How much of this is the similar, but not quite the same idea of "cell phones cause brain cancer, or, we'll, they used to, kinda".
Aka, people who can afford a cell phone in 1997 have good healthcare and don't get shot in street gang violence. They die of the hard to cure stuff. Once everyone got cell phones, the effect went away.
I've heard the "cell phones cause cancer" conspiracy theory before, but I've never heard that explanation of the phenomenon--very interesting! I think it's in the same ballpark as Scott's essay, but also think it's more neatly summarized as a case of the causal fallacy.
I remember anecdotal evidence of "oh yes he was always using his mobile phone and now he has a giant brain tumor on the right side of his head". I think there might have also been a few studies showing weak correlations, which is exactly what you'd expect if there were no true relationship and a whole lot of studies.
Did cell phone has that big of correlation to cancer to warrant that explanation? Because even with that phenomenon I think there would be too much noise to manifest in that kind of correlation. Because otherwise every single item that's slightly pricier correlates with brain cancer.
It might provide a way to tell apart tradeoffs from purely bad stuff.
If you increase the slack in a system (better healthcare, cheaper food) then the bad things that are caused by tradeoffs involving those things will increase (more cancer, more obesity).
> People may be poor because of “failures” - negative qualities with no counterbalancing advantages. For example, they may be unintelligent
Slightly tangential to your broader point: I'm not convinced that lower intelligence is strictly worse.
How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe? Personally, I know ~0 high IQ people who are content with such a life, but maybe that is sample bias?
It seems plausible to me that there is a tradeoff with IQ and it is not a categorical good (from an evolutionary perspective) because you can't have both the powerful trained/practiced pattern matching capability without having it always engaged, and when it is always engaged you run into problems being productive in the traditional sense of actually *doing* things that need to get done, because you are always looking for patterns and trying to "improve" things.
Historically, a society needs most of its people doing actual work and very few of them trying to think their way to a better future. If everyone was spending all of their energy thinking hard your society would invent a bunch of stuff in the first 6 months and then the all die of starvation.
Is that really so unnatural? The human brain is an very complex system that's effectively procedurally generated. Even extremely small changes to the starting conditions should produce large differences over time.
It is very rare for people to be born who are genuinely too stupid to live. Which is equivalent to, say, missing limbs. Human intelligence has already been selected to be, at the very least, usable. There is clearly room for improvement, but no method to actually select for that.
On farming, see Jon and his wife on "FarmCraft101" YouTube channel. They seem to find plenty of projects to occupy their minds, hands, and time on their nearly off-grid farm in rural Virginia.
These are modern humans in a modern context, not an evolutionary context. A modern set of humans with a YouTube channel, 600k subscribers, and a small family to feed definitely doesn't need to spend 14 hours a day hunting and gathering to make it through the winter.
I am not familiar with their channel, but my guess is that they aren't spending the majority of their life doing labor necessary for survival.
It *may* be correct to say that in the current evolutionary context, intelligence is more valuable than it used to be and our genes haven't caught up (we don't have the optimal ratio of "thinkers" vs "doers" in society). I still think it may be incorrect to say that intelligence is a categorical good, as I suspect our current society would fall apart if 100% of humans were "high IQ" today.
I will admit, I am very far from an expert on hunter gatherer tribes. While I do agree that they didn't spend 14 hours a day on any single task, I have been lead to understand that they didn't exactly have a huge amount of free time either. Perhaps not exclusively hunting or gathering, but also all of the other things a human needs to do to survive both summer and winter.
I initially read this as "They tend to spend a lot of time time traveling" and was briefly captivated by this new theory of hunter-gathering through time travel.
If you're singing and telling stories while weaving baskets, I call that both "useful work time" AND recreational time. You seem to have trouble with this concept, however.
Basic rundown, not the biased "it was only 20 hours a week." You can expect that during lean times, there was a lot more work, and a lot more starving.
Well, yes. But on the average, people worked less before agriculture was invented. Of course, they also tended to die more. Still, agriculture is strongly correlated with reduction is skeletal size.
I think labor done by hunter-gatherers is not super repetitive and boring, not in the way some modern jobs are. I don't think it would be particularly hard for a high-IQ person to spend an hour a day, say, stripping husks off some vegetable (presumably while talking to other people). And hunting and gathering benefit a lot from intelligence, as do many other tasks.
I agree that intelligence makes most tasks easier, and at the least it doesn't make any task more difficult. The question is whether higher intelligence makes a person less likely to engage in some tasks, and whether that effect is strong enough to cause an evolutionary pressure against maximizing intelligence.
I would also argue that the skills required to succeed at a hunting and gathering lifestyle were at least as complex and required as much individual knowledge as anything we do today. I argue this because they had the same brains we do, and people tend to push their brains to capacity regardless of the situation they are in. They had less institutional knowledge, to be sure, and what they had was less objectively accurate. The difference between us and them wasn't intelligence, it was shared media.
> The kind of work that is absolutely critical for a society to function
"But in Epsilons," said Mr. Foster very justly, "we don't need human intelligence."
> without having it always engaged
It’s 2026 and we have solved this problem; each of us has in our respective pockets a device capable of providing a continuous supply of chewing-gum for human brains at need. Welcome to our brave new world where apathy’s a tragedy and boredom is a crime!
But this is just another version of the same problem. If you're bored at work and spend all your time scrolling (or commenting on ACX) instead, it's probably not great for your career.
The initial assumption starting this thread is that there will always be some number of boring, dead-end jobs where you have to keep yourself amused somehow. I don't see a problem with e.g. the convenience store cashier checking his phone between customers and I don't think he or society would be better off if he were too dumb to be bored.
I'm not suggesting he should be too dumb to be bored, just that neither boring jobs or large amounts of scrolling are particularly good for people or society, even if we're stuck with those realities.
Perhaps ironically, this particular high IQ ADHD brain found the puzzle of "figure out the best way to bag this particular set of groceries, and get the answer as quickly as possible" interesting for the three days I worked as a supermarket cashier before getting fired for insubordination. 😅
> How many high IQ people do you know who would be content doing monotonous work every day for their whole life?
I'm not quite sure how many low IQ folks are "content" with this, as opposed to just not having better options. Isn't that why the lottery is so popular? Low IQ people who aren't content with their monotonous work?
Gambling tourism is a huge industry in both Israel and China. The chinese go to Macau while israelis gamble on cruise ships or travel to places like Georgia. And unlike most americans who tend to place small bets for fun israelis play huge amounts often ending up in debt to loan sharks.
My guess is that has nothing to do with IQ but with cultures that value money very highly.
Might just be in the sweet spot for forbidden fruit. Rabbinic Judaism very strongly frowns upon gambling while just barely not prohibiting it altogether.
Eating treif is also not completely prohibited by Rabbis, but apparently that requires specific rabbinical dispensation (as well as an explanation as to why you're doing it, and you're only allowed to do it in public, from what I understand.)
do you have a reference for that? I'm not a Jew, but my understanding was that eating treif was theoretically only allowed if it was necessary to preserve life-- nothing to do with "dispensation" or publicity.
Imagine a religion that was codified during the Middle Ages. The dispensation (when given by proper rabbinical authorities) is for spywork, where you have to blend in. There may be other reasons to give it, but that's the one I'm familiar with.
Remember, Jews only save lives of non-Jews (by breaking rabbinical laws like the sabbath) under the theory that if they don't, they'll get lynched (there's a very real argument that breaking the sabbath to save a Jewish life means more sabbaths in general get celebrated).**
**I'm sure someone has a reading that disagrees on this whole syllogism, but... this was the sort of thing I learned in Hebrew School.
I suppose this cuts to the core of the situation. A thing can be a "tradeoff" in one context while being categorically bad in another context. Perhaps IQ is a tradeoff at the evolutionary scale, but at the individual level it is categorically bad?
I suspect a *lot* of things are bad for the individual but good for the tribe, but then again I do tend to favor "group selection" theories of evolution in general.
Why would IQ be "categorically bad" at an individual level? I know that brains are metabolically expensive, but that's not a matter of IQ itself being bad such that if you could achieve a higher IQ without expending more calories that would hurt you.
The claim I was making is that maybe low IQ (not high, as your comment seems to imply) is categorically bad at an individual level. It may *also* be good to have lower IQ at an individual level in some cases, but that is a much harder argument to make since it does seem that high IQ seems to be generally better for the individual.
I am familiar with the high metabolic demand of the brain, but I'm not aware of any difference in energy demands between high and low IQ populations. A quick AI assisted search *suggests* that the two aren't correlated.
If we wanted to hypothesize, perhaps higher IQ leads to more anxiety? Someone whose brain is able to grasp problems of larger scales may suffer more anxiety about the future than people whose brains don't "wander" in such directions? This is purely speculative, but the very smart people I know do seem to fret a lot more about problems that are incredibly hard to solve, while the low IQ people I know fret about much simpler (and generally more transient) problems.
Anxiety is a problem of midwits. High IQ people learn that "randomness exists" and get used to dealing with it. When you let a midwit's brain run too hard, it's bad for them (leads to stress and strokes and heart disease). Ever wonder why our guidelines for salt usage are to "lower it"? It's not because it's bad for you... it's rather good for your brain, and geniuses need salt, rather a lot of it in fact.
High IQ (and I mean,HIGH) brains do run more (how many parallel thoughts can you run in your brain at once?), and to compensate, high IQ people train their bodies to run at a lower body temperature (running brains at too high a temperature leads to inefficiency).
