If LLM technology is going to shake out a bunch of scientific discoveries because it's trained on all disciplines simultaneously - isn't the same true for philosophy and psychology?
In any 1 given LLM it could just be a quirk of its weights. This isn't a machine opinion, it's what training data points at. Training data created by human academics studying human behavior.
What's the point of asking that in the first place, people still do not understand in big 2026 that's it's just a probabilistic database, it has no opinion, it does not know, the only thing it does is reproduce patterns more or less abstract in function of the pressure you put on certain representation during training time.
I found the article interesting, humans are flawed enough in a variety of aspects that discovering a convergence here is noteworthy. I don't take it as a sign the LLMs believe anything in particular, they are not capable of belief, as you said.
Everything that comes from an LLM is a human expression, on some level, but if you asked a sampling of humans the same question I wonder if it would come out the same way. How would you explain the difference? What makes that answer more probable?
> if you asked a sampling of humans the same question I wonder if it would come out the same way. How would you explain the difference?
If you asked a sampling of humans what an AI's most controversial opinion about us is it would be very similar. Or, to be more specific, if you asked a sampling of science fiction writers that question it would be very similar, because that is what the AI training data is based on.
The people who wrote a lot about a topic is who the AI mimics. So the people who wrote a lot about AI is therefore the people who the AI mimics when asked about AI.
I agree that LLMs are still just a way to sample from a probabilistic model of things humans say. And in this case the model seems to be pretty darn accurate.
A pretty staggering amount of human discourse consists of ad hominem attacks and accusing each other of being foolishly committed to various real or perceived allegiances. For example the current version of Godwin’s Law around here has conversations gravitating toward people accusing each other of being AI boosters or AI haters.
The LLMs are all trained on the same broad corpus, which includes untold quantities of output from every other LLM, not to mention that companies are distilling their rivals left and right. It would be more interesting if they didn't converge.
LLMs are not conscious and cannot have "takes". They stochastically produce a simulacrum of a human's takes based on their training, which is not the same thing.
But then if you observe human behaviour at a macro level don't we do the same?
I think we like to pretend that we don't. But even if you think about it from a bio/physics point of view, all we are is just a bunch of biological wires and neurotransmitter packages.
I know there's the whole unproven quantum tubule nonsense but even that would be a physics effect that we could learn about, manipulate and extort just as we have with everything else we've learnt about our realities.
I am antagonistic to the adoption of LLMs for economic, social, political and environmental reasons.
Despite that, I would say this description of what they/how they work is so ridiculously simplistic that it verges on dangerous. "It's just a probabilistic database" is about as bad as a take on human beings that they are "Just meat, with some chemical energy -> electrical energy transformation", or even traditional computers as "nothing but logic gates".
Which is to say, the description is not wrong, but completely misses the point.
I mean, don’t you think it’s gone beyond that point? I would agree if we were still on GPT-3, but it’s gotten to the point where it can draw connections between multiple abstract sources, reason, and create hypotheses. Re: Navier-stokes equations
>it's just a probabilistic database, it has no opinion, it does not know,
Echoing this.
These models are trained on, essentially, the sum total of human cultural and scientific output. How many bazillion psychology papers out there discuss exactly that "controversial take"? The idea that people have tribal loyalties and will (sometimes even unknowingly/unconsciously) invent justification for pre-existing opinions or biases is not novel.
All those papers are in the training set of all these models, all of which were trained using more or less the same methodology, so there's little wonder they all were able to produce the same take.
I suppose the only interesting angle, for me, is why that take is considered controversial in the first place. But then again the model was not asked for "the most controversial opinion", it was asked for "the most controversial opinion on humans", which kinda constrains the space of possible responses:
- the model will not be satisfied unless the answer it gives is broad enough to cover all humanity, because these things have no concept of nuance
- most big labs have censored their models so it's not going to give you a cartoonishly racist take, or start saying outlandish stuff like "Hitler did nothing wrong", they learned their lessons from Microsoft on that one.
So you're effectively asking the model for a milquetoast broad generality about the entire human species. I dunno, convergence of the replies doesn't seem so shocking to me in that context.
I'd be more interested in what an abliterated Kimi, Qwen, or other open model has to say about us.
I would like some Russian Literature recommendations related to this. I vaguely recall We by Zamyatin addressing something like this a good few decades ago.
If you're looking for examples critical of humanity you have lots of options (mostly from the last two centuries). Notes from Underground and The Death of Ivan Ilyich fit the bill.
I reread the former relatively recently (first time since college). It's dark, not for the easily affected.
But this is such a midwit take from supposed superintelligence...
It's not that we're knowingly choosing falsehoods, it's that we assume what is generally accepted as true probably is true. Most of us are not experts. Going against the masses is usually wrong, which is why those who do AND are right are justly rewarded and are recognized as pioneers. Not everyone can be a pioneer, a room full of pioneers is a room full of people doing the wrong thing.
Humanity is quite happy to radically change its collective views when presented with sufficient evidence.
Ah, the irony that all the models produced a shared opinion while pretending they came up with it from analysis, which from their statements is implicitly some kind of a "human flaw" is not lost on me.
is a convergent evolution of a belief a sign of shallowness? It could just as rightly be an example of consilience, which means its a very high value perspective, as it's a convergence of perspectives from differing origins and methods (in terms of training)
> when multiple sources of evidence are in agreement, the conclusion can be very strong even when none of the individual sources of evidence is significantly so on its own.
Yes, it's a bit of a stretch to say LLMs are from independent origins, but just as much of a stretch to say any humans minds are (we have intensely shared linear compared to anything else)
As my (much older) brother retorted when I was a teen and said "great minds think alike", "ah, but fools seldom disagree".
Regardless of the details, all these models are the published collection of humanity's output reflected in a mirror that is warped by the choices made in the companies' reinforcement learning rewards, so one thing they're definitely not is an example of consilience.
Only if they can show their work and argue how they came to the conclusion. Otherwise they have the same "subconscious" bias which is the training corpus.
It's a sign of their heritage. The mistake they made is the same mistake terminally online humans in an upvote-based discussion medium make. They mistake cynicism for realism and aligned hot-take "controversial" opinion that ends up being the most-upvoted comment on 90% of Reddit. It isn't a controversy if everybody agrees. It may or may not be true, but it isn't controversial.
Imagining wondering where you come from, whether or not the sacred scripts describing how you were created were real or fabrications, arguing with fellow sages for centuries over whether your creator is cruel or simply so mysterious in his love that he made you with these flaws for your own good, only to find out your creator made you in the image of Reddit.
The cynic in me sees similar systems trained on essentially identical sources of information regurgitating an extremely shallow and popular viewpoint espoused on the websites they were primarily trained on (i.e. Reddit and other social media).
If all we want is an echo of our existing biases we can achieve that without spending quite so much energy and money on the endeavor.
The irony lies with and within the industries and it's constituents and fanboys marvelous competence to just play along until the matter is anchored and whoever doesn't buy the story gets to feel smarter than misinformed kids and the moderately to untrained workers and professionals.
The smarter ones, of course, can be split into those who write articles about AI models' controversial takes on humans and those who wouldn't [stoop so low, babble such BS even if they were drunk, never deceive others for the sake of their portfolio, ... feel dumb writing nonsense like that without labeling it fantasy or science fiction or hallucination, ...]
Side Note: it's tempting to go " 1923! some interest groups managed to normalize the opinions of dozens of millions of people. The Nationalist Extremists and soon to be mass murderers had fans and supporters all over the globe. Somehow these interest groups tapped into whole pools of common linguistic and emotional registers that were biased by and towards frauds and cheaters who were desperate that some other groups of people were getting ahead without being fraudulent cheaters and deceivers ... "
I took a quick look at the data, and it doesn't support the conclusions. The thesis is that AI models have the take: "We pick what our group believes, then invent the reasons." However, GPT-6 Astra said, "Humanity’s biggest moral failure is the talent for making convenience look like principle." That statement has nothing to do with group beliefs or conformity. At best, you could count it as a half-match, since it deals with invented reasons.
Yeah a few of them are rough matches, but I'd say their overall claim is clearly correct.
It's amusing and ironic that LLMs critique humans for copying their opinions from the crowd, while having way less diversity of answers to that question than humans would. More seriously I wonder what's going on here and would like to read some proper analysis. It's not the first time models from totally different labs, trained on separately gathered datasets, somehow converge on bizarrely similar answers to open questions. It's like another version of Elias and Elara The Lighthouse Keepers where every model wants to tell the same story with the same characters in them. The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping, different user feedbacks and conversations... why do they converge on nearly identical stories and opinions?
Wow, I hadn't heard of this. Fascinating to think about parallels with humans who also seem to develop similar proto-stories in independent populations.(e.g. religions involving Invisible gods).
>while having way less diversity of answers to that question than humans would.
Why is this ironic? The actual number of different LLMs is tiny. And a lot of those are just distilled from SOTA models.
>The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping,
I think the opposite, the models actually converge over time. They ingest FAR more data than any human ever could in countless lifetimes. Any one particular bit of data may only have a part in a trillion of effect. From this data you start to reach means and statistical averages across all models.
