there is an study in which the part of the brain used to recognize words is also used for recognizing letters, and when one increases taking more space the other shrink. That study used brain scanners to measure and detect brain activity.
It's not completely insane, the part of the brain that gets used for recognising words is very close to the part of the brain that recognises faces. The brain likely cannibalises the part of the cortex that's used to recognise faces to recognise words and letters instead. See this study[1] where the visual word form area reacts much more strongly to faces in illiterates than in people who have learnt to read.
It's the one that immediately set off my alarm bell. I always try to put myself in the shoes of a scientist and imagine how it would be possible to design a study to test a claim. To me, this one implies humans of today are worse at recognizing faces compared to humans of the past who did not read as much or at all. That one cannot possibly be tested because you cannot test the cognitive capabilities of people of the past who no longer exist.
On the more productive side, this suggests we might develop standardized tests of human capabilities and limits that would allow people of the future to compare themselves to us.
> He explains how monogamy is to blame for a lot of our western views today, and how Mormon towns in Utah were affected by not having monogamy as the basis of society (women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man).
I haven't read the book but it sounds really interesting. Regarding tone though,
> monogamy is to blame for a lot of our western views today
Does the author use the word "blame" to mean "the reason for" or do they present it as a critique of monogamy? Not a big deal, just made me curious when I saw that.
I meant it as “is responsible for” or “explains”. The author doesn’t seem to make any judgement in over the other, but he presents polygamy in a society as a causation for male violence. Sorry for that, English isn’t my first language
> One of the wildest claims in there is the one that the north of Italy is more developed today because it was part of the Holy Roman Empire while the south wasn’t. About a thousand years separate these and he finds effects still.
I would note that the north and south of Italy have very different geography and climate. Which can be upstream of all sorts of things, culturally. The geography of Italy's two halves support different types of economic activity; and the social realities of living within these different economies, naturally evolves into major differences in culture. (Compare/contrast: the differing cultures of coastal vs midwestern America. Now imagine that split with a few thousand more years for the divergence to take hold.)
History happens once; but geography is always affecting a nation, all throughout its evolution. So if you're looking for reasons that two sub-populations within a country might have noticeable differences today, differing geography is going to be the "horse", while history is more of the "zebra."
That being said: geography can also constrain history.
Southern Italy is almost entirely coastline, in a part of the world where, for much of the last ~2000 years, everyone was constantly invading everyone else by sea. Northern Italy was relatively-more immune to amphibious assault, as its capitals could be situated more inland. (Rome itself — the exception that proves the rule — was located in south Italy, but was defended from amphibious assault mostly by the Roman Empire's huge naval home-fleet being docked to the southern-Italian coast; not by anything inherent to its location. Once the Roman Empire itself went away, big rich cities in southern Italy suddenly became juicy targets for conquest and/or sacking.)
Let’s not forget another data point. South was richer before unification than the north.
The north regions regularly at war with France and Austria were pretty much debt fuelled, whilst the south was considered the bank of Italy, solvent and very rich due to flourishing economy.
After unification, Piedmont dumped its war debts on the whole country and drained the south’s cash reserves, using them to modernise the north while the south was left weakened.
Yes, your point and other points around the web I’ve seen make his argument about north and south Italy very controversial to say the least. He does have data to back it up, where he presents distance to nearest church as a predictor for how well a population will fare, and south Italy didn’t have the churches that north the Italy had
Would love to see that link and any other posts others might have run across here as well. I feel like pre covid it was common to see high Quality ask hn discussions with niche/prestigious book recommendations often. I dont see that as much now.
Got to love calvinist societies like northern europe. "oh you are poorer than me due to centuries of colonization? That's your own moral failing!".
And while nowadays they are mostly non religious and don't directly express the idea in these terms, the disparity of treatment is still incredibly strong.
And basically it's all down to having better weapons at a certain point in history.
I haven’t read it but I think that’s genuinely interesting and not obvious.
And probably could change based on the roles of first and second wives and, yes, how male status plays out and how it influences the life of the wife.
We don’t have legal polygamy but in many places there’s not much stopping people from living in an unmarried multi-woman household with a man (or vice versa). But it’s not a very common arrangement, and it’s interesting to think about why.
The first level "why" at least is straightforward: Christianism. Even when not directly imposed, it's still the basis of the Western system of values and morals.
But it's fascinating to think about the second level "why": what made people encode monogamy and heterosexuality into their cultural canons (including their mainstream religion)? Was it property and rules about property? Was it to maximize the number of children, so that the group/tribe/kingdom would be militarily stronger than the neighbor? Or maybe it was to prevent some sort of very specific and concrete problem, real or perceived, that arose from tolerating free love, and that we today have no clue about?
> But it's fascinating to think about the second level "why": what made people encode monogamy and heterosexuality into their cultural canons (including their mainstream religion)?
I dunno about heterosexuality being encoded[1] into cultural canons, but for monogamy it's actually quite simple: violence.
Do you really want half your testosterone-fueled 18-28 year old males unable to attract a mate? There'll be continuous fighting to kill of the excess males.
===========
[1] As far as heterosexuality goes, it's not "encoded by wilful intention" so much as "this is the default". IOW, most people are happy going with the default, so if you make something opt-out, the majority won't opt-out. Same for opt-in. This is why countries that have opt-out organ donors have more organs donated, while countries that have opt-in organ donations have a fraction of he opt-out countries.
Literally one of the main points of the book: the church’s mandate and enforcement of monogamy in Europe lowered violence so much that a different kind of society emerged.
Violence about having an exclusive mate could be a purely cultural construct, reinforced from childhood. I can say that I always, personally, considered the idea of controlling a person abhorrent, as much as I found uninteresting the idea of orbiting my intimate life around a single person with special and very high privileges over whatever I do and think, including outside of bed. However, all my relatives were very insistent that I should date girls and marry. 100% culture, 0% nature. Of course, this is just a single data point.
The heterosexuality part being the "default" is a bit naive, because it ignores the lengths to which some people go to force their kids to be heterosexual. Again, anecdotally, my father sired two gay sons, who had to go to great lengths to have a less traumatic life. Sometimes I suspect my father wasn't that hetero himself, and was only ensuring the next generation inherited his cultural legacy/trauma.