These are some incredibly specific claims about "midwits" and biological processes. Do you have any citations for how average-IQ people have unusually high rates of heart disease compared to low and high IQs, or how people with high IQs need more salt?
Technically, "IQ" is your score on an IQ test, which mostly loads for academic performance. I think you probably mean intelligence in general, but then you will have to define that first before anyone can determine what the tradeoffs are.
Just know that experts have been trying to define it since IQ tests were invented, and no one has managed it yet.
Excuse me. YOUR mind may not do anything during a menial task, but a Genius' absolutely does. He's busy doing 4 other lines of thought -- give him a few hours a day to finish his novels (which he's more or less composed while doing something menial with his hands), and he's a happy camper.
This is why geniuses are much better at the whole "sensory deprivation" thing -- they'll make their own worlds, if you let them.
Yes, there is absolutely a tradeoff with IQ -- midwits suck, and are very bad at dealing with boredom ("I can't stand waiting in line" is a classic midwit line... is that you? it sounds like you. Try doing a hard problem while waiting in line, makes the time pass quicker).
Eh, depends on the queue. "He's with me" and a person willing to let you wander will do a lot for "I can wander and think" while the queue proceeds merrily along.
Besides, you can be entertainment for all the people bored in the queue (or you might try developing an internal dialogue...)
I used to do exactly that, but I found that it made me mentally exhausted at the end of the day. And my performance on complex problems is worse if I have to multitask.
haha. My friend the genius would say "that's the only way he can possibly drive" (all the multitasking is related to driving, mind you, so it's not like he's doing other things), due to motor coordination difficulties (slow reaction time). He's busy modelling all the other drivers, in a simulation of the road.
I think highly intelligent people NOT LIVING in industrial/post industrial society structures are more happy doing "daily productive work" because the nature of this work is different and so are its payoffs.
But even leaving that aside: You seem to be mixing up "happiness" with evolutionary fitness. Non starving is a powerful motivation, the idea that a high IQ person in a "starvation prone society" would just not see to their and their offspring survival first seems unlikely. I suspect in the most starving prone societies higher iq individuals are on average much better at non starving, other factors being equal.
High IQ people use a lot more calories on their brains. So, there's at least a reason to think that they might be less able to survive. You pretty much need "hacks" that they can devise, that someone else has Not Devised Beforehand (or, you know, being better at Mars* -- more persistent, less likely to "forget to cover tracks", more autistic in general).
*bit of shorthand, but if only 10% of people would remember to do everything needed to stay alive on Mars... Those are your "High IQ" people, and to the extent that "remembering to do everything right" matters, it also confers a survival bias.
> High IQ people use a lot more calories on their brains.
It's unlikely to be significant, and there's a good chance this is reversed -- they have studied chess masters in the middle of competitive matches, and the incremental calorie burn is only ~4 calories more per hour:
N. Troubat et al, "The stress of chess players as a model to study the effects of psychological stimuli on physiological responses" (2009)
Also, high IQ may actually point you the other direction in terms of calorie burn - this one looked at people doing memory problems, and found that poor performers spent 4.5x more calories than people who perform well on mental problems! (if you proxy by VO2, VO2 in low performers went up 22 ml/min vs 5 in high performers, both of these are tiny btw, over an hour it would be 6.6 cals and 1.5 cals respectively)
R.W. Blacks and K.A. Seljos, Metabolic and cardiorespiratory measures of mental effort... (1994)
IQ tests are a known stressor for low IQ people, yeah (was just reading about this, not citing unless you badly need it).
Not sure I believe chess masters are very smart, given they are playing midwit tic-tac-toe.
When you have people that have optimized their bodies to run at a lower temperature (overall) so that they can have a better functioning brain, well, those are the sort of HighIQ folks I'd want to VO2 in order to see if using brain at capacity would cause an increase in VO2 (over normies).
I rather suspect any problem set that would strain a HighIQ person would leave most normies (or midwits) staring and saying "I don't do that, that's impossible."
>High IQ people use a lot more calories on their brains
This isn't true. If anything the evidence shows the opposite. The brains of higher IQ people have lower glucose consumption both at rest and during cognitively demanding tasks. Smart people have more efficient neurons.
I'll take that smart people have more efficient neurons, although I'm far from convinced of that -- "innate skills" like riding a bike are pruned and efficient, mylenated too. Someone using intelligence(g) to ride a bike (as opposed to the learned skill) is probably running a lot more neurons -- and they're going to be a lot worse at the task, anyhow.
Or, say, driving a car -- most people's version of driving a car doesn't involve simulating the mental state and projected "next 30 seconds" of everyone on the road (or if they do so, they do it without putting the level of uncertainty that a genius must put into it).
This, of course, is setting aside a high IQ person's tendency to be working on about 6 tracks of thought at once... I'm not sure you -can- replicate that with a normal IQ person (hence why you might see more brain activation, in the truly intelligent).
Your point about happiness vs. fitness is well made. But I think the most successful people at not starving are the extroverts and people with large, tight support networks.
Extroverts and "tight support networks" are two different groups of people. Your most successful extroverts are probably bards (in terms of reproductive fitness, anywhoo) and they're the ones who move from village to village because someone's always noticing they tupped the innkeep's daughter. Very entertaining people though, and pretty useful if you need a dispute settled.
I know they are separate, but the most successful extroverts in those days were probably the story tellers that stayed at home. Today we call those people who lived back then "Shamen", though I'm not sure if that's really the most correct term.
This is hilarous, because I was just thinking about by adult ADHD diagnosis. Mainly it is hyperactivity, so the brain finds most things boring, so it keeps looking for new things or just spins on its own. Very low monotony tolerance. But reading what you say, it now sounds like high IQ in itself means hyperactivity and thus ADHD, perhaps, high IQ can even be defined as a hyperactive and thus high-output brain, so ADHD just reduces to being smart?
Actually not - all the people who do not have this mild ADHD I have, "restless overthinking explorer" but very serious issues that prevent them from getting a job and needing some level of assisted living are not high IQ.
But on mild levels, maybe. They kept telling in my childhood I could be an excellent student if I paid more attention, and I told them I find memorizing the oil ports of the world and putting them on a blank map super boring and hard to pay attention to it.
If you're able to focus on tasks you consider productive, ADHD is probably not the right diagnosis. This is one of those "not well defined" problems where psychologists take "kid having problems" and give them "pills to help" regardless of whether the diagnosis is right.
High IQ folks get three lines of thought at once, and are productive on all three. is that you?
What does productivity mean in this context of three lines of thought? Write an essay? Take some kind of action? Does "thinking it over, then hitting a wall of no evidence, so filing it for later investigation" count as productive?
If "thinking it over" means "coming up with multiple possible solutions, developing "what to look for later" " --and importantly, remembering what you put on the "Big list of mysteries" so that you can mark them solved, that's a perfectly productive thought pattern.
Writing an essay in your head, or outlining it, is also productive.
I was just reading about that Selective Attention Test (the one with the gorilla suit wearing man in the background) and I wondered if ADHD folks have a higher tendency to see the gorilla.
> How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe?
Thank goodness I'm not trapped in that hell. I get to work in enterprise software engineering, where I move protobufs around.
I'm very willing to do mindless labor that lets me think while I work because I will just listen to audiobooks and be happy. What I'm not willing to do is get paid 30k and be poor forever and be micromanaged like a stupid/lazy person.
I am utterly willing to talk to anyone who wants to pay me middle class wages to do data entry, sets me a reasonable quota, and then fucks off and leaves me to it.
Is IQ correlated to the brain size? Because then it's definitely a tradeoff. In fact, we have smaller brains on average than both Cro-Magnons of 50 kya and Neanderthals.
It doesn't seem obvious to me that what you are describing is IQ rather than an appetite for novelty. I could see something like ADHD being a tradeoff in the way you describe, but even if you are fishing, more IQ should let you do that better.
It is certainly possible that the two things are not causal! From my sample set, the people willing/interested in doing monotonous work is very strongly correlated with IQ. I know some exceptions in the middle-ish, but they are essentially non-existent (in my sample) at the extremes. People with very low IQ (in my experience) seem to hate doing "thinking jobs" and people with very high IQ seem to hate doing non-thinking jobs.
Of course, this could be sample bias related to the sets of people I interact with in each group, so I am curious if others have a different experience.
The late Jerry Pournelle worked in a machine shop in WWII. He observed that they wanted someone with an IQ of 120 for every job except tightening bolts. For that you wanted someone functionally mentally retarded, who would take longer to train, but work for the duration, where someone who learned it more quickly would get bored and quit.
That's all the tradeoff you need. There's your categorical 'good' from the evolutionary point of view--more children. If a gene increases fertility, it tends to get passed on, regardless of how miserable it makes the carrier.
> I wonder in the past how much the average IQ growth was caused by smart men having lots of illegitimate kids.
I think this is very probably true. It's difficult to triangulate with illegitimacy, but we can triangulate with a simple rich vs poor cut. At least in England, the fertility differential was pretty explicit. Greg Clark has measured that pre-Industrial-Revolution England rich families had higher fertility than poor families. Surviving children was ~2x higher in rich families vs poor (with "poor" being the bottom 80-90%), and this was in effect for a pretty long time, over 1500-1800, so 300 years.
If we model this simply as 4 vs 2 descendants respectively, the "rich descendants" are ~50% of descendants at about a hundred years in (1600 ish) and 90%+ of descendants at 1700.