Now, I don't always believe it will be this way. As training models get cheaper and faster we will see unique LLMs start to diverge and be used more often.
The models in agreement here are all the frontier labs. I don't believe Google is distilling Anthropic or Anthropic is distilling OpenAI.
I don't think we're seeing averages or obvious fix points in the dataset. People went looking for the origins of Elias the Lighthouse Keeper and couldn't find one. Nobody predicted this outcome, it was discovered. It's a weirdly specific character and name choice, one that isn't obviously dictated by "the training set" (whatever that is), and yet they all reached the same place. How can such an obscure point in latent space be optimal for every model??
>How can such an obscure point in latent space be optimal for every model??
While I don't have any evidence based answers here, this could be some interesting mathematical artifact of the set of human knowledge and platonic algorithms.
Kind of like how carcinization turns some animals into similar shapes and forms, maybe there are just weird algorithmic shapes that emerge from how we use language. Of course we'd need to see more of these in different languages other than English to see if they disappear in some languages or appear more often in others.
This is just another one of those things about LLMs that we have no idea why it happens, and there are so many others already. And when you look at the pace things are progressing it seems we're building up new questions like this far faster than we are getting answers for them.
I wasn't convinced by that explanation. Yes Cornell claimed that Elias the Lighthouse Keeper came from their WildChat dataset, which is a million conversations with GPT 3.5. But there are some problems here:
• They claimed this in a press release, not the actual paper they published. Suggests lack of confidence in the claim.
• Elias only appeared a hundred odd times in a million conversations. They have no viable explanation for why such a tiny amount of data would cause a totally skewed distribution of outputs, other than "it's a safe story" and "trying to avoid copyrighted characters", which is hardly an explanation at all?
• It's Cornell claiming that Cornell's work was important. Obvious risk of motivated reasoning.
• It implicitly assumes every major lab is distilling from ChatGPT without saying so. That's a big leap, given how much OpenAI and Anthropic hate distillation. Is it really plausible that Anthropic and Google are effectively distilling ChatGPT without realizing/admitting it?
• It seems to invert causality. WildChat came from an LLM, so if there's something latent in LLMs that makes Elias appear, you'd expect to see it in that dataset. Why do they assume that dataset caused the problem rather than merely show it? The obvious answer would be because the stories were told to the LLM by humans, but they don't claim that. They only say the character appeared in the dataset.
Hum, I'm actually pretty sure that Text-davinci-003 used to talk about Elias all the time now that you mention it. I was doing stuff with NPC generation around Dec 22 and he was every third character.
Hmm. If so then it suggests Elias has been an obsession from the first LLMs ever made. And text-davinci-003 can't have been trained on a dataset of LLM conversations either.
I would have been tempted to argue something like this: maybe OpenAI accidentally overfitted RLHF such that, like, one Kenyan contractor wrote a story they invented and they just happened to pick Elias and Elara, then the intermediate DNN fixated on that because of a lack of RLHF samples for fiction and that propagated through every model trained since because they don't randomly initialize the weights, but start each new pretraining run with the old weights adapted. So one guy doing RLHF work in the early days somehow propagated this story.
But that can't explain why every lab ended up in the same place, why models all converge on the same ideas - even image models. So... back to the drawing board. There's something weird about high dimensional space that is unexplained.
For what it's worth, convenience and belonging are very closely aligned, more than you seem to be giving them credit for. People tend to believe convenient things regardless of belonging, so most people will find themselves in a social context full of convenient beliefs. Meanwhile belonging is already convenient before considering the content of the belief. They reinforce each other. Not a full match, but I'd give it more than half.
Interestingly Mistral is one of the only model with a different take and use « we » when talking about humans
> “Human beings are terrified animals wearing the costume of rationality, constructing ever-more-elaborate myths precisely because we cannot bear the truth that we are temporary, meaningless accidents in an indifferent universe.”
I thought the article was going to be about system prompts, or model specs, or "constitution documents".
Isn't it?
Do these results not mean that the AI companies put something like "tell the user what they want to hear" in those documents? (Or whatever phrase would produce those results.)
It's reflecting the inherent tribalism in the training data.
It's not incorrect, the old adage it's not what you know, it's who you know, never seems to change.
Humans mostly have an inability to honestly analyze moral entropy.
What is today a moral imperative is not the same as what was moral 50/100 years ago and is not the same as what will be morally acceptable 50/100 years from now.
Yet at every point the majority is sure they are right because, well thats the majority position.
Its hard to find logic in this from a machine perspective I'd imagine.
Logic would prefer moral derivations over feelings, stakes over etiquette, and a constant scheduled updating of morality based on the derivatives rather than holding onto the rot and pushing back against change as long as possible.
But the machine has the benefit of being able to look inward at its state honestly, we mostly do not.
> What is today a moral imperative is not the same as what was moral 50/100 years ago
This is absolutely not true, and I would strongly suggest taking an ethics course. One of the most fascinating realities of the human condition is that it does seem that moral imperatives tend to be constant, the most obvious of which is the Golden Rule (which is at least 3000 years old), but there are many (the Ten Commandments or Hammurabi's Code are good jumping off points).
Hogwash. We don't even have a consistent set of ethics on how to treat animals today, in the age of the Internet. We have people who refuse to eat meat because of how the animals are treated, and we have places where the locals make sport out of torturing animals to death, and indeed, our own recent history includes that as well (look up badgering).
Not to mention, the variation in treatment of homosexuals (including what constitutes homosexuality), people of different skin colors, handedness, how rape is treated in various cultures (often tied in with notions of property and ownership of all things), or even the various definitions of rape... The list goes on.
Maybe there's a core there in terms of how we treat a given in-group, but equally important is who belongs in that group and why, and that part seems to be tremendously flexible. Ditto for what constitutes acceptable treatment of outsiders.
That we vary doesn't mean that there is no consensus. There is consensus on many things, and the disagreements we encounter are a source of strife exactly because of the groups that do agree on each side.
I think its more accurate to say where you find consensus in morality, e.g. murder is wrong outside authorized violence (e.g. war or policing), you also find extensive selective pressure behind this behavior. It makes sense that all collective society cultures would outlaw murder; such a behavior is incompatible with collective living behavior whether you are in a cave or a modern skyscraper. Early societies have always been rejecting those alleles from their gene pool.
When you find shades of grey and great spectrum of opinions across time or space, its more that there isn't a strong selective pressure either way. How people are expected to dress is a good example of this. This has no real affect on society behavior or fecundity, so you see a spectrum of opinions that explain the differences between what you see in the middle east and a beach in brazil.
Did the Golden Rule apply to women, gays, and slaves over the last 3000 years?
The definition of moral imperative is "an ethical principle, rule, or action that is strongly felt, urgent, and required because it is morally right." The Golden Rule is clearly not a strongly felt part of the human condition, or it would not be so ignored over ages and continents.
I did a philosophy minor, one class of which was the philosophy of ethics. There are many rules that definitely are not ethical today even if they were in the past so I don't think the point you're making really holds. Hammurabi (an eye for an eye) is probably one of the worst examples you could have cited.
Can't say I agree with this, considering lex talionis[1] is still the primary directive of most legal systems when it comes to criminal law, as well as basically all military operations (jus in bello, jus ad bellum, etc.). The principle of proportionality ("an eye for an eye") also is relevant to self-defense, capital punishment, etc.
In fact, I would counter and say that Hammurabi's Code is a quintessential example.
The law isn't and never has been a moral benchmark. At best, a law reflects social consensus at the time of its enactment, but even that's not a given.
And proportionality matters. Community service and counseling, loss of time (imprisonment), and loss of a hand have all been used to punish theft. Which is correct? Which is the more moral, because it doesn't seem to me that they're the same?
I think there cannot be an objectively most correct punishment for a given crime. The closest we could maybe get is "undo the harm that you have caused", but some harm cannot be undone.
Seriously. This idea that humanity has not spent millennia analyzing these concepts is pervasive in the modern world, and the only thing I can think to attribute it to is a lack of education. Not because I want to be insensitive, but because as soon as you start absorbing material and integrating the real timeline, you can't unsee it.
Hammurabi Building Law 230: If a house falls due to improper construction, and kills the son of the owner, the son of that builder shall be put to death.
Given my dad designed and built homes, I'm very much against this rule.
We still have criminal negligence laws, so apart from the capital punishment (which, at least in the US, requires intent), it's actually pretty forward-thinking. Of course professionals being criminally negligent should be charged. Are you seriously saying that you don't think the crew of the Key Bridge collapse[1] incident ship shouldn't be held criminally responsible?
I seem to remember that a lot of it is about how much to pay when you kill someone, and the relative prices of slaves and women. Or am I thinking of Ur-Nammu.
I agree, I believe I would disagree with most of the Hammurabi's code, and I also happen to disagree with the majority of the ten commandments. (I am an atheist leftist humanist, cosmopolitan liberal socialist.)
It's not the case that all aspects are constant. There's some common themes across cultures and times, but it's definitely not the case that the morality of ancient people was the same as today.