I could consider an argument that a majority (heterosexual) imposes a cultural canon on a non heterosexual minority. But the problem with that is that we don't really know if that majority/minority split would exist without the very strong cultural conditioning. And, as I said before, I don't really believe that homophobia is something the Canaanites invented out of spite. Most likely, it was a cultural trait that conferred advantages to groups, particularly after the agricultural revolution locked human population in a cycle of growth and war for land--but that's just a pet theory of mine.
There are even theories (read the controversial book "Sex at Dawn" if you want the details) that our current cultural canons about sexuality run against what was our nature for hundreds of thousands of years.
> Do you really want half your testosterone-fueled 18-28 year old males unable to attract a mate? There'll be continuous fighting to kill of the excess males.
This is nonsense. Non-monogamy is relinquishing exclusivity. If a man can have multiple women, but a woman can't have multiple men, it's just a different form of oppression.
Monogamy is possessiveness, and possessiveness is what drives violence.
Sounds a lot like Tom Holland - Dominion (to help the robots - "...broad history of the influence of Christianity on the world, focusing on its impact on morality – from its beginnings to the modern day.").
There is undoubtedly a real effect here, but IMHO one problem with the original article is that it treats the US as the only reference point.
On the one hand, this reflects the US dominant position in world affairs and the fact that probably most of the training materials come from there.
But on the other hand, there are some outlier results that are left unexplained. For example, ChatGPT is even more aligned with Japan than with the US.
Industrialised, as opposed to back-breaking labour.
Rich, as opposed to poor.
Democratic, LOL.
The exact type of acronym a self-important academic would create, neatly dividing the world into the good side, and the savages. What is this, the 1850s?
I’ve seen a massive uptick in the use of ‘weird’ as an insult (charitably because all the old insults get you shadowbanned on social media, less charitably because conformism is what the mainstream values more than anything), so the author isn’t even pretending to hide their agenda here.
I think they chose the WEIRD acronym to challenge the western centrism. For most of the readers of the book, the culture described by the WEIRD acronym is not only the normal culture, but is in many ways considered universal.
By calling it WEIRD, the author is trying to drive home the point that the vast majority of people in the world are NOT in that culture that many westerners feel is 'normal', which would make it 'weird' in the sense that it isn't actually the norm.
Now, I have a lot of problems with the book and his arguments, but I don't think there is anything sneaky or nefarious about the word choice, it is very up front and straightforward as to the reasoning behind it.
WEIRD is not pejorative in TFA. There's no problem being WEIRD. I am WEIRD. What's alleged in TFA is that AI, as it's currently deployed, is implicitly chauvinistic towards perspectives other than WEIRD. This sort of thing has historically been a problem with AI/ML and automation in general. The classic example is cameras deploying autofocus features that fail on non-white faces (which has happened several times).
Poorly considered automation can create frictionless experiences for some and Kafkaesque experiences for the rest, where systems refuse to accept your atypical name, your atypical style of speaking is flagged as an indicator of fraud, etc. Automating processes involving people necessarily makes assumptions about those people, and such assumptions are often brittle.
For example, it's easy to imagine a resume filtering AI being implicitly prejudiced against people from Fictionalstan, because it was only trained on a few resumes from Fictionalstan and most of those happened to be classified as "unqualified". This is a danger anytime you have a small number of samples from any particular group, because it's easy for small sample sizes to be overwhelmed by bad luck.
In general I think these types of issues are best viewed as software bugs. It's a clearer and more actionable perspective than as ideological issues. If the software isn't serving some of our end users properly, let's just fix it and move on.
> The classic example is cameras deploying autofocus features that fail on non-white faces (which has happened several times).
I'm a Caucasian living in Asia and the facial recognition systems that they recently required all banking apps to use struggles massively with my face.
I fully agree with you here that this isn't systemic racism, it's just a bug. It only becomes racism if they don't put any effort into fixing it.
When a member of a group says that a term used to refer to them is pejorative, is it now an acceptable response to simply say ‘no it isn’t’? That hasn’t been the case for decades.
Keep in mind that I am also WEIRD. Please help me to understand by referring to the specific part of the article where WEIRD was used in the pejorative.
You know, are you seeing it as pejorative or just want to silence the whole idea? Because people who came up with WEIRD acronym were ... WEIRD. They were literally talking about themselves and about weakness of studies that were made.
The idea that “mainstream” values “conformism” seems like a relic of the 1980s. Have you looked around at public figures in the news? There’s less Debbie Boone and more Dennis Rodman going on. The freaks are flying their flags out there for everyone to see.
I heard a congressman in a town hall meeting last night call a colleague "crazy liberal" - a psychotherapist called in and said don't use that word "crazy" - language is being perverted here
>In fact, this paper found that more than that, it thinks American.
I think that's because it seems to be primarily trained on reddit and therefore mirrors everything reddit stands for. Not a good thing considering just how overrun the site is with bots and political activists of all kinds.
You're absolutely right! Social media like Reddit are overrun with bots, sycophants, and trolls trying to provoke reactions by engaging in controversial topics. This forms echo chambers, which is a sub-par source for training data, and reflects those biases in LLM responses.
“Do non-American LLMs (e.g. DeepSeek, Mistral, Apertus) perform better or worse here? Do they have their own cultural biases in-built?”
I'm wondering the same thing, in addition to the related question of “Would an LLM perform better or worse if prompted with languages other than English?”.
ChatGPT is worse in Russian. Example: after accurately noting that a name appeared in a particular Russian book, it asked if I wanted the direct quote in Russian. I said yes. At this point it switched to Russian output but could no longer find the name in that book, and then apologized for having used what seemed to have been "approximations" about the book before.
(I did then go and check the book myself; ChatGPT in English was right, the name is there)
I was using Qwen3 locally in thinking mode, and noted that even if it is talking to me in Japanese, it is doing it's "thinking" steps in English. Not having a full understanding of how the layers in an LLM handle language connections I can't say for sure, but for a human this would result in subpar outcomes.
For example (not actual output):
Input: "こにちは"(konichwa)
Qwen Thinking: "Ah, the user has said "こにちは", I should respond in a kind and friendly manner.
Qwen Output: こにちは！
It quiiiickly gets confused in this, much quicker than in English.
I'm kind of wondering when will it become a universal understanding that LLMs can't be trained with equal amounts of Japanese and Chinese contents in training data due to Han Unification, making these two languages incoherent mix of two conflicting syntax in one. It's remarkable that Latin languages is not apparently facing issues without clear technical explanation as to why, which I'm guessing has to do with the fact of granularity of characters.