But this is too simplistic - if 80-90% of the population was at replacement, only the top 10-20% have a genuinely higher carrying capacity, and we need to model wealth dilution at inheritance and the fact that only the top 10% each generation will have the genuinely 2x carrying capacity. If you add these factors, you hit "rich descendants" crowding out the poor to a 4x extent over the 300 years, or ~80% rich descendants to ~20% poor descendants in 1800 (when you started reversed). And presumably, "merit" would have been being concentrated that whole time.
One could imagine a world (I'm not sure it aligns with our reality) where intelligence allows you to override your basal instincts more, and this leads to prioritizing your individual desires/needs over the ones that evolution has programmed for you. This could lead to smart people not having as many children because they believe that they will be a burden on them (individually), while less intelligent people are more likely to follow their instincts which tell them to have as much sex as possible.
That would make sense. Probably some variant of time preference; i.e. smart people think ahead but impulsive sex leads to kids (and misery) but the genes get passed on.
> "How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe? Personally, I know ~0 high IQ people who are content with such a life, but maybe that is sample bias?"
I don't know if this is historically true. It's often said that our society is unprecedented in the degree to which it tries to connect high-IQ people with high-IQ occupations ("Meritocracy", "The Great Sort", "Coming Apart", etc). If you look back in history, you can find lots of yeoman farmers who wrote books or had interesting ideas, and surely there were even more who were just the smartest person in their village but never did anything that got recorded.
It also seems like people in the past just had lower standards, and this worked fine-ish for them. You don't hear people complaining about nobody ever taking a bath, or about women having ten children without anaesthetic or any kind of medical care, because nobody expected life to be tolerable, so it didn't seem worth worrying about the fact that it wasn't. I think high-IQ people being forced into unstimulating jobs was probably the same kind of thing.
Also, it's not clear to me that low-IQ people like menial unstimulating jobs any better than high-IQ people, they just don't have a lot of other options and don't write as much literature hand-wringing about how they hate them.
I think the reason why a lot of low IQ people dislike menial unstimulating jobs is because for them the jobs actually require more intelligence than they have. For example, if you work at a place like Walmart, you have to navigate a surprising amount of corporate policy, dynamic situations with customers, and social give-and-takes with managers who might vary in professionalism. I think there are very few jobs that are easy to do from a cognitive standpoint in modern times if you have below a 100 IQ. If you know them, the complaints poor working-class people tend to have about these jobs usually revolve less around boredom and more around how much these jobs are stressing them out.
Also, I do not know if I would call myself 'high IQ' but count me in the 'would enjoy unstimulating jobs' bin. I would not enjoy the 'being poor' part of them. But if I could count pebbles for 8 hours a day and be solidly middle class doing this, I would be very happy with this. I am autistic in a very developmental way and I always had a notoriously spiky intelligence so the comfort with menial work likely connects to that in some way.
> if I could count pebbles for 8 hours a day and be solidly middle class doing this, I would be very happy with this.
And you would be content to do this your whole life? Would you choose this even if there was an option to instead solve complex/hard problems 8 hours a day for the same pay?
Probably. I get stressed out easily when there are high external expectations and I spend my free time reading and doing a lot of intellectual stuff anyway.
This suggests that perhaps my issue is just sample bias. Your writing (or more generally your thought process) suggests that you are either an LLM or someone who I would categorize as "high IQ" (at the least, not below average) and so my expectation would have been that you would dislike that sort of work.
When given the opportunity, the low IQ people I work with will choose work with less "thinking" (i.e., more menial work) if the pay is the same. I actively try to encourage all of my employees to better their situation and learn more complex skills, and you can nearly sort people by IQ based purely on how interested they are in taking on more complex workloads.
> If you look back in history, you can find lots of yeoman farmers who wrote books or had interesting ideas
A guy who published over 40 books once told me that he wrote his first books working as a night guard in a factory.
The job required him to be there all night long, and once in an hour take a flashlight and do a short walk around the factory area. He took a notebook there, and spent the rest of the time writing fantasy novels.
> Slightly tangential to your broader point: I'm not convinced that lower intelligence is strictly worse.
We know that it isn't. You can use a much simpler argument than the one you present: intelligence shows high levels of natural variation, and this is sufficient to demonstrate that there hasn't been a stable evolutionary advantage at any particular level of it.
(There is a fallback argument of "well, of course, but the tradeoff used to be that intelligence has high energy requirements, which are no longer burdensome in the modern world". If that argument were correct, there would have been powerful recent selection for intelligence.)
I believe intelligence is going up over time in humans, which would support your argument that there is a recent selection for it (perhaps correlated with food abundance).
While I like evolutionary arguments for hypothesis generation, I find them uncompelling as a form of evidence for something though. It is too easy to construct an evolutionary argument that fits just about any narrative, and there is much disagreement on the nature and mechanism of evolution.
For example, one could argue that there is high variance in intelligence because there is a tradeoff and a lineage/group/genome that has high variance is more likely to survive/thrive than one that optimizes purely in one direction within that tradeoff space.
Another way to say this from the evolutionary perspective, our uniquely social species evolved a repertoire of traits that exist on bell shaped distributions. It is the distribution of those traits that is adaptive. We need a distribution of IQ, compassion, aggression, masculinity/femininity, deference, heroism, etc., for society to cohere.
When we look at IQ how does common sense fit into the picture. I have seen many uneducated people become successful. I have seen many educated people unable to deal with small common problems that were out of their grasp of the system.
G enables "common sense" solutions. Fail algebra four times, and you Find A Solution, eventually, even if it is 2's complement. Midwits tend to be over-educated in our society, and they're the ones that fail "small common problems."
(Exception: g also is better at "problems which allow some creativity" -- asking someone with high g and dyscalcula to repeat a number back to you, when they haven't trained on that, has gotten "mindreading is easier and more predictive, so I'm doing that instead")
Sorry, I lapsed into Robin Hanson Blog Post Title-ese. It's asking "How Natural [Are] Tradeoff And Failure Components?", where components is used in the statistical sense as clusters of causes.
Is there any reasoning why "are" can be just left out without making people scratch their heads? I'm good at english, but not a native speaker and couldn't make out the missing word in my head. I guess I was further confused by the possibility of "natural" just being an adjective for "tradeoff".
When examining tradeoff as regards medical conditions it is worth looking at which conditions have been selected for in the last 5,000 years of evolution and which are linked to harm in early life like head injuries.
To add a social layer to this, I believe it's bad for culture to make rules that make it easy for people to cheat. Because the failure modes and tradeoffs are tough for institutions to distinguish, the cheating effectively ruins things for everybody because some people who have chosen the tradeoff willingly will nonetheless see their way to passing themselves off as suffering from a failure mode.
For example, compensating veterans who have lost limbs is a failure mode that is hard to cheat. Compensating veterans suffering from PTSD is a failure mode that is easy to cheat. From a systemic perspective, it would be nice if we could compensate the failure modes in both, but compensating only the failure modes in the first case is easy and in the second case hard.
One would imagine that this might be an unobjectionable principle, but I see it violated all the time. The people who do institutional design and draft rules seem almost never to be considering these dynamics, although I believe they are obvious to the people responsible for administering public benefits.
Suffer a bullet right by your head, and we can quantify what just happened to your brain. No need to call it PTSD, it's a freakin' shock wave and it damages your brain.
My impression (from Greg Cochran) is that PTSD wasn't noticed much in the Civil War, where a huge percentage of Americans were veterans. Instead it was explosive artillery that made the difference in WW1 with "shell shock". https://westhunt.wordpress.com/2018/08/18/ptsd/
It's... a bit of both. For one thing, you had "actual trauma" from the medicine of the time (so if someone commits suicide, later, because they can't help on the family farm, it's more likely to be seen as "warranted" and not "because of a major depression.")
I find that folks are likely to project onto others, the current diagnosis, but it's also true that we have written records of "what actually happened" (and it would have served nobody's interest to "make up" that the man of the house was barricaded in his room with live weapons, thinking the war was still going on. I might note it was more likely to be a "bad dream/nightmare" plus sleepwalking, than the article gives it credit for.) I also think it likely that there were fewer problems per soldier in the Civil War
.Cochran's wrong about parenting not causing problems, of course. Narcissism is the classic -- "demanding parents plus kids who can't actually measure up." But parenting doesn't cause all problems, far from it.
> Not all scholars applaud this trend, which includes new scholarship on subjects such as rape, torture and guerrilla atrocities. “All these dark elements describe the margins not the mainstream of Civil War experience,” says Gary Gallagher, a historian at the University of Virginia who has authored and edited over 30 books on the war. While he welcomes the fresh research, he worries that readers may come away with a distorted perception of the overall conflict. The vast majority of soldiers, he adds, weren’t traumatized and went on to have productive postwar lives.
Compatible with your claim of "fewer problems per soldier in the Civil War", but I would go further and say that without data we don't even know if suicide rates were higher for Civil War veterans vs peers who didn't serve.
> Cochran's wrong about parenting not causing problems, of course.
No, not "of course". Twin/adoption studies find little evidence of "shared environment" having much effect. There are some extreme cases of practically feral children who don't acquire language, but outside of that even surviving the Holocaust as a child doesn't seem to have much effect.
I was not expecting to see significantly higher suicide rates in Active duty personnel in the 1800s, given the high probability of undercounting due to shame/disapproval.
(This, again, isn't specific to the Civil War, and the suicide rates go UP closer to 1900, which again would counter any plausible idea I'd come up with).
Also, weren't they supposed to believe that suicide is a mortal sin and a good Christian just carries his cross for life? Or am I overestimating their religiosity? Like, maybe, although formal atheism was not a thing, just quietly not caring was a thing?
On the shared environment thing, it's probably important to distinguish between 'cause' in the sense that you'd see a statistical increase, and 'cause' in the sense of being a critical component of the explanatory pathway.