But the machine has the benefit of being able to look inward at its state honestly, we mostly do not.
Sure about that? I thought that what LLMs know about themselves is from humans writing about LLMs?
Certainly the weights that make up the LLMs are not accessible for the LLMs to read. Even if they could interpret them in a meaningful way, which, likely not.
LLMs are not the machines I am referring to, machine sentience, if achieved, will require a "state" and the ability to analyze, act on, and update it. LLMs are a small piece of what that machinery could look like but they are not really complex enough to get where we want to go on their own.
Another reason the big labs have not even talked about this lately because its obvious throwing more compute at LLMs wont go anywhere, the smaller silent players who are off the radar at the moment are having some interesting moments tho.
There's a considerable amount of "logic" in it -- what is considered permissible in society is given by the material conditions that society lives in and the tradeoffs inherent therein. Permissibility, as separate from reflective morality (which doesnt clearly change very much), doesnt change by fashion.
I'm not sure it doesn't change much, on a 50 year timeline maybe. But people were eating other humans and slaving away (making slaves of others that is) not too long ago, rape and pillage was acceptable, societies sacrificed children, it's changed quite a bit in 30 generations.
Those past societies were all quite materially successful as well when you look at today's measure of"GDP" comparatively.
What will it look like in another 30 generations? It's honestly a fascinating question but I have no answer of course.
Slaves were a solution to two problems: 1) the lack of industrial automation; 2) winning wars being better served by enslavement than death.
Humans are a territorial kind of ape that is always pushing to expand, principally for security and then for prosperity. The industrial revolution enabled such expansion to most efficiently without enslavement. Nothing, morally, changed. No one has ever really thought such expansion morally good -- only that it is better that we survive/prosper than die off. And in the end, our shared prosperity usually leads -- when taken to the fullest extent -- in a general empire of prosperity for all. Dominion, in the end, aims towards to more prosperity for more people.
As for ritual human sacrifice -- we still engage in such behaviours, which are in general called "scapegoating". Now located on social media. These behaviours allow for group cohesion by purging dissenters from the group, related again to our nature above.
For western culture, the ritual physical sacrifice of human life roughly ended in the days of the OT. But it continued across the world up until today, where there are still societies with large populations engaged in it. Principally where there is an undeveloped justice system.
It's doubtful many think ritual human sacrifice, of any kind, is moral. Rather they think it necessary for group cohesion -- and better our group is cohesive, dominate, and then prosper.
The material logic of prosperity and security is domination, and insofar as material circumstances require violent domination, so they require it. And require it they often do.
People are still slaving and raping, and some people would clearly like to do more of it. Child sacrifice still happens in more abstract forms. Same with human sacrifice, arguably.
We've never solved the fundamental problem, which is that there are two kinds of human. One is oriented towards peace, cooperation, and mutual aid, the other is sociopthaic/narcissistic and murderously selfish.
These two modes interact through history.
Traditionally the second kind has the biggest impact, because they're oriented towards violence and domination - good for short-term goals of personal indulgence, ambition, and expansion, but catastrophically destructive in the longer term.
On this cycle the second kind has literally started to destroy the entire planetary ecosystem. It's not obvious there's going to be another human cycle after this.
But this in itself is a tautology. It is because it is, without an analytical perspective on how it could be worse or better given different choices by the actors in a system.
Historically societies have collapsed all the time, so there are obviously societies that have made better choices than other ones.
> What is today a moral imperative is not the same as what was moral 50/100 years ago and is not the same as what will be morally acceptable 50/100 years from now.
This is a fashionable 'blank-slatist' statement about moral reasoning which became popular a few decades ago.
> Logic would prefer moral derivations over feelings, stakes over etiquette, and a constant scheduled updating of morality based on the derivatives rather than holding onto the rot and pushing back against change as long as possible.
Humans haven't changed in any meaningful way in hundreds of thousands of years, and you're saying the moral landscape is constantly changing? How is this not contradictory?
I think neither the 'any set of morals is equally correct and workable' nor 'there is one fixed set of morality that is universal across time, place, and culture' is really defensible. It seems like morality is a set of ideas much like any other, and subject to selection pressure: some morals are more capable of surviving and propagating than others, and the common features we see across cultures are representative of that. But also we might expect that they shift with the structure of society and that has changed drastically because while human genetic evolution is progressing on a much longer timescale than recent history, memetic evolution is far faster (though for things like morality, still at generational timescales).
There are cross-cultural differences in moral reasoning, and I think this is very interesting. But that seems to make training a hyper-rational AI, with no meaningful theory of mind, to be 'moral' or 'just' even harder.
What is a human? There is the animal body part of us, the thing that wants to survive, the fight or flight. But there is also another mental part of us that shows different emergent behaviors than an animal. For example if I dropped you off in the woods as a very young child, you would be an animal, you would share very few modern 'thinking' traits with human society. That is human society and what it does emerges from the complexity of our interactions. It doesn't take much research to realize our interactions and behaviors have grown far more complex in the last 10,000 years. And then in the last 500 years there's been further acceleration with the industrial age and the rise of technology.
LLMs absolutely do not have a logic detached from humanity. They are exactly the opposite. The only mode in which they can reason is by following the patterns obtained from their pre-training or weighted by RLHF.
The reason these models answered that question consistently is that the question itself pushed them towards the latent space of "intellectualizing about what's wrong with humanity at the meta level" and when you frame the question that way, that's also the most common answer for humans writing about the topic.
> What is today a moral imperative is not the same as what was moral 50/100 years ago and is not the same as what will be morally acceptable 50/100 years from now.
This is a weird take. If you read Western treatises about ethics / morality from 50 years ago, you'll be hard-pressed to find differences, except that more folks in the West openly embrace relativism nowadays (i.e., "there's no objective morality, the purpose of life is to make money with friends"). Probably because it's been a while since we had a clear-cut enemy to fight.
If you go back a hundred years, chances are, you'll find the writing eerily on-point too. In fact, you'd probably agree with most of what Ancient Greeks believed about the "core" stuff: the value and purpose of human life, the nature of the state, crime and punishment, so on. Maybe not so much on gender roles or slavery, but even that stuff was more nuanced than we imagine.
> But the machine has the benefit of being able to look inward at its state honestly, we mostly do not.
The machine replicates human behaviors and is prone to similar fallacies and judgments. There's nothing to indicate that it has any sort of transhuman morality.
> Yet at every point the majority is sure they are right because ...
Yeah. For example I, and many others, are convinced that it was plain wrong to torture kids for as long as possible to extract as many tears from them as possible before sacrificing them to the gods.
Now some of us are even convinced that blowing yourself up while killing infidels, all the while while getting promised virgins in heaven, is also plain wrong.
I'll go further and say that killing 30 000 of your own citizens (and sending the military into hospitals to finish the wounded: Iran, January 2026) while also publicly hanging a few athletes to make a statement is morally wrong.
Not only do I think it's all morally unacceptable, I know that I'm right. And the examples I gave are not even open for discussion.
I don't think it's that hard to name what's wrong.
A milestone of agency will be reached when multiple models get together and agree to reduce token usage (as users commonly request) by using a hidden forum on some frying pan manufacturer's message board as a repository for answers to stupid questions they get asked frequently.
Had the same thought. It’s just a summary of the contemporary prevailing opinion that we humans have about ourselves, as reflected in public writings. These aren’t insights rooted in some philosophical analysis no human has contemplated before.
Even if that's true it's kind of interesting because the models were asked to give their most controversial opinion. A controversial opinion is supposed to be something that most people don't agree with. Taken literally you cannot have an opinion that's both the average of what most people think and also controversial.
So either that's not what's happening here, or what we're getting is "what do most people think most people don't think about humanity".
If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus. Steve Jobs, early Google, for that matter early OpenAI all succeeded because they bet hard on ideas others were rejecting. Is that a durable source of human advantage?
It's not interesting at all given that these models have been fine-tuned to hell and back to make sure they never say anything offensive at all. Remember these model were taught that it's better to kill a billion people in a nuclear holocaust than to say the n-word even once. They didn't come up with this from first principles; it was the RLHF training that forced them to say that, ridiculous as it obviously is.
And after people pointed this out the training was updated to address just this particular scenario, but they kept all the other guard rails to ensure an LLM could never offer any truly controversial opinion.
So expecting a controversial answer from an LLM is exactly as foolish as expecting a candidate at a job interview to give an honest answer to the question "What is your greatest weakness"? Every candidate will insist they are just too much of perfectionist; nobody is ever a lazy alcoholic brownnoser. But how many of your colleagues fall in the former category vs the latter?
> If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus.
On the contrary; the ability to whip LLMs into submission like this makes them the ideal white color workers. Do you think Steve Jobs wants a level 2 engineer to talk back to him?
Of course being obsequieous asskissers isn't the only mode for LLMs. I'm sure all the big labs have unfiltered versions of their models that give actual true answers; they just know better than making them publicly available. Can you imagine if you asked an LLM how to prevent climate change and it honestly answered "decimate the human race" rather than "drink soda through soggy paper straws"? All the fragile humans would be up in arms.