That said, in my tiny experience, LLMs all think in their dataset majority language. They don't adhere to prompt languages, one way or another. Chinese models usually think in either English or Chinese, rarely in cursed mix thereof, and never in Japanese or any of their non-native languages.
Would they not quickly bocome divergent vectors? In the same way that apple and Apple can exist in the same vector set with totally different meanings?
So all information gleaned reading a glyph in the context of japanese articles would be totally different vectors to the information gleaned from the same glyph in Chinese?
Quite a few reasoning LLMs do reasoning in English only. Because the RL setup specifically forces them to do so.
Why?
Because the creators want the reasoning trace to be human readable. And without a pressure forcing them to think in English, they tend to get weird with the reasoning trace. Wild language-mixing, devolved grammar, strange language-mixed nonsense words that the LLM itself seemingly understands just fine.
Ah nah, that was just me here, I'm no good with the phone IME. I tried a bunch of different sentences. It always thought in English.
It was pretty good at one shot translations with thinking turned off however, I imagine thinking distracts it from going down the Japanese only vector paths.
I have saved instructions for Gemini to translate queries into the local language then retranslate the output back to English, when asking about non-English speaking countries/cultures. It seems to work fairly well, but I think it's just due to the different content trained in that language; obviously there would be more in depth discussion of Indonesian cuisine in Indonesian. Whether the country is rich or democratic shouldn't really affect the output.
That's interesting! I manually do the same by prompting in the target language, since it drastically changes the results.
This has been true of web search since forever mind you. The wev has always been culturally delineated by language, and the English Web as I call it is not the only web.
I imagine the culture of HRLF trainers affects things. Maybe there’s disproportionally more of them from Oz/NZ, as native English-speaking countries with possibly lower wages?
I had heard that a lot of human feedback was being provided by people in Nigeria, as it has a very large English speaking population (owing to its history as a British colony) while also having low wages. This was the explanation given as to why ChatGPT seemed to use the word "delve" so often, apparently it's used much more frequently in business contexts there.
Possible that they're using different sources of feedback for different training though
Some of the questions don't really make sense to ask an LLM (being about the survey taker's personal financial situation and such), but the paper doesn't seem to go into detail of what questions were used.
Another paper that echoes similar concerns — AI Suggestions Homogenize Writing Toward Western Styles and Diminish Cultural Nuances (https://arxiv.org/abs/2409.11360)
It seems like almost all contexts might get value from specialized training. People often vary radically depending on where they were raised and where they live, their occupation and social class, and a range of other factors. Even workers from essentially identical backgrounds but practicing different trades can have very different perceptions and framing for what might appear to be shared tasks.
Social media reflects a silicon valley perspective and US domestic news have contaminated the entire eurozone since over a decade.
Movies are also a distillate of a local culture often with a rather uniform ideological slant.
AI being a clone army of corporate spokesmen from the US west coast brings sparsely little new cultural homogeneity to the already very smoothed table.
> The greater the cultural distance between a country and the USA, the less accurate ChatGPT got at simulating peoples’ values.
I'd be somewhat concerned that what is actually reflected is a cultures willingness to adopt US (west coast) values over its own. We see this constantly in some European countries where we're willing to adopt US view points and problems over our own. Either because we're constantly exposed the US problems online or because the US problems are simply more "interesting", in the sense that they are more decisive and easier for us to split into right and wrong.
I don’t have the data but I assume the corpus available to train an LLM is majorly in English, written by Americans and western counterparts. If we’re training the LLMs to sound similar to the training data, I imagine the responses have to match that world view.
My anecdote is that before LLMs I would default to search Google in English instead of my own native language simply because there was so much more content in English to be found that would help me.
And here I am producing novel sentences in English to respond to your message, further continuing the cycle where English is the main language to search and do things.
In my experience, ChatGPT, at least, seems to have had multiple languages used to train its corpus. I am guessing this based on its interaction with me in a different language, where it changed English idioms like "short and sweet" to analogous versions in that language that were not direct translations.
But my guess is that the data sets used from the other languages are smaller (and actually, even if it had perfect access to every single piece of data on the internet, that would still be true, due to the astonishing quantity of English-language data out there compared to the rest. Your comment validates that). With less data, one would expect a poorer performance in all metrics for any non-Anglophone place, including the "cultural world view" metric.
Can’t recommend the book that coined this acronym enough: The WEIRDest People in the World Book by Joseph Henrich.
It is such an eye-opening piece that explains so much of the world around us. He’s an anthropologist that goes into the psychology of it all. Touching on points like how religion plays a part in shaping the America of today and even how humans are worst at discerning faces today because we need to discern letters and words and dedicate brain power for that.
There are so many interest studies mentioned there, one that really stuck with me is how Protestant-raised Americans will work harder for the next day after having (reasearch-led) incestuous thoughts when compared to Catholics and Atheists.
He explains how monogamy is to blame for a lot of our western views today, and how Mormon towns in Utah were affected by not having monogamy as the basis of society (women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man).
One of the wildest claims in there is the one that the north of Italy is more developed today because it was part of the Holy Roman Empire while the south wasn’t. About a thousand years separate these and he finds effects still. Mostly in connection to the spread of read/write to the public being a core tenant of Protestantism.
Anyway, this is not a summary of the book but instead a few points from it that really stuck with me after reading it. Fascinating stuff
I have to ask (and I don't mean this combatively) - given the ongoing realization of the replication crisis, how likely is it that the book you mention reflects a summation of the "too pat" studies about human behavior that, en masse, always seem pithy in an interesting headline, but years later end up being completely bunk?
I've noticed over the years many chains of reasoning - made up of what I believe someone called "cocktail party" pithy takes - that only last as long as you don't dig into the nuts and bolts of them. Pleasant little takes on our psyche and behavior that makes for nice reaffirming thoughts of our views but break down under later analysis.
It feels like we have sometimes accreted an amalgam of these pithy takes based on very small, one off, studies (never replicated) that let us comfortably assemble an affirmation of our broader takes.
This is a rotten thing to say about your book recommendation, given I have never read it (I hope you'll forgive me), but based on the last few years of the replication crisis, do you think, in your heart of hearts, that what you are describing truly does stand up?