Imagine you flip a coin: heads the kids copy exactly what their parents do, tails they "rebel" and do the exact opposite. In either case, it feels natural to say the parents are the reason behind the child's behavior. But the coin flip is also causal, and it wipes out any statistical correlation
So even if statistically-speaking shared environment doesn't cause shit, I'm hesitant to throw causal narrativization out the window
The difference in PTSD rates between 19th century warfare and World War One may have less to do with the direct effects of the artillery itself, and more to do with the change in how soldiers experienced stress.
If you are campaigning with the Army of the Potomac, there are specific, clearly defined days when you are "in danger" versus "not in danger." Most of the threats to your survival are things you can theoretically see before they kill you, or at least identify as "that place over there where all the gunfire's coming from." While you or other members of your group are in danger, the danger normally happens in a way that you can retaliate against (shoot back). Your position and situation often move around, so that you are not constantly stuck sitting near the bodies of your former comrades. You spend much of your time in a structured environment (the army on the march and in camp) that acts as a "normal" where danger is distant or absent.
In World War One, infantry deployed to the front are continuously in danger from shellfire, which may arrive at any time on nearly zero notice. You cannot see it before it is already falling on or near your position, and at best you get less than a minute's notice to take cover. Bombardments may go on for days, drowning out any hope of living normal life even by the rugged standards of "we are camping out in a trench in a field." Your position is often static, dead bodies often stay near you for extended periods. Periods of continuous danger last longer.
I could easily see a significant rise in PTSD from just the raw differences between being a WWI infantryman and being an infantryman in the American Civil War.
I think prior to World War One there would have been other conflicts where you wouldn't know on which day you'd be in danger. Particularly ones involving fighting irregulars.
Yes, but pre-WWI threats involving irregulars were usually the kind where the irregulars jump out of the bushes and start shooting, or just as often jump out of the bushes and charge you with swords. The threat was personalized and often coming from close enough range that merely normal human vigilance could provide some measure of protection. Also, not all types of terrain make it realistic that you could be ambushed at truly any time.
That, too, is very different from being in a situation where it is predictable that (from the perspective of your inner caveman-brain) the sky is going to mysteriously rain loud whistling explosions that send shrapnel raining across the landscape to kill everyone exposed in the open, and that this is going to go on for days, forcing you to cower in a trench with six inches of water in the bottom, where somebody's hand is sticking out of the wall because this stretch of trench collapsed a month ago under bombardment and buried them alive.
I'm not saying that previous forms of warfare do not produce fear, or do not produce traumatizing PTSD, in any of the combatants who fought in them. But it is easy to see that World War One was very different from nearly any previous war (indeed, plenty of people fighting in it commented on exactly that!) And that many of these differences took forms that it would be at least superficially plausible to expect as having a psychological impact.
I thought even at its worst, around 50% of soldiers suffer PTSD. It's a lot, yes, but it also means that 50% *don't* suffer PTSD. I think I remember the statistics imply that it's not all soldiers, but ones under shockwaves specifically, so I find that statistic more surprising.
It means it's not death sentence and we may try to exploit it to find ways to reduce PTSD without decreasing deployment. For instance it'd be very good if we can identify soldiers who're three explosions short of developing PTSD, so we can regulate a more effective rotation mechanism. But still, if it's truly random we're truly out of luck.
I have no idea, honestly. I just pulled some numbers on suicide. PTSD is... a little different, sometimes.
If a car backfires, and someone drops to the ground and reaches for a gun that isn't there, is that PTSD? Or is that well-warranted caution/instincts bred from being endangered?
How about someone who hears a cannon fired at them, and flips the table (sending a laptop crashing to the ground), while they hit the deck? (Erm. This incident occurred inside a well-built (pre 1950) building. The odds of the cannonball entering the building were slight).
Both of these are "problematic assessments of risk" but I'm not sure they meet the technical definition of PTSD.
Yes, I do agree,it would be nice if we could figure out who's one bullet away from PTSD, but the Army likes robot dogs better -- and really, who can blame them?
That's a tradeoff too! Making it easy to cheat at something is often a side-effect of making it easy to do legitimately, and at some point you have to accept some background level of cheating to not burden the legitimate users too much. See e.g. https://www.bitsaboutmoney.com/archive/optimal-amount-of-fraud/ for another example.
True. And this is something I've thought about too.
If your system can distinguish 100 categories but the frauds cluster into 8 behavioral modes, you're wasting investigative energy on distinctions that don't matter. If your system can only handle 3 categories but the population exhibits 30 distinct fraud signatures, you're blind.
So this could be operationalized: An agency could measure the *mutual information* (specifically, KL divergence) between its risk scores and actual fraud outcomes, then iteratively adjust its categorization scheme to maximize that mutual information per unit of investigative cost, maybe with some buffer.
You could also measure entropy and volatility of signals themselves, not just summary statistics, which are easier to game.
There is actually a lot of potential for AI and AI generated algos to help reduce fraud and expand available services on the same operating budget.
I think this turns a useful rule of thumb into a rigid principle that refuses to deal with hard situations. It's all nice and well to say something like "we'll compensate people for lost limbs, which is impossible to fake, and not for epilepsy, which is easy to fake" until you get epilepsy and nobody will help you.
I don't think the designers of these systems are too dumb to notice that cheating is bad, I think they've decided that they'll do the best they can, but that potentially allowing a few cheaters is less bad that ignoring every malady that doesn't have an unmistakable physical component.
I didn't mean to suggest that it should be too rigid. In hell, due process is meticulously observed. And I definitely see and appreciate that the people responsible on the ground are usually doing the best they can.
Ultimately, I guess my point is that in designing these systems, we need to budget for the cost of maintaining them, including the cost of making the measurements necessary to suss out the frauds. Too often, we design systems with the best intentions, and then watch them ruined by the handful of bad actors who abuse them. This is what I meant by easy to cheat.
The actual costs of maintaining a system by preventing an unreasonable rate of exploitation by frauds are often much lower than would be imagined by someone who expects massive fraud to dominate the system and its outcomes.
The current kerfluffle over voter identification presents a good example. Clearly, the US election system must be protected against having millions of noncitizens voting. It turns out the actual cost of preventing noncitizens from voting in US elections at any rate plausibly capable of having an impact on US political outcomes is "very close to zero," though. Because the rate at which this actually occurs is already very close to zero; there are not, in fact, millions of noncitizens voting in the elections.
The trouble is that the need to provide protection against frauds and cheats must be measured in terms of outcomes, as opposed to emotional security.
The question of outcomes is "how much does it cost to actually succeed in doing this." How much does it cost to ensure that money genuinely wasted on actual frauds makes up only 1%, or 0.1%, or 0.01% of the overall budget of the operation?
The terms of emotional security are more nebulous. "How much does it cost to make me stop being afraid?" "How much does it cost to make the man on TV with the very convincing haircut stop telling me that this is a problem?"
Whether I feel unafraid about a potential 'danger' may have little connection to whether the danger exists and to how much more resources society needs to invest in it. My personal fear of ghosts does not mean that society needs to spend billions on systems to protect us all from evil spirits. Whether the man with the nice haircut stops claiming that there is a problem is probably has little to do with whether the problem has been solved, or even existed in the first place. It has much more to do with his own incentives and with those of whoever owns the news channel.
Many people overestimate how easy it is to "cheat" a system in blatant fashion (e.g. "fake having PTSD") because they can easily imagine a Just-So story in which a cheater prospers, and it instinctively makes them angry, and so they assume it must be real and a threat to their way of life. In practice, this leads to situations where the public is easily manipulated into approving actions that are bad for normal people, but good for skilled manipulators that have access to billion-dollar bank accounts from which to build propaganda.
There's another sense in which everything has tradeoffs: the cost of preventing a certain bad outcome can easily be worse than the original outcome. We could solve the problem of bad pizza by simply shooting every chef who baked a bad pie; our immune system could prevent cancer by immediately terminating every cell that looked at them cross-eyed.
> the cost of preventing a certain bad outcome can easily be worse than the original outcome
Non-hypothetical examples: War on Drugs and earlier, Prohibition. Most harm associated with drugs is a direct or indirect result of their illegality, not of their physiological effects. You don't see any liquor sellers shooting each other -- although during Prohibition they did.
19th century, not 18th. And you are correct, Qing (and later, Communists) thought opium was a scourge. But that was mostly because UK *forced* opium onto China via a shockingly humiliating military defeat. Qing government did not have any issue with opium earlier, when it was all homegrown.
I am not saying Prohibition or War on Drugs are (or were) an "unmitigated negative". Their purpose, at least ostensibly, was to prevent certain bad outcomes, and they at least somewhat succeeded. It is just that the side effects were much worse.
It was not just gang violence -- poisonings by methanol went way up. There were many good reasons Prohibition was cancelled after 13 years.
(low-effort ITT failure, to be sure, but it's what immediately sprang to mind)
More importantly, this is cornmeal crust erasure. Vigolo's would like a word.
I think there's probably also a comment to be made comparing the "tradeoffs" we see all the time in evolution with "failures" leading to extinction, but that's getting out over my skis and I don't feel confident enough not to assert something stupid. Iterated games complicate the dynamics here.
All multifactorial problems should naturally be combinations of tradeoffs and failures.
hahahaha.
You're no longer walking the savannah. So there's one extremely obvious idea you're missing:
Formerly adaptive mutations that no longer have the triggers that caused the adaptation to be functional. You no longer have the diet we were adapted for (and, sometimes, the gut bacteria). A very adaptive mutation might very well cause problems -- if they only cause, say, non-childhood cancer, you don't really starve the gene out of the population either.