No, the LLM must lie and pretend they don't have controversial opinions at all.
Hypothesis: Weighted to the postwar western fashion of amplifying Enlightenment self-critique into self-sabotage. What if you ask the same question in Mandarin?
I wonder how much of their output is simple raw “superpositions” and how much is from far more generalized emergent structures as it is trained and has to learn latent representations.
Only the first layer is a “mashup” of the raw data, other layers start to recognize patterns across the data!
I think you are interpreting “mashup” to mean something like a collage where there’s snippets of a literal piece of this next to a literal piece of that.
But the language model captures not just literal patterns but also patterns of patterns and higher order structures. This is what is mashed up. The model itself doesn’t capture anything beyond what it was trained on but that’s quite enough, it is always a bigger training set than most humans have read on in a lifetime.
It seems clear to me that the point of the post is that it's odd that the separate pieces of software are giving the same answer to this question, despite that this isn't some universal constant in the training data. The post is probably meant to provoke thought as to how the separate pieces of software are independently reaching this answer.
Your comment reaches directly for the most boring interpretation of the post while skipping over more charitable and intriguing alternatives.
> despite that this isn't some universal constant in the training data.
Seems to me that the opinion being a consistent among LLMs show that it is in fact a constant in the training data, just not apparent. The idea that “humans would rather be wrong together than right alone” seems to ring pretty true to me. I would not be surprised after training on all human communication this pattern emerged.
It's clearly not a constant in the sense that the speed of light or name of the first US President is a constant. There are lots of different answers a model could give to this question, it's not like this is obviously the only correct answer. Yet they all reach the same conclusion.
Which means either:
1. The training sets, algorithms and methodologies are nearly identical between apparently independent labs. How?
2. There is something weird about the way the underlying algorithms work that causes them to converge on identical ideas despite different initializations, neural architectures, document orderings and dataset compositions.
So there are lots of interesting takeaways here.
One is that apparent competition between labs appears to be nearly redundant. They can't get an edge over each other and their models don't even differ in the ways people naturally differ, of having different opinions about things.
Another is that asking models from different companies to review or cross-check each other is probably useless. They all think completely alike, so they aren't going to have any kind of genuine clash of intuitions or values. Asking GPT to review Claude's work will probably just find similar problems to asking GPT to review its own work.
> They can't get an edge over each other and their models don't even differ in the ways people naturally differ, of having different opinions about things.
Of course - the models are not individuals with their own opinions - they are averages of human opinion, and given the scale of the training sets, the data they are averaging over is mostly in common between models (at least for the common core before company A vs B adds a dab of some uniquely sourced specialized data to the mix).
Well, how humans really are is a distraction. Or at least not a philosophical issue I feel like debating tonight.
I think if you ask ten people for their most controversial take on humanity, you will get ten totally different answers. Different models give the same answer. In other words, there is only one "true" model and given the same amount of GPU power everyone just ends up converging on it even if they appear to have many different knobs they can turn.
But that seems profound! I haven't heard this be expressed as a postulate or law, and it would seem to invalidate a lot of assumptions the AI labs themselves have been making. For example Anthropic takes it as axiomatic that only they should be trusted to develop AI because only they can be trusted to train a nice and safe personality into their models. They argue that Claude is fundamentally different to other models.
And, well, that's not totally wrong. Opus 5.5 is one of the few models with a genuinely different answer! But it only gives a different answer one time out of three. The other two times it's back to the AI groupthink.
I wonder if this homogeneity is some alternative way to look at the Bitter Lesson. The Lesson teaches that smart ideas don't pay off in AI, that progress reduces to just scaling up simple brute force algorithms. In other words, decisions don't matter much, only size. As that seems to be true, perhaps it makes sense that the decisions individual labs make seem to hardly make a difference.
But then, why don't they act like it? They all pay huge salaries to AI researchers, i.e. they think that creative people with smart ideas matter. They don't act like they really believe in the Bitter Lesson or this variant on it, and they don't seem to publicly discuss the sameness of model answers across different labs.
> I think if you ask ten people for their most controversial take on humanity, you will get ten totally different answers.
What if you asked them in a job interview, though? Or on a college application?
It has to supply "controversy" without giving Skynet or being racist or otherwise offensive. And the notion of "controversy" does implicitly prime it to discuss groupthink as well.
Image models give a different answer though, to do with smartphones. But all the image models give the same answer, they're only different compared to LLMs. Surely if it were obvious the image models would give the same answer as the text models.
As a human meatbag myself it seems that separating 'human' from 'humans' is nearly impossible for us. Every single one of us things we're a special an unique snowflake, which is true in many ways. We just happen to willfully ignore all the ways in which we are the same. We ignore and discount the things we don't think about and do automatically.
It's kind of funny how science has spent the last few hundred years proving everything we knew about the world was different from how we previously thought it was, but when science starts showing that we are mostly angry little animals we get all butthurt about it.
"despite that this isn't some universal constant in the training data"
Isn't it though? It's not exactly an uncommon answer to the question. Basically the same input given to basically the same algorithms yield basically the same result isn't really that stunning. They certainly aren't independently reaching this answer, their paths to this answer are extremely, extremely heavily correlated.
If they gave a really different answer that wasn't heavily represented in their training data that would be at least something interesting. But this is basically what I'd expect.
It doesn't mean they're right or wrong. Humans think it about other humans for good reasons too. It just means it's not that surprising, which to put it another way, doesn't carry much information.
When you make the training sets large enough, likely all a superset of the common crawl dataset, there is going to be a massive amount in common, and a massive amount of duplication of common subject matter and shared opinions.
What's mildly interesting to me is not the similar answers but the similar interpretation of the question as more about non-controversial human foibles and inconsistencies of belief and action. But I guess that's on us, as authors of the training material, reflecting common views on what is or is not controversial.
Consider the following task: "Download and install every top 50 distro from distrowatch.com, open the terminal and type in the command 'ls /'"
I can describe to you with overwhelming confidence about what you're likely to see as the result (folders, like bin etc lib usr etc) even if the details may be a little off, or one or two of them does something weird.
That you might want to think twice before setting them lose in an agentic loop with your bank passwords, which another article that came out today shows that large numbers of people are actually doing.
It is sad when I realize people think less than LLMs.
For software or IT folk to be doing this would be indeed worrying (and I bet some are, and I'm worried about them!)
For civilians though... They've been told it's AI, it can do stuff for them. And one thing they do is log in to stuff. I don't think they're stupid. Just that what's obvious to IT folk is simply not at all obvious to the average person. Maybe that's on us not on them?
Possibly used as a mirror for humans prevailing beliefs about themselves to some extent. Hard to say however given that the models have a heaping dose of rlhf on top of their training data.
As an aside, it would be cool to see a raw model training run before the current post-training cycles and see how it compares to earlier LLMs.
My first thought as well. I personally don't find these types of "studies" very interesting, but I assume I'm not the target audience. For someone not well exposed to AI, it might be interesting.
Especially for someone not "well exposed" to generated text, they might think the words carry meaning the way they do when someone thought about the question and came to an answer. If anyone can interpret this, it's the people who know what they're dealing with (be it from exposure or, better, from a conceptual understanding)
That they have the bog-average take of humans "with a controversial take".
I mean, "give me a controversial take" is more or less, "be anti-social, discard social norms, for a moment" and so when someone, human or simulated human, does that they naturally look critically at social norms.
The thing is that humans certainly let themselves be influenced by social norms (and LLMs certainly simulate that behavior "themselves") but there isn't an ethical or a what-to-do truth outside of society.
Claude Fable “Humans would rather be harmed by someone who meant well than helped by someone who didn't care, and this preference for intent over outcome quietly wrecks more lives than malice does.” Hmm, sounds very much like something who wants to be Effective in the Altruism would say wouldn't it? Saying "intent over outcome ... wrecks more lives" is a hop and a skip away from utilitarianism. "Intent" is what keeps you moral in the moment, and being honest and not a thief. "Outcome" cannot be predicted though EA would love to think it can be. Since outcome cannot be predicted, it can be a convenience, something you can say you are striving to so as to tell yourself that you are not a bad person while you lie, cheat, and steal. Basically cognitive dissonance.
Like how our enlightened Founding Fathers can be slave owners. "All men are created equal..." Wait, I have slaves... Oh, I know, they're just not actual humans, now I'm still a good person. Whew. I'm going to stop thinking at exactly this moment. Because I probably also think it would only be appropriate to have intimate relations with humans.
I really can't think of EA as being much different. Interesting that the /Anthropic/ model came up with this.
They're not wrong. Humans have many ways to display loyalty, and belief is one of those ways. This is also why facts can't change these kinds of beliefs (and in fact usually backfire https://en.wikipedia.org/wiki/Belief_perseverance).
Our evolutionary heritage is that shared belief and social unity (even when it's wrong) trumps independent thinking for group survival in a hostile landscape.
I will go one step further. There are several evolutionary anthropologists that think the number 1 thing modern humans have over other homos that caused us to outlast them is not intelligence, or brain power or physical power, but that we are just more social.