I do not have the time now to craft you a full answer as I don’t have the book on hand and have only been commenting from memory so far. But to give you a quick answer: I don’t think all of it is shallow, especially given the format: the book is mostly a prose re-writing of the author’s own peer-reviewed anthropology scientific papers. Most of the authors claims are backed by actual papers for reference on the footnotes. As for replication it seems that the author himself replicated some of his studies with different hunter gatherer societies in the world. It’s been a good while since I read it.
I can tell you from my personal experience that the info there has helped me understand the differences between how people think in Brazil (where I come from) and how people think here in the US. Could it be me pattern matching? Possibly
I wouldn’t expect all of it to be true, but I would be very surprised if most of the sources the author provide are false or lack theory and tests, since he explain control groups and experiments in details.
I’m not that married to the book either, as I find some claims rather bold (like the Italy divide)
The title does sound catchy tho
Edit; the author’s main point is how the papal rule on monogamy changed Europe and its colonies to this day, which I didn’t capture on my main comment. Lots to unpack there
> I don’t think all of it is shallow, especially given the format: the book is mostly a prose re-writing of the author’s own peer-reviewed anthropology scientific papers.
That's not the issue. The replication crisis is the phenomenon that many scientific results and conclusions which originate from serious, peer-reviewed research, couldn't be replicated by other researchers, and sometimes not even by the original scientist. This is especially concerning because many results with strong statements – unintuitive ones as well as bias-confirming ones – turned out to be non-existent [1]. It doesn't necessarily have anything to do with "shallowness" or "cocktail-party takes", although the strength of the purported effects, combined with pop-science simplifications and reductions, lend themselves well to such memetically spreading factoids.
[1] The "softer" sciences tend to be much more affected by this than the "harder" sciences.
No idea why this is getting downvoted. Everything I wrote is true and directly on topic. Further reading: https://en.wikipedia.org/wiki/Replication_crisis
I know we are not supposed to talk about karma (and therefore a downvote ironically may be appropriate on this comment), but I agree vote behaviour on HN lately shows a still-small but growing tendency toward suppressive downvoting.
This seems to happen generally for two reasons: Even a neutral comment is evaluated for what stance it most closely aligns with, and then downvoted to suppress the opposing view just in case. Or alternatively, a comment that appears low-key combative (but really isn't directly so) gets downvoted in an attempt to ensure harmony.
Both moves to me have "culture war vibes", and come from either adopting those habits or feeling very tired from strife.
I think it's increasingly easy to fall into either bracket, but let's not do that on HN! If a comment is generally polite enough, the only bar to meet is adding new information or new thought into a conversation. None of us come here to be pandered to, and getting challenged by viewpoints that force you to consider the corner cases of your own views is half the fun.
https://danluu.com/cocktail-ideas/
Here are the cocktail ideas. Hits the spot.
>No one thinks about moving the starting or ending point of the bridge midway through construction.
it is a common rhetorical device to phrase something as an absolute when the negation of it is only an edge case.
hence
>Hillel interviewed a civil engineer who said that they had to move a bridge! Of course, civil engineers don't move bridges as frequently as programmers deal with changes in software but,
The book has been recommended and discussed here before.
I noticed years ago that majority Catholic and Orthodox countries are generally less prosperous than Protestant countries and have higher tolerance for corruption, etc. why is that - perhaps cultural due to difference in theology is a factor?
If only the protestant countries' secret services would stop arming rebels every time they democratically elect someone that wants to stop funneling riches to said protestant countries…
The book suggests two main reasons for this.
1. The reformation increased literacy/education in the populace to a greater extent in protestant areas, because you no longer needed clergy to talk to God, or understand the bible. Protestant countries have had better education for longer and it has a compounding effect.
2. The "Marriage and Family Program" (the "MFP")... protestant areas discouraged cousin marriage and levirate marriage much earlier than catholic countries, and it is still very common in the rest of the world. Consanguineous marriage is ludicrously prevalent in the middle east, it makes most of the rest of the world more tribal and you end up with compounding genetic defects. By making cousin marriage taboo, it encouraged children to move to a different town and made people less clannish.
I don't think that is a fruitful line of reasoning, given that the vast majority of the world is not Catholic/Orthodox/Protestant! You should consider digging deeper for underlying causes that go beyond localized religions.
Put another way - that would seem to be an effect, not a cause.
> Put another way - that would seem to be an effect, not a cause.
Protestantism does seem to become the preferred variant of Christianity in areas of bottom-up power systems, such as the UK (at least wrt Magna Carta), which does make perfect sense given the Vatican being the ultimate in top-down thinking.
> I noticed years ago that majority Catholic and Orthodox countries are generally less prosperous than Protestant countries
Based on what? I can't think of a single pairing of Protestant and Catholic/Orthodox countries that genuinely had a similar enough history, geography, geopolitical situation to be able to compare them easily. Maybe the closest would be Belgium and Netherlands, but even then Belgium spent a few hundred years more under a faraway empire milking them, and as a frontier for lots of fighting with the French, and occupied by the Germans. And Belgium is also smaller in land and population, and had a very different colonial way of working. It also had resources (coal) that allowed it to industrialise quickly, while the Netherlands didn't so focused mostly on trade and trade posts.
And... Belgium has a slightly worse economy by most classic metrics (GDP, GDP per capita, etc.).
Maybe the only other even remotely comparable countries with different religions are the Baltics (Estonia was Lutheran when they were religious, Latvia is very mixed, Lithuania is mostly Catholic). Estonia and Lithuania have pretty similar GDP per capita, with Latvia a bit behind.
But seeing this through a religious lens is missing the forrest for the trees at best.
The book examines this in multiple different ways, not just at the national level, but even within countries (provinces that are more catholic vs more protestant, and even within Germany, how far the city was from Wittenberg), as well as comparing third world countries that encountered catholic missionaries vs protestant missionaries.
The cause is the Spanish Armada did not conquer England.
> It feels like we have sometimes accreted an amalgam of these pithy takes based on very small, one off, studies (never replicated) that let us comfortably assemble an affirmation of our broader takes.
The patterns are there and are hard to deny. The reasoning and explanations of these types of books? Don't take them for granted, do your own research if anything is of particular interest, think for yourself, etc. The books can be of value without being 100% correct.
> humans are worst at discerning faces today because we need to discern letters and words and dedicate brain power for that.