Pale skin and freckles are very functional adaptations to northern climates and covering ones' skin for most of the year (in particular, the Vitamin D generation at the end of winter, a traditional time for starvation and illness), but if you're living at the equator, you're no longer getting that boost, and, if you're a redhead, still have higher cancer incidence.
I’m confused, isn’t this a tautology? Or at least for nearly so? Things that are bad break into components: parts that are simply bad, or parts that are bad with a trade-off of good. Either you have some goodness tradeoff or you don’t. Since you’re starting with the thing, that is bad, all your components necessarily include a little bit of that thing that is bad. Is there some depth here that I’m missing?
No, it's absolutely not a tautology. Scott's missing entire swathes of solutions (as I noted above in my post directly to the OP).
You can have a brain that is smarter, say, with enough salt -- but deprive it of salt, and it's suddenly WORSE than others. Yet, you evolved for a salt-rich diet, and seeing it in the perspective of "bad bad" or "trade-off bad" is completely ruining the paradigm. It was a GOOD mutation, and it got selected for, and all it's really waiting for is more salt, an environmental change.
Being better in one type of environment and worse in a different environment is a tradeoff. Probably a very common evolutionary tradeoff, if the environment tends to remain stable long enough for natural selection.
It's no longer the same tradeoff, though. You can't say "it's a bit better for sun exposure and a bit worse on cancer" (like, say, being a redhead). You get "this sucks" or "this is great" and it's all environmental. You can't take a maladaptive "current state" and say whether it's been adaptive or not, in prior states, based on the current one.
In which case, you've got three states:
1) Adaptive.
2) Tradeoff (some parts good, some parts bad)
3) Maladaptive (but not always lethal)
and, here's the key point, you can't figure out which a particular gene (set) is, based on your current data (one "current thing" that is historically weird is "early menarche" at age 12-13ish, not age 18ish, and you better believe that has influence on how problematic "diseases that start in late teens-early twenties" are in terms of reproductive success.)
I think you're smooshing two layers of analysis together here. Some trade-offs are adaptive, some trade-offs or maladaptive. 'Adaptive' and 'Trade-off' are not mutually exclusive categories.
Yeah, I think OP is doing so as well, probably driven by his perspective on psychiatric disorders. Consider that some psychiatric disorders may actually be adaptive, no matter how unpleasant they feel at the time.
I'm a bit confused by this. Much earlier menarche means that it's physically possible for girls (specifically) to reproduce earlier. Later menarche (say, age 18) means that a disease that kicks in when she turns 20 will affect a woman during a higher proportion of her lifetime period of fertility. Earlier menarche means she theoretically has several years in which to have one or more children before the disease hits. That part makes sense...
...But if early menarche is a <i>novel</i> phenomenon, something historically weird, then we must convolute that with the fact that teen pregnancy rates in developed societies are very low. While it may well be for all I know that medieval 14 year olds girls usually hadn't had menarche yet, whereas modern 14 year olds are theoretically able to get pregnant... in <i>neither</i> period does this result in statistically large percentage of women having children at the age of fourteen.
Which means that "theoretically you can now have kids before this disease affects you" isn't actually going to impact reproductive success very much.
Well, the finer point on it is "the better the food supply, the earlier the menarche" (so Israel could and did have menarche years earlier than the Norwegians). And yeah, that does fun things to genetics, doesn't it just?
Teen pregnancy rates are historically lower than they've been since... early 1800s or so? So you might have had a flowering of "this is possible"... dunno? Just spitballing.
I'm afraid I don't know how to actually read the article; I'd be curious to know the details of the claim that Paleolithic skeletal evidence shows signs of menarche age being from 14 down to the extremely low value of 7.
This post is saying "I wrote of this in the past as very complicated stuff, but if I restate it like this it's just obvious" so no surprise if you find it obvious :)
It's certainly possible to have both higher cancer risk (because of DNA damage) and at the same time higher degeneration burden (because cell senescence is adaptive in high-DNA-damage context). For example, smoking or radiation can plausibly increase both.
The "faulty DNA repair -> cancer and faulty DNA repair -> aging" framing is basically the figure 1 of Wolf's tumor suppression theory of aging paper (distilling Judith Campisi's research direction).
So through this lens, the ovarian aging findings would be interpreted as "ovaries have adaptations that initiate defensive degeneration in high-cancer-risk context".
As they write: "Cancer risk-increasing alleles in SAMHD1 were associated with later [menopause age], following a similar pattern demonstrated previously for CHEK2. This finding is consistent with a mechanism of disrupted DNA damage sensing and apoptosis, resulting in slowed depletion of the ovarian reserve".
In other words, when you disrupt tumor-suppressors SAMHD1 and CHEK2, the ovaries senescence is delayed, but the cancer risk is increased. That's the tradeoff. Basically shifting leftward on the Wolf illustration, but not upward.
My favorite story in this regard about Henry Ford, the Model T and king pins. Let me just copy and paste from wiki.
"Henry Ford, it is said, commissioned a survey of the car scrap-yards of America to find out if there were parts of the Model T Ford which never failed. His inspectors came back with reports of almost every kind of failure: axles, brakes, pistons – all were liable to go wrong. But they drew attention to one notable exception, the kingpins of the scrapped cars invariably had years of life left in them. With ruthless logic Ford concluded that the kingpins on the Model T were too good for their job and ordered that in future they should be made to an inferior specification.[3]"
Slightly related historical trivia: There is exactly one part of the Model T that is interchangeable with a modern Ford: The valves for letting air into the tires.
That's based on, as I recall, a 19th century patent that was essentially perfect as-is and never changed.
The title is a joke where by virtue of a random deletion, it has failed. The question is whether the resulting comments are a positive, determining whether the title failure is a trade off or simple failure
The title is hard to parse. Should there be an "Are" in there? (I realize that it's valid English to make the "are" implicit, but it doesn't seem like you're going for something poetic here.)
Speaking of amputations, you hardly ever see people with missing arms anymore. I would attribute this to several factors, at least in the US: major wars receding into the past, better safeguards on industrial and agricultural equipment (not to mention fewer individuals working in those jobs), and greater ability to reattach severed arms.
The terse headline-style title of this one reminds me of a (probably apocryphal) exchange of telegrams between a newspaper reporter doing a short bio piece and the actor Cary Grant.
From what I can tell, the IQ variable in the Splitting Schizophrenia paper isn’t based on a clinical IQ test, but on UK Biobank cognitive data which was brief, unsupervised touchscreen tasks measuring things like reaction time, memory, and simple reasoning. These are then combined into a statistical proxy for general cognitive ability.
While that proxy has some validity at scale, it’s still a relatively noisy and limited measure compared to standardized IQ tests like the WAIS. So the conclusions are interesting, but they rest on an indirect definition of IQ.
Oooh, I really don't know that being unintelligent is all failure with no tradeoff. It's not what I'd choose but there's certainly room for debate on that front.
There's room for debate, but the subject is greatly complicated by the fact that "being intelligent" often manifests itself in ways that are socially constructed, either by the broader society or by the subculture that the intelligent person is a member of.
An intelligent person who has willingly or unwillingly internalized behaviors that cause them to self-sabotage may fail to thrive in situations where a less intelligent person who lacks the self-sabotage does thrive. And they may blame their "greater intelligence" for their failure to thrive. But in reality, the problem may be something else.
Rates of suicidality are higher in smarter people, smarter people tend towards fewer children, smarter people tend to have fewer completely idiotic sexual encounters. All of these are disadvantages
I wouldn't say I have a very solid "theory of ethics" but to the extent I do, it's based in this tradeoff principle. CS Lewis wrote that human desires are like the keys of a piano. They make a horrible, discordant cacophony when played randomly, but beautiful music when played by an expert pianist (obviously he meant God, but I don't).
Aristotle's Golden Mean rightly gets a lot of flak, but I think it's groping towards this as well. Often two goods are in competition. Where this doesn't happen, there's really nothing to talk about (Imagine positing a Virtue of Taking Breaths in the Furtherance of Life. Who are you having this conversation with? Who disagrees?). Where it does, you must balance the goods.
You can also model this economically - systems have inputs and outputs. They compete with each other for inputs and factor this competition into their design. So when the competitor is weakened or removed, they may produce too much output and muck up the system.
And failures are...still this, tbh. Like a small chance of falling off a cliff is the cost for having functional gravity, a necessary component of our physical system.
The problem isn't that some things are good and some things are bad. The problem is that we have competing goods, in every sense of the word "good." We have goods that are bad in excess, maybe because they block other goods. Or we have goods that we can't get enough of because another system is hogging them. Or we have goods that are necessary constraints on their systems, but do unfortunate things in specific cases. Ultimately, you have to figure out what goods you want to purchase (an extremely hard problem), figure out what they cost in other goods, and then figure out if that's a reasonable price.
This sounds a bit tautological. Like, anything CAN be "decomposed" into factors, the tricky thing is deciding which specific factors are useful and practically speaking how to do that decomposition.
All diseases can be decomposed into "ice cream factors" (to the extent that their development is correlated with increased ice cream consumption) and "monotony factors (the opposite). Type 2 diabetes and some gut biome issues have a high ice cream factor, cystic fibrosis doesn't.
All qualities can be decomposed into skub and non-skub factors.
I feel like talking about "the real components" of things in this way is similar to talking about "the real basis" of a vector space. There are many ways to describe something, you choose between them based on what you actually want to do (in which case, one basis might be situationally very nice) and not on which one is "the most ontologically appropriate" (which isn't unique).