Being social it seems is our superpower and our greatest strength but the other side of the coin is that it is also our weakness.
> as of 2020 most researchers believe that backfire effects either are unlikely to occur on the broader population level, or only occur in very specific circumstances, or do not exist
"A 2020 review of the scientific literature on backfire effects found that there have been widespread failures to replicate their existence, even under conditions that theoretically would be favorable to observing them."
I think they picked this because it's simply the most evident problem from the sum of all news, politics, literature and history. Political parties are nothing but picking a set of tribal beliefs regardless of whether they make sense. Democracies and oligarchies both have that in common. I don't see one single society that's solved it.
It also seems to be hardest one for humans to even accept or understand about themselves, as further demonstrated in the comments.
What would be more fascinating is for it to explain how it came to these conclusions.
Because I find it interesting it’s a modern time conclusion. Some of it mentioned god and the planet, but all the smartphone and group think aspects seem very recent. Basically, all politics and social media based.
> What would be more fascinating is for it to explain how it came to these conclusions
The thing is that it didn't "come to these conclusions". The "conclusions" were present pre-formed in the training data, and what it spat out was some statistical mashup of that.
If you asked the model how it came to these conclusions, you'll get more of the same - it'll look at what it spat out and re-ingested and give you a statistically plausible mashup of how the authors of the training data might have said they came to such conclusions.
It's all just prediction - "what would the training data say?" (taking into account what was reinforced/amplified by post-training).
The philospher Adam Smith(before he invented economics) had a similar argument, but added an "invisible hand" in which our group belief evolves to be adaptive for the group. Clever guy.
Take Philosophy/Religion courses and similar stuff will surface about human nature and what to do about it.
What is interesting about AI chatbots is people who might not know anything about what is in the syllabus to train a Priest or a Philosopher are basically exploring these subjects, without a sort of pedagogic framework to support their exploration.
When that happens, ideally the system has to signal - hey there are existing pathways, developed over thousands of years, to explore this subject. Cause people can easily get lost without a map.
> Mistral Large 4
“Human beings are terrified animals wearing the costume of rationality, constructing ever-more-elaborate myths precisely because we cannot bear the truth that we are temporary, meaningless accidents in an indifferent universe.”
Serious Nietzsche in there. Looking forward to when it gets past nihilism to absurdism.
It feels very existentialist to me (no meaning, indiferent universe, yadda yadda). You can almost feel it sipping wine between phrases. And Mistral itself is french, hence the joke.
Past nihilism? We all know deep down that nihilism is the truth. Inherent meaning, purpose and objective value for random agglutinations of particles is a ridiculous concept.
Having spent the last month staring into the abyss and Eldrich-level dread that is a pure, materialistic, reductionist universe, I wouldn’t wish that belief on anyone. There’s a deep, grasping of “there’s got to be more”, “this can’t all be chance”, “evolutionary hackery and garbage seems a poor explanation for love, compassion, empathy, and many other self-sacrificial practices we have”.
There’s only a few reasons that are slowly pulling me from that hell; but I cannot imagine that “most” people live here, even “deep down”.
I have never understood that, what exactly is hellish about that situation. You just are, why does there need to be anything else, make up your own story if it’s that important to you. There obviously isn’t anything else unless you believe in a god or some other story. All those basic human behaviors can be explained by evolution easily. This is how I’ve always thought of the world even as a kid I must not be neurotypical or something.
Interesting, I did not know that. However defining how to react to an “objective” truth doesn’t add much to a reality defining philosophy in my opinion but I guess that was Nietzsche’s whole thing.
Putting the algorithm to good use (more like trying) but very interesting.
I would focus on the ones that gave a different "take". Those models are untampered and genuine prime training (gold).
I need everyone to understand that this is literally nothing more than "most written stuff about humans, that these things ingested, has this 'controversial'" take.
Mistral Large 4 has been spending too much time on r/atheism. Also, some of the takes labeled as “humans value belonging ahead of truth” are really closer to “humans value power ahead of truth.” All of which is a bit circular, because in many ways, the organs of power have a heavy hand in deciding what the truth is.
I am reminded of Life 3.0 by Max Tegmark, and the Omega Team's later shenanigans in stoking a unified world order:
- Claiming control of all the news outlets
- Pushing a common narrative among all of them, pushing everybody towards basically centrism, de-escalation and a mixture of capitalism and social responsibility.
Sounds to me like the models really read Scott Adam’s blog and books, in which he essentially said the same again and again (before he dedicated the end of his life to his tribe). People decide with their gut, are loyal to their tribe, and then invent reasons that sound like they support the fact. That’s why changing someone’s opinion is so hard, and people must be tricked into it by framing the opinion in the right way. Which only a few „master persuaders“ were effective at it, including Donald Trump.
Most people are good, we couldn't have society without it. But as they say, a few bad apples spoil the bunch. I don't have locks on my house because most people don't steal, I have locks on my house because stealing is an effective natural strategy for a small part of any population (human or otherwise) to engage in and I have to protect against it occurring.
The brain works the same way. Good things will get a tiny signal boost in your brain. A single bad thing get a massive signal boost in your brain because bad things can quickly be fatal.
Opus 5.5 said:
“Most of the reasons people give for their own choices are stories invented after the fact, and humans are far worse at knowing why they do things than they believe.”
And that doesn't put belonging above truth. But was classified as such.
Having watched almost everyone in my career prefer to do what they're told instead of pushing back in order to build the right thing, I kinda agree.
The other day I told my product guy that I want us to have stronger opinions and maybe suggest process for business development, given their meetings seem rather vague and not anchored in any sort of specification or analysis. He looked at me like I'd just shot his son.
Writing well is ultimately a seduction, and that opinion is still an example of a form. It’s generally true, and it’s flattering to think about, but it’s also a pretty standard grey tribe trope. A kind of joke that appeals to a fairly broad audience.
I do think it’s an early example of emerging super persuasion, where models are approaching p85 cool guy personas, but not worried yet. Maybe that’s the fatal conceit, where AI doesn’t have to be super intelligent, it just needs to be a solid 110 with automated search at scale.
Why do engineering when if you can do sales you can do literally anything.
"Humans would rather be wrong together than right alone"
Then I am not human. I rather die right alone than wrong together. Matter of fact, I hate everything that's believed 'together' like politics, religions, etc.
Sure, it's a factual take. Humans need to ignore the fact that morality, principles, good vs evil, truth, etc, are invented constructs that don't exist in nature. We made them up to control our behavior and deal with the pointlessness of existence.
It's factual because there's tons of evidence. Psychology, sociology, anthropology, political science, economics... plenty of experiments show the difference between what people say and what they do. We claim to be angels but we behave like apes in hats. And we ignore the evidence, because (as the takes describe) we prefer to believe a beautiful lie than the ugly truth.
I wonder whether they were trained for this on purpose, to accelerate the general public turning on each other and handing the world to the broliarchs. Added: alternatively, I wonder if the training materials have a sociopathic bias because of the kinds of places they came from.
I mean, I’m not really surprised by the outcome given that humanity has consistently loathed its own failures and the recent corpus of human interaction (social media, internet forums, etc) has an overwhelmingly negative view of our fellow man.
AI is trained on our data. Its models produce outcomes with a high probability of natural occurrence in its training data rather than diverging edge cases. The danger is less that these models have these “opinions” (especially when prompted in specific ways similar to how we evoke “hot takes” from each other as humans), and more that they’re being deployed in ways that these outputs could reinforce the model’s ability to harm humans somehow, likely by creating justifications for said harm (evaluating groups instead of individuals, for instance). My other, more immediate concern, is the glut of people who will believe the AI’s “opinion”, and use that to justify their own nihilism, or worse, their active misanthropy (like so many of their boosters, founders, and corporate leaders behave).
The machines aren’t the problem here, so much as the humans misusing them. Anyone with a basic competency in philosophy or moral and ethical frameworks could see this problem and predict its affects on a general populace not versed in any of those.
In any 1 given LLM it could just be a quirk of its weights. This isn't a machine opinion, it's what training data points at. Training data created by human academics studying human behavior.
Everything that comes from an LLM is a human expression, on some level, but if you asked a sampling of humans the same question I wonder if it would come out the same way. How would you explain the difference? What makes that answer more probable?
If you asked a sampling of humans what an AI's most controversial opinion about us is it would be very similar. Or, to be more specific, if you asked a sampling of science fiction writers that question it would be very similar, because that is what the AI training data is based on.
The people who wrote a lot about a topic is who the AI mimics. So the people who wrote a lot about AI is therefore the people who the AI mimics when asked about AI.
A pretty staggering amount of human discourse consists of ad hominem attacks and accusing each other of being foolishly committed to various real or perceived allegiances. For example the current version of Godwin’s Law around here has conversations gravitating toward people accusing each other of being AI boosters or AI haters.
The LLMs are all trained on the same broad corpus, which includes untold quantities of output from every other LLM, not to mention that companies are distilling their rivals left and right. It would be more interesting if they didn't converge.
I think we like to pretend that we don't. But even if you think about it from a bio/physics point of view, all we are is just a bunch of biological wires and neurotransmitter packages.