I have absolutely nothing to back this up, but my gut tells me this risks being one of those bold claims that grows legs and runs for a while until we debunk it.
Yeah, has that Malcolm Gladwell knowledge porn vibe. A book that empowers its reader with secret knowledge of explanation that all fits together a little too neatly and loses nuance or is often just plain wrong.
Are you saying the theory of igon values is not so universal after all?
I think the “secret explanation” can simply be things you don’t know (or that most people don’t know) because they’re not interesting per se, but when combined they make an interesting whole.
I enjoyed some books that don’t have anything unknown in its parts but that brought a lot of shift in perspectives for me, such as “Man’s Worldly Goods”[1] and “The Drunkard’s Walk”[2].
[1] https://www.amazon.com/Mans-Worldly-Goods-Wealth-Nations/dp/... [2] https://www.amazon.com/Drunkards-Walk-Randomness-Rules-Lives...
After finding out even Think Fast, Slow (a book from a very creditable researcher and nobel laureate) is full of replication crisis, I approach pop-sci as entertainment instead of self-education.
*nobel memorial laureate. This is exactly why people get annoyed with the branding of the bank of Sweden’s economics prize. We have yet to see the prize for chemistry awarded for research that does not reproduce.
As one of the professors I had undergrad classes with liked to say "Economics is the only field where you can be awarded the Nobel prize for showing A and then next year someone gets a Nobel prize for showing not A".
The peace prize is awarded to warmongers all the time though.
Yeah, well, you know, that's just, like, your opinion, man.
You can find, maybe, three or four such recipients out of 100. And they usually did make peace, even if they previously or later made mistakes.
Obama.
> We have yet to see the prize for chemistry awarded for research that does not reproduce.
Maybe, but e.g. Millikan's prize for physics was on the basis of results that appear to have been at least partially fabricated.
Was it? I thought Millikan's measurement had a minor error from an incorrect viscosity of air, and several other researchers' subsequent measurements were fabricated to agree with Millikan's.
I suppose to be fair to the field of economics, the replication issues were mainly with research in psychology (as I recall).
On the other hand, does economics have less of a replication issue because it’s basically unreplicable?
There are people who have face-blindness (inability to discern faces) or dyslexia (visual processing disorder that leads to severe difficulties with reading). The two aren't strongly correlated.
Dyslexia seems to be tied to some broader visual processing issues, which impair the ability to discriminate faces somewhat. But not the other way around.
If the two skills were strongly related, you'd expect a very strong and obvious link. Maybe in form of both performing poorly, if damage to the same pathways impairs both. Or as one performing poorly while another performs unusually well (super-recognizers? children who learn reading at 2?) - if the two skills compete for brain real estate and create a performance tradeoff, as claimed.
Or perhaps something that it's true even after we "debunked" it
I'm struggling to think of any way to test the hypothesis which is (A) practical and (B) accurate.
For example, suppose you sampled a group today and found an inverse-correlation between "good at recognizing many faces" and "good at recognizing written text"... That still wouldn't show that one facility grew causing the other to shrink, because maybe people are just born (or early-development-ed) with a certain bias.
The author uses mostly hunter-gatherers tribes/societies in different continents as the control groups, usually. Most of his work is in anthropology.
The part of the brain that recognises faces quite literally shrinks in literates compared to illiterates
https://www.unicog.org/publications/1-s2.0-S1364661311000738...
I'm quite willing to believe that human brains do Weird Stuff with respect to reusing circuits and development, but I still don't see how they concluded "the skill encourages the brain state" by disproving "the brain state encourages the acquisition of the skill."
It would be ethically difficult to randomly assigned children to groups (A) taught to read versus (B) forced to remain illiterate while ensuring both groups had the same number of people's faces in their social circles.
IIRC the author compares brain scans and recognition abilities of children of hunter gatherers that where sent to school vs same age relatives that were not. I’m bringing this up from memory now and I’m not so sure of this, but this claim stems from some studies of the author
It seems that there is an study in which the part of the brain used to recognize words is also used for recognizing letters, and when one increases taking more space the other shrink. That study used brain scanners to measure and detect brain activity.
It's not completely insane, the part of the brain that gets used for recognising words is very close to the part of the brain that recognises faces. The brain likely cannibalises the part of the cortex that's used to recognise faces to recognise words and letters instead. See this study[1] where the visual word form area reacts much more strongly to faces in illiterates than in people who have learnt to read.
[1] https://www.unicog.org/publications/1-s2.0-S1364661311000738...
https://en.wikipedia.org/wiki/Visual_word_form_area
https://en.wikipedia.org/wiki/Fusiform_face_area
It's the one that immediately set off my alarm bell. I always try to put myself in the shoes of a scientist and imagine how it would be possible to design a study to test a claim. To me, this one implies humans of today are worse at recognizing faces compared to humans of the past who did not read as much or at all. That one cannot possibly be tested because you cannot test the cognitive capabilities of people of the past who no longer exist.
On the more productive side, this suggests we might develop standardized tests of human capabilities and limits that would allow people of the future to compare themselves to us.
There are people who are illiterate today who you can test.
> He explains how monogamy is to blame for a lot of our western views today, and how Mormon towns in Utah were affected by not having monogamy as the basis of society (women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man).
Must suck to be a lower-ranking man lol
> women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man
I bet you think "2nd" means "secondary."
I bet the parties to the marriage think "2nd" means "most recently allocated."
I haven't read the book but it sounds really interesting. Regarding tone though,
> monogamy is to blame for a lot of our western views today
Does the author use the word "blame" to mean "the reason for" or do they present it as a critique of monogamy? Not a big deal, just made me curious when I saw that.
I meant it as “is responsible for” or “explains”. The author doesn’t seem to make any judgement in over the other, but he presents polygamy in a society as a causation for male violence. Sorry for that, English isn’t my first language
> One of the wildest claims in there is the one that the north of Italy is more developed today because it was part of the Holy Roman Empire while the south wasn’t. About a thousand years separate these and he finds effects still.
I would note that the north and south of Italy have very different geography and climate. Which can be upstream of all sorts of things, culturally. The geography of Italy's two halves support different types of economic activity; and the social realities of living within these different economies, naturally evolves into major differences in culture. (Compare/contrast: the differing cultures of coastal vs midwestern America. Now imagine that split with a few thousand more years for the divergence to take hold.)