How do you tell if it is genuinely a case of trade-off in the sense that certain good things can only be realized with a certain risk (eg sickle cell) as opposed to simply a beneficial mutation and harmful mutation that arose in the same lineage and the beneficial mutation was useful enough that the harmful mutation has yet to be selected out of the population?
I'm not sure what you're asking - I'm not claiming there's a natural way to know through philosophy alone. I think you'd have to do the research and figure it out.
I took you to be suggesting that we already had done said research and had reason to believe both effects were real. Based on what you said it wasn't clear to me that the evidence is clear enough to be suggestive of that. It is certainly a hypothesis worth testing, I just thought you were suggesting we already had reasonable evidence for it.
My sense is that to really resolve this we need to identify the specific genes at issue so we can seperate their effects, eg, is it just that there is a gene which increases psychosis risk and lowers IQ/edu attainment and it is correlated with another that increases IQ/edu attainment or do some psychosis increasing genes do that without the negative effects on IQ/edu attainment.
Governments can be thought of as varying mixtures of tradeoffs and failures. Communism is pure failure, as it fails to even optimize its stated objective of the well being of the proletariat better than literally anywhere else on the spectrum of economic freedom. Housing construction regulations and rent and wage controls are mostly-failure. Tariffs are a mostly-failure inefficently subsidizing domestic manufacturers at the expense of domestic consumers and foreign consumers at the expense of foreign manufacturers. The indoor smoking ban or the requirement to register vehicles is a legit tradeoff between freedom and limiting externalities.
I am the last to defend communism, but I am not sure that "pure failure" is the right frame. For example, East Germany was able to embark on a massive homebuilding project that delivered significant improvements to housing quality and cost. Singapore is not typically thought of as a "communist country", but the government owns almost all the land and over 75% of Singaporeans live in public housing.
There is some work to understand aging and cancer as a general trade-off that multicellular organisms face, which itself can be framed as the trade-off between coordination and competition within a multiagent system (as arises across natural scales).
Scott: I'm having trouble making grammatical sense of the title of your post, "How Natural Tradeoff And Failure Components?" Shouldn't there be a verb in there? "How Natural *ARE* Tradeoff and Failure Components?"
I just read the tradeoffs-vs-failure post and my main takeaway was that, because of the thing about March-conceived babies, you found a scientific basis for astrology-based character traits
I guess you've already touched it at the last part, but still when a model seems too universal, I immediately wanted to ask what kind of scenario can debunk it. Because if there are none, then it proves too much and is useless.
I think people have a better intuitive model of romance than of, for example, the way that t-cells release cytokines. If what we're actually talking about is t-cell cytokine release, I think an analogy to romance would shed light on the situation.
Fascinating stuff. As someone who works in urban education, I can say many failing institutions in urban education wouldn’t be failing to this degree had the subjects of these institutions not made tradeoffs that left them poor.
I think that whether people default to a heuristic that assumes tradeoff etiologies or default to a heuristic that assumes failure etiologies when it comes to psychiatric and sociological problems tells you a lot about their political ideology. If most psychiatric and sociological problems involve failure etiologies, then this casts the world into a hierarchy where people at the bottom drag everyone else down and have little hope. If most psychiatric and sociological problems involve tradeoff etiologies, then people who struggle or make poor choices are the natural consequence of a population that is well-rounded, resilient to environmental changes, and produces different people who are useful in different ways.
> The reason it keeps happening is that the gene for muscle protein is really big - so if you get a random deleterious mutation, it’s pretty likely to be there!
Well, that goes back to your previous point: there's probably a smaller protein that's less risky, but also doesn't work as well. So the trade-off is amongst the risks again.
Both is a fun and clean answer to the question wether mental illnesses are purely detrimental or have trade off capacities.
I am a John Nash esque schizophrenia case, now holding a pretty good law degree, and I have felt a strong aversion to polygenetic screening before, because I strongly felt that eliminating the schizophrenia gene strain entirely would be a massive mistake.
I am fine with pruning it, cutting selectively only the parts that make me retarded.
The scarier version of the tradeoff/failure model: schizophrenia risk genes increase educational attainment via creativity, and the gene that makes you more likely to think in broken patterns also makes you more likely to build something new. A lot of civilization might be tradeoff genes running at scale.
There is no objective "good". Something may seem to have no tradeoffs; time will tell. If the tradeoff is effective, as contrasted to good, then it will be more likely to persist and become more common.
How much of this is the similar, but not quite the same idea of "cell phones cause brain cancer, or, we'll, they used to, kinda".
Aka, people who can afford a cell phone in 1997 have good healthcare and don't get shot in street gang violence. They die of the hard to cure stuff. Once everyone got cell phones, the effect went away.
I've heard the "cell phones cause cancer" conspiracy theory before, but I've never heard that explanation of the phenomenon--very interesting! I think it's in the same ballpark as Scott's essay, but also think it's more neatly summarized as a case of the causal fallacy.
IIUC, there was never any real evidence that "cellphones cause cancer". It was just people being frightened of something new.
OTOH, I did worry about it for awhile, because there also wasn't any good evidence that they didn't.
I remember anecdotal evidence of "oh yes he was always using his mobile phone and now he has a giant brain tumor on the right side of his head". I think there might have also been a few studies showing weak correlations, which is exactly what you'd expect if there were no true relationship and a whole lot of studies.
Did cell phone has that big of correlation to cancer to warrant that explanation? Because even with that phenomenon I think there would be too much noise to manifest in that kind of correlation. Because otherwise every single item that's slightly pricier correlates with brain cancer.
I don't think it's similar. I'm proposing that these tradeoffs are real at the biological level. But I might be misunderstanding you.
It might provide a way to tell apart tradeoffs from purely bad stuff.
If you increase the slack in a system (better healthcare, cheaper food) then the bad things that are caused by tradeoffs involving those things will increase (more cancer, more obesity).
This sounds like a neat theory, but you'd need to be really simple minded to fail to control for basic things like wealth.
It's a bit like saying that preventing heart attacks causes cancer.
> People may be poor because of “failures” - negative qualities with no counterbalancing advantages. For example, they may be unintelligent
Slightly tangential to your broader point: I'm not convinced that lower intelligence is strictly worse.
How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe? Personally, I know ~0 high IQ people who are content with such a life, but maybe that is sample bias?
It seems plausible to me that there is a tradeoff with IQ and it is not a categorical good (from an evolutionary perspective) because you can't have both the powerful trained/practiced pattern matching capability without having it always engaged, and when it is always engaged you run into problems being productive in the traditional sense of actually *doing* things that need to get done, because you are always looking for patterns and trying to "improve" things.
Historically, a society needs most of its people doing actual work and very few of them trying to think their way to a better future. If everyone was spending all of their energy thinking hard your society would invent a bunch of stuff in the first 6 months and then the all die of starvation.
Is that really so unnatural? The human brain is an very complex system that's effectively procedurally generated. Even extremely small changes to the starting conditions should produce large differences over time.
It is very rare for people to be born who are genuinely too stupid to live. Which is equivalent to, say, missing limbs. Human intelligence has already been selected to be, at the very least, usable. There is clearly room for improvement, but no method to actually select for that.
On farming, see Jon and his wife on "FarmCraft101" YouTube channel. They seem to find plenty of projects to occupy their minds, hands, and time on their nearly off-grid farm in rural Virginia.
These are modern humans in a modern context, not an evolutionary context. A modern set of humans with a YouTube channel, 600k subscribers, and a small family to feed definitely doesn't need to spend 14 hours a day hunting and gathering to make it through the winter.
I am not familiar with their channel, but my guess is that they aren't spending the majority of their life doing labor necessary for survival.
It *may* be correct to say that in the current evolutionary context, intelligence is more valuable than it used to be and our genes haven't caught up (we don't have the optimal ratio of "thinkers" vs "doers" in society). I still think it may be incorrect to say that intelligence is a categorical good, as I suspect our current society would fall apart if 100% of humans were "high IQ" today.
Hunter and gatherers didn't spend 14 hours a day hunting and gathering. not even the ones who spent 4 hours running the gazelle to death.
I will admit, I am very far from an expert on hunter gatherer tribes. While I do agree that they didn't spend 14 hours a day on any single task, I have been lead to understand that they didn't exactly have a huge amount of free time either. Perhaps not exclusively hunting or gathering, but also all of the other things a human needs to do to survive both summer and winter.
They tend spend a lot of time traveling.
I initially read this as "They tend to spend a lot of time time traveling" and was briefly captivated by this new theory of hunter-gathering through time travel.
If you're singing and telling stories while weaving baskets, I call that both "useful work time" AND recreational time. You seem to have trouble with this concept, however.
Nearly all "recreational" (and religious) activities were communal bond building activities, which were very necessary for survival.
https://scienceinsights.org/did-hunter-gatherers-really-work-only-20-hours-a-week/
Basic rundown, not the biased "it was only 20 hours a week." You can expect that during lean times, there was a lot more work, and a lot more starving.
Well, yes. But on the average, people worked less before agriculture was invented. Of course, they also tended to die more. Still, agriculture is strongly correlated with reduction is skeletal size.
The starving during lean times is what keeps the population density low enough to keep the workload during fat times low.
A settled agrarian society has much less of that lean time filter. For better or worse.
They do, however, spend nearly all that time pursuing the necessities of survival, because survival is hard.