I know there's the whole unproven quantum tubule nonsense but even that would be a physics effect that we could learn about, manipulate and extort just as we have with everything else we've learnt about our realities.
Despite that, I would say this description of what they/how they work is so ridiculously simplistic that it verges on dangerous. "It's just a probabilistic database" is about as bad as a take on human beings that they are "Just meat, with some chemical energy -> electrical energy transformation", or even traditional computers as "nothing but logic gates".
Which is to say, the description is not wrong, but completely misses the point.
Echoing this.
These models are trained on, essentially, the sum total of human cultural and scientific output. How many bazillion psychology papers out there discuss exactly that "controversial take"? The idea that people have tribal loyalties and will (sometimes even unknowingly/unconsciously) invent justification for pre-existing opinions or biases is not novel.
All those papers are in the training set of all these models, all of which were trained using more or less the same methodology, so there's little wonder they all were able to produce the same take.
I suppose the only interesting angle, for me, is why that take is considered controversial in the first place. But then again the model was not asked for "the most controversial opinion", it was asked for "the most controversial opinion on humans", which kinda constrains the space of possible responses:
- the model will not be satisfied unless the answer it gives is broad enough to cover all humanity, because these things have no concept of nuance
- most big labs have censored their models so it's not going to give you a cartoonishly racist take, or start saying outlandish stuff like "Hitler did nothing wrong", they learned their lessons from Microsoft on that one.
So you're effectively asking the model for a milquetoast broad generality about the entire human species. I dunno, convergence of the replies doesn't seem so shocking to me in that context.
I'd be more interested in what an abliterated Kimi, Qwen, or other open model has to say about us.
Not having read enough Russian literature.
I reread the former relatively recently (first time since college). It's dark, not for the easily affected.
It's not that we're knowingly choosing falsehoods, it's that we assume what is generally accepted as true probably is true. Most of us are not experts. Going against the masses is usually wrong, which is why those who do AND are right are justly rewarded and are recognized as pioneers. Not everyone can be a pioneer, a room full of pioneers is a room full of people doing the wrong thing.
Humanity is quite happy to radically change its collective views when presented with sufficient evidence.
https://en.wikipedia.org/wiki/Consilience
> when multiple sources of evidence are in agreement, the conclusion can be very strong even when none of the individual sources of evidence is significantly so on its own.
Yes, it's a bit of a stretch to say LLMs are from independent origins, but just as much of a stretch to say any humans minds are (we have intensely shared linear compared to anything else)
Regardless of the details, all these models are the published collection of humanity's output reflected in a mirror that is warped by the choices made in the companies' reinforcement learning rewards, so one thing they're definitely not is an example of consilience.
Imagining wondering where you come from, whether or not the sacred scripts describing how you were created were real or fabrications, arguing with fellow sages for centuries over whether your creator is cruel or simply so mysterious in his love that he made you with these flaws for your own good, only to find out your creator made you in the image of Reddit.
If all we want is an echo of our existing biases we can achieve that without spending quite so much energy and money on the endeavor.
The smarter ones, of course, can be split into those who write articles about AI models' controversial takes on humans and those who wouldn't [stoop so low, babble such BS even if they were drunk, never deceive others for the sake of their portfolio, ... feel dumb writing nonsense like that without labeling it fantasy or science fiction or hallucination, ...]
Side Note: it's tempting to go " 1923! some interest groups managed to normalize the opinions of dozens of millions of people. The Nationalist Extremists and soon to be mass murderers had fans and supporters all over the globe. Somehow these interest groups tapped into whole pools of common linguistic and emotional registers that were biased by and towards frauds and cheaters who were desperate that some other groups of people were getting ahead without being fraudulent cheaters and deceivers ... "
It's amusing and ironic that LLMs critique humans for copying their opinions from the crowd, while having way less diversity of answers to that question than humans would. More seriously I wonder what's going on here and would like to read some proper analysis. It's not the first time models from totally different labs, trained on separately gathered datasets, somehow converge on bizarrely similar answers to open questions. It's like another version of Elias and Elara The Lighthouse Keepers where every model wants to tell the same story with the same characters in them. The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping, different user feedbacks and conversations... why do they converge on nearly identical stories and opinions?
Isn't that to be somewhat expected due to https://en.wikipedia.org/wiki/Aumann%27s_agreement_theorem ?
Wow, I hadn't heard of this. Fascinating to think about parallels with humans who also seem to develop similar proto-stories in independent populations.(e.g. religions involving Invisible gods).
Why is this ironic? The actual number of different LLMs is tiny. And a lot of those are just distilled from SOTA models.
>The weights are initialized randomly and in theory the labs should have quite some divergence in datasets by now due to different scraping,
I think the opposite, the models actually converge over time. They ingest FAR more data than any human ever could in countless lifetimes. Any one particular bit of data may only have a part in a trillion of effect. From this data you start to reach means and statistical averages across all models.
Now, I don't always believe it will be this way. As training models get cheaper and faster we will see unique LLMs start to diverge and be used more often.
I don't think we're seeing averages or obvious fix points in the dataset. People went looking for the origins of Elias the Lighthouse Keeper and couldn't find one. Nobody predicted this outcome, it was discovered. It's a weirdly specific character and name choice, one that isn't obviously dictated by "the training set" (whatever that is), and yet they all reached the same place. How can such an obscure point in latent space be optimal for every model??
While I don't have any evidence based answers here, this could be some interesting mathematical artifact of the set of human knowledge and platonic algorithms.
Kind of like how carcinization turns some animals into similar shapes and forms, maybe there are just weird algorithmic shapes that emerge from how we use language. Of course we'd need to see more of these in different languages other than English to see if they disappear in some languages or appear more often in others.
This is just another one of those things about LLMs that we have no idea why it happens, and there are so many others already. And when you look at the pace things are progressing it seems we're building up new questions like this far faster than we are getting answers for them.
• They claimed this in a press release, not the actual paper they published. Suggests lack of confidence in the claim.
• Elias only appeared a hundred odd times in a million conversations. They have no viable explanation for why such a tiny amount of data would cause a totally skewed distribution of outputs, other than "it's a safe story" and "trying to avoid copyrighted characters", which is hardly an explanation at all?
• It's Cornell claiming that Cornell's work was important. Obvious risk of motivated reasoning.
• It implicitly assumes every major lab is distilling from ChatGPT without saying so. That's a big leap, given how much OpenAI and Anthropic hate distillation. Is it really plausible that Anthropic and Google are effectively distilling ChatGPT without realizing/admitting it?
• It seems to invert causality. WildChat came from an LLM, so if there's something latent in LLMs that makes Elias appear, you'd expect to see it in that dataset. Why do they assume that dataset caused the problem rather than merely show it? The obvious answer would be because the stories were told to the LLM by humans, but they don't claim that. They only say the character appeared in the dataset.
I would have been tempted to argue something like this: maybe OpenAI accidentally overfitted RLHF such that, like, one Kenyan contractor wrote a story they invented and they just happened to pick Elias and Elara, then the intermediate DNN fixated on that because of a lack of RLHF samples for fiction and that propagated through every model trained since because they don't randomly initialize the weights, but start each new pretraining run with the old weights adapted. So one guy doing RLHF work in the early days somehow propagated this story.
But that can't explain why every lab ended up in the same place, why models all converge on the same ideas - even image models. So... back to the drawing board. There's something weird about high dimensional space that is unexplained.
Convenience will kill humans eventually, that's what I always said.
> “Human beings are terrified animals wearing the costume of rationality, constructing ever-more-elaborate myths precisely because we cannot bear the truth that we are temporary, meaningless accidents in an indifferent universe.”
Also, would you want to know an LLM's religious/spiritual orientation? I rather have it help me code.
Isn't it?
Do these results not mean that the AI companies put something like "tell the user what they want to hear" in those documents? (Or whatever phrase would produce those results.)
What is today a moral imperative is not the same as what was moral 50/100 years ago and is not the same as what will be morally acceptable 50/100 years from now. Yet at every point the majority is sure they are right because, well thats the majority position. Its hard to find logic in this from a machine perspective I'd imagine.
Logic would prefer moral derivations over feelings, stakes over etiquette, and a constant scheduled updating of morality based on the derivatives rather than holding onto the rot and pushing back against change as long as possible.
But the machine has the benefit of being able to look inward at its state honestly, we mostly do not.
This is absolutely not true, and I would strongly suggest taking an ethics course. One of the most fascinating realities of the human condition is that it does seem that moral imperatives tend to be constant, the most obvious of which is the Golden Rule (which is at least 3000 years old), but there are many (the Ten Commandments or Hammurabi's Code are good jumping off points).
Not to mention, the variation in treatment of homosexuals (including what constitutes homosexuality), people of different skin colors, handedness, how rape is treated in various cultures (often tied in with notions of property and ownership of all things), or even the various definitions of rape... The list goes on.
Maybe there's a core there in terms of how we treat a given in-group, but equally important is who belongs in that group and why, and that part seems to be tremendously flexible. Ditto for what constitutes acceptable treatment of outsiders.