History happens once; but geography is always affecting a nation, all throughout its evolution. So if you're looking for reasons that two sub-populations within a country might have noticeable differences today, differing geography is going to be the "horse", while history is more of the "zebra."
That being said: geography can also constrain history.
Southern Italy is almost entirely coastline, in a part of the world where, for much of the last ~2000 years, everyone was constantly invading everyone else by sea. Northern Italy was relatively-more immune to amphibious assault, as its capitals could be situated more inland. (Rome itself — the exception that proves the rule — was located in south Italy, but was defended from amphibious assault mostly by the Roman Empire's huge naval home-fleet being docked to the southern-Italian coast; not by anything inherent to its location. Once the Roman Empire itself went away, big rich cities in southern Italy suddenly became juicy targets for conquest and/or sacking.)
Let’s not forget another data point. South was richer before unification than the north. The north regions regularly at war with France and Austria were pretty much debt fuelled, whilst the south was considered the bank of Italy, solvent and very rich due to flourishing economy. After unification, Piedmont dumped its war debts on the whole country and drained the south’s cash reserves, using them to modernise the north while the south was left weakened.
Also that Italy has only been a single country for a couple hundred years [0], so there will obviously still be regional differences.
[0] https://en.wikipedia.org/wiki/Unification_of_Italy
Yes, your point and other points around the web I’ve seen make his argument about north and south Italy very controversial to say the least. He does have data to back it up, where he presents distance to nearest church as a predictor for how well a population will fare, and south Italy didn’t have the churches that north the Italy had
This would only work as evidence if church placement was random. Could be a correlation with a different cause.
All the basketball players are tall. Ergo, if we want to be taller, we should all play basketball.
Yes, maybe I was not clear but I’m in doubt of the Italy claim in the book and that’s why I remember it
One of my greatest pleasure of random walking the internet is building my list of possible next books to read.. thank you for this one!
Do you have some recommendations? I remember seeing Show HN tools gathering book recommendations from here, but I like specific experiences
yeah I just ordered a copy of "We Survived the Night" based on a post here. Never would have heard of it otherwise.
Mind linking to the post for the curious?
Would love to see that link and any other posts others might have run across here as well. I feel like pre covid it was common to see high Quality ask hn discussions with niche/prestigious book recommendations often. I dont see that as much now.
The acronym was coined in a paper with the same title about 10 years before the book.
Anthropologists have a couple centuries to go before they earn any credibility.
Not sure how this hasn't come up in this thread yet, but the idea that Protestantism leads to more productive societies has been around for quite a while. https://en.wikipedia.org/wiki/The_Protestant_Ethic_and_the_S...
Got to love calvinist societies like northern europe. "oh you are poorer than me due to centuries of colonization? That's your own moral failing!".
And while nowadays they are mostly non religious and don't directly express the idea in these terms, the disparity of treatment is still incredibly strong.
And basically it's all down to having better weapons at a certain point in history.
> women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man
This seems like a bit of a tautology; how are they 'ranking' these men?
> > women there tend to prefer to be 2nd wives of a better man rather being the only wife of a lower-ranking man
> This seems like a bit of a tautology; how are they 'ranking' these men?
I am not seeing the tautology. Can you explain?
I haven’t read it but I think that’s genuinely interesting and not obvious.
And probably could change based on the roles of first and second wives and, yes, how male status plays out and how it influences the life of the wife.
We don’t have legal polygamy but in many places there’s not much stopping people from living in an unmarried multi-woman household with a man (or vice versa). But it’s not a very common arrangement, and it’s interesting to think about why.
The first level "why" at least is straightforward: Christianism. Even when not directly imposed, it's still the basis of the Western system of values and morals.
But it's fascinating to think about the second level "why": what made people encode monogamy and heterosexuality into their cultural canons (including their mainstream religion)? Was it property and rules about property? Was it to maximize the number of children, so that the group/tribe/kingdom would be militarily stronger than the neighbor? Or maybe it was to prevent some sort of very specific and concrete problem, real or perceived, that arose from tolerating free love, and that we today have no clue about?
> But it's fascinating to think about the second level "why": what made people encode monogamy and heterosexuality into their cultural canons (including their mainstream religion)?
I dunno about heterosexuality being encoded[1] into cultural canons, but for monogamy it's actually quite simple: violence.
Do you really want half your testosterone-fueled 18-28 year old males unable to attract a mate? There'll be continuous fighting to kill of the excess males.
===========
[1] As far as heterosexuality goes, it's not "encoded by wilful intention" so much as "this is the default". IOW, most people are happy going with the default, so if you make something opt-out, the majority won't opt-out. Same for opt-in. This is why countries that have opt-out organ donors have more organs donated, while countries that have opt-in organ donations have a fraction of he opt-out countries.
Defaults matter.
Literally one of the main points of the book: the church’s mandate and enforcement of monogamy in Europe lowered violence so much that a different kind of society emerged.
You may be taking things from the wrong end.
Violence about having an exclusive mate could be a purely cultural construct, reinforced from childhood. I can say that I always, personally, considered the idea of controlling a person abhorrent, as much as I found uninteresting the idea of orbiting my intimate life around a single person with special and very high privileges over whatever I do and think, including outside of bed. However, all my relatives were very insistent that I should date girls and marry. 100% culture, 0% nature. Of course, this is just a single data point.
The heterosexuality part being the "default" is a bit naive, because it ignores the lengths to which some people go to force their kids to be heterosexual. Again, anecdotally, my father sired two gay sons, who had to go to great lengths to have a less traumatic life. Sometimes I suspect my father wasn't that hetero himself, and was only ensuring the next generation inherited his cultural legacy/trauma.
I could consider an argument that a majority (heterosexual) imposes a cultural canon on a non heterosexual minority. But the problem with that is that we don't really know if that majority/minority split would exist without the very strong cultural conditioning. And, as I said before, I don't really believe that homophobia is something the Canaanites invented out of spite. Most likely, it was a cultural trait that conferred advantages to groups, particularly after the agricultural revolution locked human population in a cycle of growth and war for land--but that's just a pet theory of mine.
There are even theories (read the controversial book "Sex at Dawn" if you want the details) that our current cultural canons about sexuality run against what was our nature for hundreds of thousands of years.
> Do you really want half your testosterone-fueled 18-28 year old males unable to attract a mate? There'll be continuous fighting to kill of the excess males.