I think labor done by hunter-gatherers is not super repetitive and boring, not in the way some modern jobs are. I don't think it would be particularly hard for a high-IQ person to spend an hour a day, say, stripping husks off some vegetable (presumably while talking to other people). And hunting and gathering benefit a lot from intelligence, as do many other tasks.
I agree that intelligence makes most tasks easier, and at the least it doesn't make any task more difficult. The question is whether higher intelligence makes a person less likely to engage in some tasks, and whether that effect is strong enough to cause an evolutionary pressure against maximizing intelligence.
I would also argue that the skills required to succeed at a hunting and gathering lifestyle were at least as complex and required as much individual knowledge as anything we do today. I argue this because they had the same brains we do, and people tend to push their brains to capacity regardless of the situation they are in. They had less institutional knowledge, to be sure, and what they had was less objectively accurate. The difference between us and them wasn't intelligence, it was shared media.
> The kind of work that is absolutely critical for a society to function
"But in Epsilons," said Mr. Foster very justly, "we don't need human intelligence."
> without having it always engaged
It’s 2026 and we have solved this problem; each of us has in our respective pockets a device capable of providing a continuous supply of chewing-gum for human brains at need. Welcome to our brave new world where apathy’s a tragedy and boredom is a crime!
But this is just another version of the same problem. If you're bored at work and spend all your time scrolling (or commenting on ACX) instead, it's probably not great for your career.
The initial assumption starting this thread is that there will always be some number of boring, dead-end jobs where you have to keep yourself amused somehow. I don't see a problem with e.g. the convenience store cashier checking his phone between customers and I don't think he or society would be better off if he were too dumb to be bored.
I'm not suggesting he should be too dumb to be bored, just that neither boring jobs or large amounts of scrolling are particularly good for people or society, even if we're stuck with those realities.
Ideally, he or she should be helping manage the store, but then they would have to pay them more.
Perhaps ironically, this particular high IQ ADHD brain found the puzzle of "figure out the best way to bag this particular set of groceries, and get the answer as quickly as possible" interesting for the three days I worked as a supermarket cashier before getting fired for insubordination. 😅
I'm sure there's a lesson in there for all of us. ; )
> How many high IQ people do you know who would be content doing monotonous work every day for their whole life?
I'm not quite sure how many low IQ folks are "content" with this, as opposed to just not having better options. Isn't that why the lottery is so popular? Low IQ people who aren't content with their monotonous work?
I see a lot of jews playing the lottery, and they're supposed to have quite high iqs. have statistics led me wrong again?
Gambling tourism is a huge industry in both Israel and China. The chinese go to Macau while israelis gamble on cruise ships or travel to places like Georgia. And unlike most americans who tend to place small bets for fun israelis play huge amounts often ending up in debt to loan sharks.
My guess is that has nothing to do with IQ but with cultures that value money very highly.
Might just be in the sweet spot for forbidden fruit. Rabbinic Judaism very strongly frowns upon gambling while just barely not prohibiting it altogether.
Eating treif is also not completely prohibited by Rabbis, but apparently that requires specific rabbinical dispensation (as well as an explanation as to why you're doing it, and you're only allowed to do it in public, from what I understand.)
do you have a reference for that? I'm not a Jew, but my understanding was that eating treif was theoretically only allowed if it was necessary to preserve life-- nothing to do with "dispensation" or publicity.
Imagine a religion that was codified during the Middle Ages. The dispensation (when given by proper rabbinical authorities) is for spywork, where you have to blend in. There may be other reasons to give it, but that's the one I'm familiar with.
Remember, Jews only save lives of non-Jews (by breaking rabbinical laws like the sabbath) under the theory that if they don't, they'll get lynched (there's a very real argument that breaking the sabbath to save a Jewish life means more sabbaths in general get celebrated).**
**I'm sure someone has a reading that disagrees on this whole syllogism, but... this was the sort of thing I learned in Hebrew School.
I suppose this cuts to the core of the situation. A thing can be a "tradeoff" in one context while being categorically bad in another context. Perhaps IQ is a tradeoff at the evolutionary scale, but at the individual level it is categorically bad?
I suspect a *lot* of things are bad for the individual but good for the tribe, but then again I do tend to favor "group selection" theories of evolution in general.
IQ can be very good for an individual. They can become a "specialist" due privileges, for example.
Why would IQ be "categorically bad" at an individual level? I know that brains are metabolically expensive, but that's not a matter of IQ itself being bad such that if you could achieve a higher IQ without expending more calories that would hurt you.
The claim I was making is that maybe low IQ (not high, as your comment seems to imply) is categorically bad at an individual level. It may *also* be good to have lower IQ at an individual level in some cases, but that is a much harder argument to make since it does seem that high IQ seems to be generally better for the individual.
I am familiar with the high metabolic demand of the brain, but I'm not aware of any difference in energy demands between high and low IQ populations. A quick AI assisted search *suggests* that the two aren't correlated.
If we wanted to hypothesize, perhaps higher IQ leads to more anxiety? Someone whose brain is able to grasp problems of larger scales may suffer more anxiety about the future than people whose brains don't "wander" in such directions? This is purely speculative, but the very smart people I know do seem to fret a lot more about problems that are incredibly hard to solve, while the low IQ people I know fret about much simpler (and generally more transient) problems.
Anxiety is a problem of midwits. High IQ people learn that "randomness exists" and get used to dealing with it. When you let a midwit's brain run too hard, it's bad for them (leads to stress and strokes and heart disease). Ever wonder why our guidelines for salt usage are to "lower it"? It's not because it's bad for you... it's rather good for your brain, and geniuses need salt, rather a lot of it in fact.
High IQ (and I mean,HIGH) brains do run more (how many parallel thoughts can you run in your brain at once?), and to compensate, high IQ people train their bodies to run at a lower body temperature (running brains at too high a temperature leads to inefficiency).
These are some incredibly specific claims about "midwits" and biological processes. Do you have any citations for how average-IQ people have unusually high rates of heart disease compared to low and high IQs, or how people with high IQs need more salt?
Technically, "IQ" is your score on an IQ test, which mostly loads for academic performance. I think you probably mean intelligence in general, but then you will have to define that first before anyone can determine what the tradeoffs are.
Just know that experts have been trying to define it since IQ tests were invented, and no one has managed it yet.
Is anyone out there actually "content?" Is that a thing humans do?
It should be noted that the advent of farming seems to be associated with an increase in PGS measures for IQ in ancient DNA.
"Where your mind is not doing anything."
...
Excuse me. YOUR mind may not do anything during a menial task, but a Genius' absolutely does. He's busy doing 4 other lines of thought -- give him a few hours a day to finish his novels (which he's more or less composed while doing something menial with his hands), and he's a happy camper.
This is why geniuses are much better at the whole "sensory deprivation" thing -- they'll make their own worlds, if you let them.
Yes, there is absolutely a tradeoff with IQ -- midwits suck, and are very bad at dealing with boredom ("I can't stand waiting in line" is a classic midwit line... is that you? it sounds like you. Try doing a hard problem while waiting in line, makes the time pass quicker).
LOL. But thinking about hard problems requires pacing around like a lion in a cage, and loudly arguing with myself, hard to do in a queue :)
Eh, depends on the queue. "He's with me" and a person willing to let you wander will do a lot for "I can wander and think" while the queue proceeds merrily along.
Besides, you can be entertainment for all the people bored in the queue (or you might try developing an internal dialogue...)
I used to do exactly that, but I found that it made me mentally exhausted at the end of the day. And my performance on complex problems is worse if I have to multitask.
Oh, and please don't do this while driving.
haha. My friend the genius would say "that's the only way he can possibly drive" (all the multitasking is related to driving, mind you, so it's not like he's doing other things), due to motor coordination difficulties (slow reaction time). He's busy modelling all the other drivers, in a simulation of the road.
I think highly intelligent people NOT LIVING in industrial/post industrial society structures are more happy doing "daily productive work" because the nature of this work is different and so are its payoffs.
But even leaving that aside: You seem to be mixing up "happiness" with evolutionary fitness. Non starving is a powerful motivation, the idea that a high IQ person in a "starvation prone society" would just not see to their and their offspring survival first seems unlikely. I suspect in the most starving prone societies higher iq individuals are on average much better at non starving, other factors being equal.
You make a good point here, differentiating between happiness and fitness!
High IQ people use a lot more calories on their brains. So, there's at least a reason to think that they might be less able to survive. You pretty much need "hacks" that they can devise, that someone else has Not Devised Beforehand (or, you know, being better at Mars* -- more persistent, less likely to "forget to cover tracks", more autistic in general).
*bit of shorthand, but if only 10% of people would remember to do everything needed to stay alive on Mars... Those are your "High IQ" people, and to the extent that "remembering to do everything right" matters, it also confers a survival bias.
> High IQ people use a lot more calories on their brains.
It's unlikely to be significant, and there's a good chance this is reversed -- they have studied chess masters in the middle of competitive matches, and the incremental calorie burn is only ~4 calories more per hour:
N. Troubat et al, "The stress of chess players as a model to study the effects of psychological stimuli on physiological responses" (2009)
Also, high IQ may actually point you the other direction in terms of calorie burn - this one looked at people doing memory problems, and found that poor performers spent 4.5x more calories than people who perform well on mental problems! (if you proxy by VO2, VO2 in low performers went up 22 ml/min vs 5 in high performers, both of these are tiny btw, over an hour it would be 6.6 cals and 1.5 cals respectively)
R.W. Blacks and K.A. Seljos, Metabolic and cardiorespiratory measures of mental effort... (1994)
IQ tests are a known stressor for low IQ people, yeah (was just reading about this, not citing unless you badly need it).