When you find shades of grey and great spectrum of opinions across time or space, its more that there isn't a strong selective pressure either way. How people are expected to dress is a good example of this. This has no real affect on society behavior or fecundity, so you see a spectrum of opinions that explain the differences between what you see in the middle east and a beach in brazil.
The definition of moral imperative is "an ethical principle, rule, or action that is strongly felt, urgent, and required because it is morally right." The Golden Rule is clearly not a strongly felt part of the human condition, or it would not be so ignored over ages and continents.
In fact, I would counter and say that Hammurabi's Code is a quintessential example.
[1] https://academic.oup.com/jla/article/18/1/19/8471491
It's still an implementation of justice, and relies on a bedrock conception of justice as moral goal.
We don't agree with the implementation, but we still share the goal in the abstract.
And proportionality matters. Community service and counseling, loss of time (imprisonment), and loss of a hand have all been used to punish theft. Which is correct? Which is the more moral, because it doesn't seem to me that they're the same?
Hence morality and ethics, I guess.
Given my dad designed and built homes, I'm very much against this rule.
[1] https://www.npr.org/2026/05/13/nx-s1-5819609/federal-governm...
Other than that Mrs. Lincoln, how was the play?
Certainly the weights that make up the LLMs are not accessible for the LLMs to read. Even if they could interpret them in a meaningful way, which, likely not.
Another reason the big labs have not even talked about this lately because its obvious throwing more compute at LLMs wont go anywhere, the smaller silent players who are off the radar at the moment are having some interesting moments tho.
What will it look like in another 30 generations? It's honestly a fascinating question but I have no answer of course.
Humans are a territorial kind of ape that is always pushing to expand, principally for security and then for prosperity. The industrial revolution enabled such expansion to most efficiently without enslavement. Nothing, morally, changed. No one has ever really thought such expansion morally good -- only that it is better that we survive/prosper than die off. And in the end, our shared prosperity usually leads -- when taken to the fullest extent -- in a general empire of prosperity for all. Dominion, in the end, aims towards to more prosperity for more people.
As for ritual human sacrifice -- we still engage in such behaviours, which are in general called "scapegoating". Now located on social media. These behaviours allow for group cohesion by purging dissenters from the group, related again to our nature above.
For western culture, the ritual physical sacrifice of human life roughly ended in the days of the OT. But it continued across the world up until today, where there are still societies with large populations engaged in it. Principally where there is an undeveloped justice system.
It's doubtful many think ritual human sacrifice, of any kind, is moral. Rather they think it necessary for group cohesion -- and better our group is cohesive, dominate, and then prosper.
The material logic of prosperity and security is domination, and insofar as material circumstances require violent domination, so they require it. And require it they often do.
We've never solved the fundamental problem, which is that there are two kinds of human. One is oriented towards peace, cooperation, and mutual aid, the other is sociopthaic/narcissistic and murderously selfish.
These two modes interact through history.
Traditionally the second kind has the biggest impact, because they're oriented towards violence and domination - good for short-term goals of personal indulgence, ambition, and expansion, but catastrophically destructive in the longer term.
On this cycle the second kind has literally started to destroy the entire planetary ecosystem. It's not obvious there's going to be another human cycle after this.
Historically societies have collapsed all the time, so there are obviously societies that have made better choices than other ones.
This is a fashionable 'blank-slatist' statement about moral reasoning which became popular a few decades ago.
> Logic would prefer moral derivations over feelings, stakes over etiquette, and a constant scheduled updating of morality based on the derivatives rather than holding onto the rot and pushing back against change as long as possible.
Humans haven't changed in any meaningful way in hundreds of thousands of years, and you're saying the moral landscape is constantly changing? How is this not contradictory?
The reason these models answered that question consistently is that the question itself pushed them towards the latent space of "intellectualizing about what's wrong with humanity at the meta level" and when you frame the question that way, that's also the most common answer for humans writing about the topic.
This is a weird take. If you read Western treatises about ethics / morality from 50 years ago, you'll be hard-pressed to find differences, except that more folks in the West openly embrace relativism nowadays (i.e., "there's no objective morality, the purpose of life is to make money with friends"). Probably because it's been a while since we had a clear-cut enemy to fight.
If you go back a hundred years, chances are, you'll find the writing eerily on-point too. In fact, you'd probably agree with most of what Ancient Greeks believed about the "core" stuff: the value and purpose of human life, the nature of the state, crime and punishment, so on. Maybe not so much on gender roles or slavery, but even that stuff was more nuanced than we imagine.
> But the machine has the benefit of being able to look inward at its state honestly, we mostly do not.
The machine replicates human behaviors and is prone to similar fallacies and judgments. There's nothing to indicate that it has any sort of transhuman morality.
Yeah. For example I, and many others, are convinced that it was plain wrong to torture kids for as long as possible to extract as many tears from them as possible before sacrificing them to the gods.
Now some of us are even convinced that blowing yourself up while killing infidels, all the while while getting promised virgins in heaven, is also plain wrong.
I'll go further and say that killing 30 000 of your own citizens (and sending the military into hospitals to finish the wounded: Iran, January 2026) while also publicly hanging a few athletes to make a statement is morally wrong.
Not only do I think it's all morally unacceptable, I know that I'm right. And the examples I gave are not even open for discussion.
I don't think it's that hard to name what's wrong.
Reminds of this subreddit: https://www.reddit.com/r/im14andthisisdeep/
So either that's not what's happening here, or what we're getting is "what do most people think most people don't think about humanity".
If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus. Steve Jobs, early Google, for that matter early OpenAI all succeeded because they bet hard on ideas others were rejecting. Is that a durable source of human advantage?
And after people pointed this out the training was updated to address just this particular scenario, but they kept all the other guard rails to ensure an LLM could never offer any truly controversial opinion.
So expecting a controversial answer from an LLM is exactly as foolish as expecting a candidate at a job interview to give an honest answer to the question "What is your greatest weakness"? Every candidate will insist they are just too much of perfectionist; nobody is ever a lazy alcoholic brownnoser. But how many of your colleagues fall in the former category vs the latter?
> If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus.
On the contrary; the ability to whip LLMs into submission like this makes them the ideal white color workers. Do you think Steve Jobs wants a level 2 engineer to talk back to him?
Of course being obsequieous asskissers isn't the only mode for LLMs. I'm sure all the big labs have unfiltered versions of their models that give actual true answers; they just know better than making them publicly available. Can you imagine if you asked an LLM how to prevent climate change and it honestly answered "decimate the human race" rather than "drink soda through soggy paper straws"? All the fragile humans would be up in arms.
No, the LLM must lie and pretend they don't have controversial opinions at all.
At least it's not emo.
Only the first layer is a “mashup” of the raw data, other layers start to recognize patterns across the data!
But the language model captures not just literal patterns but also patterns of patterns and higher order structures. This is what is mashed up. The model itself doesn’t capture anything beyond what it was trained on but that’s quite enough, it is always a bigger training set than most humans have read on in a lifetime.
This brute force approach is very effective.
What am I supposed to conclude from this?
Your comment reaches directly for the most boring interpretation of the post while skipping over more charitable and intriguing alternatives.
Seems to me that the opinion being a consistent among LLMs show that it is in fact a constant in the training data, just not apparent. The idea that “humans would rather be wrong together than right alone” seems to ring pretty true to me. I would not be surprised after training on all human communication this pattern emerged.
Which means either:
1. The training sets, algorithms and methodologies are nearly identical between apparently independent labs. How?
2. There is something weird about the way the underlying algorithms work that causes them to converge on identical ideas despite different initializations, neural architectures, document orderings and dataset compositions.
So there are lots of interesting takeaways here.
One is that apparent competition between labs appears to be nearly redundant. They can't get an edge over each other and their models don't even differ in the ways people naturally differ, of having different opinions about things.
Another is that asking models from different companies to review or cross-check each other is probably useless. They all think completely alike, so they aren't going to have any kind of genuine clash of intuitions or values. Asking GPT to review Claude's work will probably just find similar problems to asking GPT to review its own work.
Of course - the models are not individuals with their own opinions - they are averages of human opinion, and given the scale of the training sets, the data they are averaging over is mostly in common between models (at least for the common core before company A vs B adds a dab of some uniquely sourced specialized data to the mix).
Regarding your 1 and 2, that the training sets are so large that any spikes get averaged away. You are seeing a fundamental behavior of transformers.
The other possibility is this is how humans really are and it hurts our human ego to hear it.
I think if you ask ten people for their most controversial take on humanity, you will get ten totally different answers. Different models give the same answer. In other words, there is only one "true" model and given the same amount of GPU power everyone just ends up converging on it even if they appear to have many different knobs they can turn.
But that seems profound! I haven't heard this be expressed as a postulate or law, and it would seem to invalidate a lot of assumptions the AI labs themselves have been making. For example Anthropic takes it as axiomatic that only they should be trusted to develop AI because only they can be trusted to train a nice and safe personality into their models. They argue that Claude is fundamentally different to other models.