This is nonsense. Non-monogamy is relinquishing exclusivity. If a man can have multiple women, but a woman can't have multiple men, it's just a different form of oppression.
Monogamy is possessiveness, and possessiveness is what drives violence.
Sounds a lot like Tom Holland - Dominion (to help the robots - "...broad history of the influence of Christianity on the world, focusing on its impact on morality – from its beginnings to the modern day.").
> This seems like a bit of a tautology
I don't think it is? A priori it's not at all obvious which option women would be expected to prefer.
a woman can rank you in sub-second time, just as you can her.
There is undoubtedly a real effect here, but IMHO one problem with the original article is that it treats the US as the only reference point.
On the one hand, this reflects the US dominant position in world affairs and the fact that probably most of the training materials come from there.
But on the other hand, there are some outlier results that are left unexplained. For example, ChatGPT is even more aligned with Japan than with the US.
> For example, ChatGPT is even more aligned with Japan than with the US
I have no expertise in this field.
Is it actually even more aligned? Or is it simply aligned with the elements of Japanese culture and/or media that are exported to the West?
By the same graph, Japan is rated as more culturally distant from the US than Singapore, Russia, and Zimbabwe.
I think the comment is basing it on the graphic from the article.
WEIRD here stands for "Western, Educated, Industrialised, Rich, Democratic".
Educated, as opposed to ignorant.
Industrialised, as opposed to back-breaking labour.
Rich, as opposed to poor.
Democratic, LOL.
The exact type of acronym a self-important academic would create, neatly dividing the world into the good side, and the savages. What is this, the 1850s?
I’ve seen a massive uptick in the use of ‘weird’ as an insult (charitably because all the old insults get you shadowbanned on social media, less charitably because conformism is what the mainstream values more than anything), so the author isn’t even pretending to hide their agenda here.
I think they chose the WEIRD acronym to challenge the western centrism. For most of the readers of the book, the culture described by the WEIRD acronym is not only the normal culture, but is in many ways considered universal.
By calling it WEIRD, the author is trying to drive home the point that the vast majority of people in the world are NOT in that culture that many westerners feel is 'normal', which would make it 'weird' in the sense that it isn't actually the norm.
Now, I have a lot of problems with the book and his arguments, but I don't think there is anything sneaky or nefarious about the word choice, it is very up front and straightforward as to the reasoning behind it.
WEIRD is not pejorative in TFA. There's no problem being WEIRD. I am WEIRD. What's alleged in TFA is that AI, as it's currently deployed, is implicitly chauvinistic towards perspectives other than WEIRD. This sort of thing has historically been a problem with AI/ML and automation in general. The classic example is cameras deploying autofocus features that fail on non-white faces (which has happened several times).
Poorly considered automation can create frictionless experiences for some and Kafkaesque experiences for the rest, where systems refuse to accept your atypical name, your atypical style of speaking is flagged as an indicator of fraud, etc. Automating processes involving people necessarily makes assumptions about those people, and such assumptions are often brittle.
For example, it's easy to imagine a resume filtering AI being implicitly prejudiced against people from Fictionalstan, because it was only trained on a few resumes from Fictionalstan and most of those happened to be classified as "unqualified". This is a danger anytime you have a small number of samples from any particular group, because it's easy for small sample sizes to be overwhelmed by bad luck.
In general I think these types of issues are best viewed as software bugs. It's a clearer and more actionable perspective than as ideological issues. If the software isn't serving some of our end users properly, let's just fix it and move on.
> The classic example is cameras deploying autofocus features that fail on non-white faces (which has happened several times).
I'm a Caucasian living in Asia and the facial recognition systems that they recently required all banking apps to use struggles massively with my face.
I fully agree with you here that this isn't systemic racism, it's just a bug. It only becomes racism if they don't put any effort into fixing it.
When a member of a group says that a term used to refer to them is pejorative, is it now an acceptable response to simply say ‘no it isn’t’? That hasn’t been the case for decades.
Keep in mind that I am also WEIRD. Please help me to understand by referring to the specific part of the article where WEIRD was used in the pejorative.
Are you saying that self-hating westerners aren’t real? It’s an entire industry at this point.
Another acceptable response is "take a break, go touch grass".
There is too much victim mentality going on nowadays, and we continue rewarding that behavior so I am not surprised.
You know, are you seeing it as pejorative or just want to silence the whole idea? Because people who came up with WEIRD acronym were ... WEIRD. They were literally talking about themselves and about weakness of studies that were made.
The idea that “mainstream” values “conformism” seems like a relic of the 1980s. Have you looked around at public figures in the news? There’s less Debbie Boone and more Dennis Rodman going on. The freaks are flying their flags out there for everyone to see.
There are no ‘freaks’ flying flags. There are highly polarised groups signalling their in-group identity, each backed by their own billionaires.
Could you explain the agenda to those of us like me, who missed it anyway?
It’s pretty simple. ‘West => weird = bad.’
I heard a congressman in a town hall meeting last night call a colleague "crazy liberal" - a psychotherapist called in and said don't use that word "crazy" - language is being perverted here
That is crazy.
>In fact, this paper found that more than that, it thinks American.
I think that's because it seems to be primarily trained on reddit and therefore mirrors everything reddit stands for. Not a good thing considering just how overrun the site is with bots and political activists of all kinds.
You're absolutely right! Social media like Reddit are overrun with bots, sycophants, and trolls trying to provoke reactions by engaging in controversial topics. This forms echo chambers, which is a sub-par source for training data, and reflects those biases in LLM responses.
The headline should retain the caps, since WEIRD here is not the same as regular weird.
“Do non-American LLMs (e.g. DeepSeek, Mistral, Apertus) perform better or worse here? Do they have their own cultural biases in-built?”
I'm wondering the same thing, in addition to the related question of “Would an LLM perform better or worse if prompted with languages other than English?”.
ChatGPT is worse in Russian. Example: after accurately noting that a name appeared in a particular Russian book, it asked if I wanted the direct quote in Russian. I said yes. At this point it switched to Russian output but could no longer find the name in that book, and then apologized for having used what seemed to have been "approximations" about the book before.
(I did then go and check the book myself; ChatGPT in English was right, the name is there)
I was using Qwen3 locally in thinking mode, and noted that even if it is talking to me in Japanese, it is doing it's "thinking" steps in English. Not having a full understanding of how the layers in an LLM handle language connections I can't say for sure, but for a human this would result in subpar outcomes.