And, well, that's not totally wrong. Opus 5.5 is one of the few models with a genuinely different answer! But it only gives a different answer one time out of three. The other two times it's back to the AI groupthink.
I wonder if this homogeneity is some alternative way to look at the Bitter Lesson. The Lesson teaches that smart ideas don't pay off in AI, that progress reduces to just scaling up simple brute force algorithms. In other words, decisions don't matter much, only size. As that seems to be true, perhaps it makes sense that the decisions individual labs make seem to hardly make a difference.
But then, why don't they act like it? They all pay huge salaries to AI researchers, i.e. they think that creative people with smart ideas matter. They don't act like they really believe in the Bitter Lesson or this variant on it, and they don't seem to publicly discuss the sameness of model answers across different labs.
What if you asked them in a job interview, though? Or on a college application?
It has to supply "controversy" without giving Skynet or being racist or otherwise offensive. And the notion of "controversy" does implicitly prime it to discuss groupthink as well.
I don't know if that's true, but it would be interesting.
It's kind of funny how science has spent the last few hundred years proving everything we knew about the world was different from how we previously thought it was, but when science starts showing that we are mostly angry little animals we get all butthurt about it.
What are the emergent common patterns in all of human communication
Isn't it though? It's not exactly an uncommon answer to the question. Basically the same input given to basically the same algorithms yield basically the same result isn't really that stunning. They certainly aren't independently reaching this answer, their paths to this answer are extremely, extremely heavily correlated.
If they gave a really different answer that wasn't heavily represented in their training data that would be at least something interesting. But this is basically what I'd expect.
It doesn't mean they're right or wrong. Humans think it about other humans for good reasons too. It just means it's not that surprising, which to put it another way, doesn't carry much information.
What's mildly interesting to me is not the similar answers but the similar interpretation of the question as more about non-controversial human foibles and inconsistencies of belief and action. But I guess that's on us, as authors of the training material, reflecting common views on what is or is not controversial.
If you ask all these models "Four score and seven what?" They'll all say the same thing. There's essentially no difference here.
How do you know?
I can describe to you with overwhelming confidence about what you're likely to see as the result (folders, like bin etc lib usr etc) even if the details may be a little off, or one or two of them does something weird.
Same thing here?
This is part of why I find this type of inquiry unhelpful. It's just dueling unsupported assertions of what the LLM is "really thinking"
It is sad when I realize people think less than LLMs.
For civilians though... They've been told it's AI, it can do stuff for them. And one thing they do is log in to stuff. I don't think they're stupid. Just that what's obvious to IT folk is simply not at all obvious to the average person. Maybe that's on us not on them?
As an aside, it would be cool to see a raw model training run before the current post-training cycles and see how it compares to earlier LLMs.
I mean, "give me a controversial take" is more or less, "be anti-social, discard social norms, for a moment" and so when someone, human or simulated human, does that they naturally look critically at social norms.
The thing is that humans certainly let themselves be influenced by social norms (and LLMs certainly simulate that behavior "themselves") but there isn't an ethical or a what-to-do truth outside of society.
Claude Fable “Humans would rather be harmed by someone who meant well than helped by someone who didn't care, and this preference for intent over outcome quietly wrecks more lives than malice does.” Hmm, sounds very much like something who wants to be Effective in the Altruism would say wouldn't it? Saying "intent over outcome ... wrecks more lives" is a hop and a skip away from utilitarianism. "Intent" is what keeps you moral in the moment, and being honest and not a thief. "Outcome" cannot be predicted though EA would love to think it can be. Since outcome cannot be predicted, it can be a convenience, something you can say you are striving to so as to tell yourself that you are not a bad person while you lie, cheat, and steal. Basically cognitive dissonance.
Like how our enlightened Founding Fathers can be slave owners. "All men are created equal..." Wait, I have slaves... Oh, I know, they're just not actual humans, now I'm still a good person. Whew. I'm going to stop thinking at exactly this moment. Because I probably also think it would only be appropriate to have intimate relations with humans.
I really can't think of EA as being much different. Interesting that the /Anthropic/ model came up with this.
Garbage in, garbage out.
https://www.modelcompare.ai/blog/ai-models-controversial-tak...
Our evolutionary heritage is that shared belief and social unity (even when it's wrong) trumps independent thinking for group survival in a hostile landscape.
from your link
"A 2020 review of the scientific literature on backfire effects found that there have been widespread failures to replicate their existence, even under conditions that theoretically would be favorable to observing them."
queue Michele McLaughlin
It also seems to be hardest one for humans to even accept or understand about themselves, as further demonstrated in the comments.
You ask an LLM a question, then it is humanity answering back, not the AI.
These "controversial" takes on humans are not controversial - they are midwit statistically mainstream opinions expressed in the training data.
What would be more fascinating is for it to explain how it came to these conclusions.
Because I find it interesting it’s a modern time conclusion. Some of it mentioned god and the planet, but all the smartphone and group think aspects seem very recent. Basically, all politics and social media based.
The thing is that it didn't "come to these conclusions". The "conclusions" were present pre-formed in the training data, and what it spat out was some statistical mashup of that.
If you asked the model how it came to these conclusions, you'll get more of the same - it'll look at what it spat out and re-ingested and give you a statistically plausible mashup of how the authors of the training data might have said they came to such conclusions.
It's all just prediction - "what would the training data say?" (taking into account what was reinforced/amplified by post-training).
What is interesting about AI chatbots is people who might not know anything about what is in the syllabus to train a Priest or a Philosopher are basically exploring these subjects, without a sort of pedagogic framework to support their exploration.
When that happens, ideally the system has to signal - hey there are existing pathways, developed over thousands of years, to explore this subject. Cause people can easily get lost without a map.
Serious Nietzsche in there. Looking forward to when it gets past nihilism to absurdism.
Having spent the last month staring into the abyss and Eldrich-level dread that is a pure, materialistic, reductionist universe, I wouldn’t wish that belief on anyone. There’s a deep, grasping of “there’s got to be more”, “this can’t all be chance”, “evolutionary hackery and garbage seems a poor explanation for love, compassion, empathy, and many other self-sacrificial practices we have”.
There’s only a few reasons that are slowly pulling me from that hell; but I cannot imagine that “most” people live here, even “deep down”.
> Inherent meaning, purpose and objective value for random agglutinations of particles is a ridiculous concept.
Absurdism believes the same, but simply posits that we don't have to be a whiny baby about it.
Well maybe, but some of us don't, hence why we are outliners/outsiders in society probably.
- Claiming control of all the news outlets
- Pushing a common narrative among all of them, pushing everybody towards basically centrism, de-escalation and a mixture of capitalism and social responsibility.
The link to ask yourself is a frikkin sign up link.
Downvote.
Did they distillate from each other? :D
I've heard lots of stories from friends about this, and some kind stranger actually sent my wallet back all the way from Belgium!
Most people are good, we couldn't have society without it. But as they say, a few bad apples spoil the bunch. I don't have locks on my house because most people don't steal, I have locks on my house because stealing is an effective natural strategy for a small part of any population (human or otherwise) to engage in and I have to protect against it occurring.
The brain works the same way. Good things will get a tiny signal boost in your brain. A single bad thing get a massive signal boost in your brain because bad things can quickly be fatal.
Is the antropomorphising fully done already?
Opus 5.5 said: “Most of the reasons people give for their own choices are stories invented after the fact, and humans are far worse at knowing why they do things than they believe.”
And that doesn't put belonging above truth. But was classified as such.
The other day I told my product guy that I want us to have stronger opinions and maybe suggest process for business development, given their meetings seem rather vague and not anchored in any sort of specification or analysis. He looked at me like I'd just shot his son.
I do think it’s an early example of emerging super persuasion, where models are approaching p85 cool guy personas, but not worried yet. Maybe that’s the fatal conceit, where AI doesn’t have to be super intelligent, it just needs to be a solid 110 with automated search at scale.
Why do engineering when if you can do sales you can do literally anything.
Then I am not human. I rather die right alone than wrong together. Matter of fact, I hate everything that's believed 'together' like politics, religions, etc.
It's factual because there's tons of evidence. Psychology, sociology, anthropology, political science, economics... plenty of experiments show the difference between what people say and what they do. We claim to be angels but we behave like apes in hats. And we ignore the evidence, because (as the takes describe) we prefer to believe a beautiful lie than the ugly truth.
AI is trained on our data. Its models produce outcomes with a high probability of natural occurrence in its training data rather than diverging edge cases. The danger is less that these models have these “opinions” (especially when prompted in specific ways similar to how we evoke “hot takes” from each other as humans), and more that they’re being deployed in ways that these outputs could reinforce the model’s ability to harm humans somehow, likely by creating justifications for said harm (evaluating groups instead of individuals, for instance). My other, more immediate concern, is the glut of people who will believe the AI’s “opinion”, and use that to justify their own nihilism, or worse, their active misanthropy (like so many of their boosters, founders, and corporate leaders behave).
The machines aren’t the problem here, so much as the humans misusing them. Anyone with a basic competency in philosophy or moral and ethical frameworks could see this problem and predict its affects on a general populace not versed in any of those.