For example (not actual output):
Input: "こにちは"(konichwa) Qwen Thinking: "Ah, the user has said "こにちは", I should respond in a kind and friendly manner.
Qwen Output: こにちは！
It quiiiickly gets confused in this, much quicker than in English.
I'm kind of wondering when will it become a universal understanding that LLMs can't be trained with equal amounts of Japanese and Chinese contents in training data due to Han Unification, making these two languages incoherent mix of two conflicting syntax in one. It's remarkable that Latin languages is not apparently facing issues without clear technical explanation as to why, which I'm guessing has to do with the fact of granularity of characters.
That said, in my tiny experience, LLMs all think in their dataset majority language. They don't adhere to prompt languages, one way or another. Chinese models usually think in either English or Chinese, rarely in cursed mix thereof, and never in Japanese or any of their non-native languages.
Would they not quickly bocome divergent vectors? In the same way that apple and Apple can exist in the same vector set with totally different meanings?
So all information gleaned reading a glyph in the context of japanese articles would be totally different vectors to the information gleaned from the same glyph in Chinese?
Why would that be an issue?
Quite a few reasoning LLMs do reasoning in English only. Because the RL setup specifically forces them to do so.
Why?
Because the creators want the reasoning trace to be human readable. And without a pressure forcing them to think in English, they tend to get weird with the reasoning trace. Wild language-mixing, devolved grammar, strange language-mixed nonsense words that the LLM itself seemingly understands just fine.
Perhaps it knows most users who misspell こんにちは are English speakers?
Ah nah, that was just me here, I'm no good with the phone IME. I tried a bunch of different sentences. It always thought in English.
It was pretty good at one shot translations with thinking turned off however, I imagine thinking distracts it from going down the Japanese only vector paths.
I assume the training dataset is mostly the same anyway. I imagine prompting in different language could have a huge effect though.
[dead]
I have saved instructions for Gemini to translate queries into the local language then retranslate the output back to English, when asking about non-English speaking countries/cultures. It seems to work fairly well, but I think it's just due to the different content trained in that language; obviously there would be more in depth discussion of Indonesian cuisine in Indonesian. Whether the country is rich or democratic shouldn't really affect the output.
That's interesting! I manually do the same by prompting in the target language, since it drastically changes the results.
This has been true of web search since forever mind you. The wev has always been culturally delineated by language, and the English Web as I call it is not the only web.
Interesting that the responses from ChatGPT on the World Values Survey correlated most closely with the responses from Australians and New Zealanders.
I expect (as TFA says) that they would most closely align with Californians, but that isn't in the data.
I imagine the culture of HRLF trainers affects things. Maybe there’s disproportionally more of them from Oz/NZ, as native English-speaking countries with possibly lower wages?
I had heard that a lot of human feedback was being provided by people in Nigeria, as it has a very large English speaking population (owing to its history as a British colony) while also having low wages. This was the explanation given as to why ChatGPT seemed to use the word "delve" so often, apparently it's used much more frequently in business contexts there.
Possible that they're using different sources of feedback for different training though
I wonder how ChatGPT and the like would do if you asked it to give a response as if they were a person from one of those other cultures.
In other words, is the issue in the defaults or is it impossible for AI to respond from other cultures?
For those interested, you can see the World Values Survey Questionnaire here: https://www.worldvaluessurvey.org/WVSDocumentationWV7.jsp
(For Wave 7 (2017-2022), which the paper used)
Some of the questions don't really make sense to ask an LLM (being about the survey taker's personal financial situation and such), but the paper doesn't seem to go into detail of what questions were used.
Oh it’s an arcronym. I was very confused for a good portion of the article
Another paper that echoes similar concerns — AI Suggestions Homogenize Writing Toward Western Styles and Diminish Cultural Nuances (https://arxiv.org/abs/2409.11360)
It seems like almost all contexts might get value from specialized training. People often vary radically depending on where they were raised and where they live, their occupation and social class, and a range of other factors. Even workers from essentially identical backgrounds but practicing different trades can have very different perceptions and framing for what might appear to be shared tasks.
Social media reflects a silicon valley perspective and US domestic news have contaminated the entire eurozone since over a decade.
Movies are also a distillate of a local culture often with a rather uniform ideological slant.
AI being a clone army of corporate spokesmen from the US west coast brings sparsely little new cultural homogeneity to the already very smoothed table.
> The greater the cultural distance between a country and the USA, the less accurate ChatGPT got at simulating peoples’ values.
I'd be somewhat concerned that what is actually reflected is a cultures willingness to adopt US (west coast) values over its own. We see this constantly in some European countries where we're willing to adopt US view points and problems over our own. Either because we're constantly exposed the US problems online or because the US problems are simply more "interesting", in the sense that they are more decisive and easier for us to split into right and wrong.
Ummm... doesn't the AI have to scrape the data of those non- WEIRD cultures to work then? What am I missing here?
There are parts of the world where constant person-electronic connection isn't a thing. Is that your point?
I don’t have the data but I assume the corpus available to train an LLM is majorly in English, written by Americans and western counterparts. If we’re training the LLMs to sound similar to the training data, I imagine the responses have to match that world view.
My anecdote is that before LLMs I would default to search Google in English instead of my own native language simply because there was so much more content in English to be found that would help me.
And here I am producing novel sentences in English to respond to your message, further continuing the cycle where English is the main language to search and do things.
And the RHLF was directed by Californians, and so the "values" are likely very California.
In my experience, ChatGPT, at least, seems to have had multiple languages used to train its corpus. I am guessing this based on its interaction with me in a different language, where it changed English idioms like "short and sweet" to analogous versions in that language that were not direct translations.
But my guess is that the data sets used from the other languages are smaller (and actually, even if it had perfect access to every single piece of data on the internet, that would still be true, due to the astonishing quantity of English-language data out there compared to the rest. Your comment validates that). With less data, one would expect a poorer performance in all metrics for any non-Anglophone place, including the "cultural world view" metric.
english is the lingua franca ;-)
“Fancy autocomplete better at completing documents similar to ones it has seen before” isn’t as headline-worthy.
Well said. It's so odd that this isn't just the general societal talking point when it comes to LLM's
[flagged]