Not sure I believe chess masters are very smart, given they are playing midwit tic-tac-toe.
When you have people that have optimized their bodies to run at a lower temperature (overall) so that they can have a better functioning brain, well, those are the sort of HighIQ folks I'd want to VO2 in order to see if using brain at capacity would cause an increase in VO2 (over normies).
I rather suspect any problem set that would strain a HighIQ person would leave most normies (or midwits) staring and saying "I don't do that, that's impossible."
>High IQ people use a lot more calories on their brains
This isn't true. If anything the evidence shows the opposite. The brains of higher IQ people have lower glucose consumption both at rest and during cognitively demanding tasks. Smart people have more efficient neurons.
I'll take that smart people have more efficient neurons, although I'm far from convinced of that -- "innate skills" like riding a bike are pruned and efficient, mylenated too. Someone using intelligence(g) to ride a bike (as opposed to the learned skill) is probably running a lot more neurons -- and they're going to be a lot worse at the task, anyhow.
Or, say, driving a car -- most people's version of driving a car doesn't involve simulating the mental state and projected "next 30 seconds" of everyone on the road (or if they do so, they do it without putting the level of uncertainty that a genius must put into it).
This, of course, is setting aside a high IQ person's tendency to be working on about 6 tracks of thought at once... I'm not sure you -can- replicate that with a normal IQ person (hence why you might see more brain activation, in the truly intelligent).
I think you're confusing IQ and intelligence. Lots of high IQ people are rather low in practical intelligence.
Your point about happiness vs. fitness is well made. But I think the most successful people at not starving are the extroverts and people with large, tight support networks.
Extroverts and "tight support networks" are two different groups of people. Your most successful extroverts are probably bards (in terms of reproductive fitness, anywhoo) and they're the ones who move from village to village because someone's always noticing they tupped the innkeep's daughter. Very entertaining people though, and pretty useful if you need a dispute settled.
I know they are separate, but the most successful extroverts in those days were probably the story tellers that stayed at home. Today we call those people who lived back then "Shamen", though I'm not sure if that's really the most correct term.
This is hilarous, because I was just thinking about by adult ADHD diagnosis. Mainly it is hyperactivity, so the brain finds most things boring, so it keeps looking for new things or just spins on its own. Very low monotony tolerance. But reading what you say, it now sounds like high IQ in itself means hyperactivity and thus ADHD, perhaps, high IQ can even be defined as a hyperactive and thus high-output brain, so ADHD just reduces to being smart?
Actually not - all the people who do not have this mild ADHD I have, "restless overthinking explorer" but very serious issues that prevent them from getting a job and needing some level of assisted living are not high IQ.
But on mild levels, maybe. They kept telling in my childhood I could be an excellent student if I paid more attention, and I told them I find memorizing the oil ports of the world and putting them on a blank map super boring and hard to pay attention to it.
If you're able to focus on tasks you consider productive, ADHD is probably not the right diagnosis. This is one of those "not well defined" problems where psychologists take "kid having problems" and give them "pills to help" regardless of whether the diagnosis is right.
High IQ folks get three lines of thought at once, and are productive on all three. is that you?
What does productivity mean in this context of three lines of thought? Write an essay? Take some kind of action? Does "thinking it over, then hitting a wall of no evidence, so filing it for later investigation" count as productive?
If "thinking it over" means "coming up with multiple possible solutions, developing "what to look for later" " --and importantly, remembering what you put on the "Big list of mysteries" so that you can mark them solved, that's a perfectly productive thought pattern.
Writing an essay in your head, or outlining it, is also productive.
ADHD people are very much able to focus on tasks their brains currently consider productive.
What they're bad at is choosing what to focus on at will.
I was just reading about that Selective Attention Test (the one with the gorilla suit wearing man in the background) and I wondered if ADHD folks have a higher tendency to see the gorilla.
> How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe?
Thank goodness I'm not trapped in that hell. I get to work in enterprise software engineering, where I move protobufs around.
I'm very willing to do mindless labor that lets me think while I work because I will just listen to audiobooks and be happy. What I'm not willing to do is get paid 30k and be poor forever and be micromanaged like a stupid/lazy person.
I am utterly willing to talk to anyone who wants to pay me middle class wages to do data entry, sets me a reasonable quota, and then fucks off and leaves me to it.
Is IQ correlated to the brain size? Because then it's definitely a tradeoff. In fact, we have smaller brains on average than both Cro-Magnons of 50 kya and Neanderthals.
> Is IQ correlated to the brain size?
It is fairly strongly correlated with brain size, at about r=0.4.
https://www.sciencedirect.com/science/article/abs/pii/S0160289616303385
Men and women have different intercepts, and normalized-IQ women vs men have smaller brains sizes.
It doesn't seem obvious to me that what you are describing is IQ rather than an appetite for novelty. I could see something like ADHD being a tradeoff in the way you describe, but even if you are fishing, more IQ should let you do that better.
It is certainly possible that the two things are not causal! From my sample set, the people willing/interested in doing monotonous work is very strongly correlated with IQ. I know some exceptions in the middle-ish, but they are essentially non-existent (in my sample) at the extremes. People with very low IQ (in my experience) seem to hate doing "thinking jobs" and people with very high IQ seem to hate doing non-thinking jobs.
Of course, this could be sample bias related to the sets of people I interact with in each group, so I am curious if others have a different experience.
The late Jerry Pournelle worked in a machine shop in WWII. He observed that they wanted someone with an IQ of 120 for every job except tightening bolts. For that you wanted someone functionally mentally retarded, who would take longer to train, but work for the duration, where someone who learned it more quickly would get bored and quit.
Well, apparently there's a negative correlation between fertility and intelligence:
https://en.wikipedia.org/wiki/Fertility_and_intelligence
And that's Wikipedia, which is PC as heck.
That's all the tradeoff you need. There's your categorical 'good' from the evolutionary point of view--more children. If a gene increases fertility, it tends to get passed on, regardless of how miserable it makes the carrier.
I wonder in the past how much the average IQ growth was caused by smart men having lots of illegitimate kids.
> I wonder in the past how much the average IQ growth was caused by smart men having lots of illegitimate kids.
I think this is very probably true. It's difficult to triangulate with illegitimacy, but we can triangulate with a simple rich vs poor cut. At least in England, the fertility differential was pretty explicit. Greg Clark has measured that pre-Industrial-Revolution England rich families had higher fertility than poor families. Surviving children was ~2x higher in rich families vs poor (with "poor" being the bottom 80-90%), and this was in effect for a pretty long time, over 1500-1800, so 300 years.
If we model this simply as 4 vs 2 descendants respectively, the "rich descendants" are ~50% of descendants at about a hundred years in (1600 ish) and 90%+ of descendants at 1700.
But this is too simplistic - if 80-90% of the population was at replacement, only the top 10-20% have a genuinely higher carrying capacity, and we need to model wealth dilution at inheritance and the fact that only the top 10% each generation will have the genuinely 2x carrying capacity. If you add these factors, you hit "rich descendants" crowding out the poor to a 4x extent over the 300 years, or ~80% rich descendants to ~20% poor descendants in 1800 (when you started reversed). And presumably, "merit" would have been being concentrated that whole time.
https://imgur.com/a/ZrtodA8
You find similar but less extreme differentials for pre-IR times for rich vs poor families in Japan and a number of other countries.
One could imagine a world (I'm not sure it aligns with our reality) where intelligence allows you to override your basal instincts more, and this leads to prioritizing your individual desires/needs over the ones that evolution has programmed for you. This could lead to smart people not having as many children because they believe that they will be a burden on them (individually), while less intelligent people are more likely to follow their instincts which tell them to have as much sex as possible.
That would make sense. Probably some variant of time preference; i.e. smart people think ahead but impulsive sex leads to kids (and misery) but the genes get passed on.
> "How many high IQ people do you know who would be content doing monotonous work every day for their whole life? The kind of work that is absolutely critical for a society to function, like fishing or farming, where your mind is not doing anything but you are consistently producing value for your tribe? Personally, I know ~0 high IQ people who are content with such a life, but maybe that is sample bias?"
I don't know if this is historically true. It's often said that our society is unprecedented in the degree to which it tries to connect high-IQ people with high-IQ occupations ("Meritocracy", "The Great Sort", "Coming Apart", etc). If you look back in history, you can find lots of yeoman farmers who wrote books or had interesting ideas, and surely there were even more who were just the smartest person in their village but never did anything that got recorded.
It also seems like people in the past just had lower standards, and this worked fine-ish for them. You don't hear people complaining about nobody ever taking a bath, or about women having ten children without anaesthetic or any kind of medical care, because nobody expected life to be tolerable, so it didn't seem worth worrying about the fact that it wasn't. I think high-IQ people being forced into unstimulating jobs was probably the same kind of thing.
Also, it's not clear to me that low-IQ people like menial unstimulating jobs any better than high-IQ people, they just don't have a lot of other options and don't write as much literature hand-wringing about how they hate them.
I think the reason why a lot of low IQ people dislike menial unstimulating jobs is because for them the jobs actually require more intelligence than they have. For example, if you work at a place like Walmart, you have to navigate a surprising amount of corporate policy, dynamic situations with customers, and social give-and-takes with managers who might vary in professionalism. I think there are very few jobs that are easy to do from a cognitive standpoint in modern times if you have below a 100 IQ. If you know them, the complaints poor working-class people tend to have about these jobs usually revolve less around boredom and more around how much these jobs are stressing them out.
Also, I do not know if I would call myself 'high IQ' but count me in the 'would enjoy unstimulating jobs' bin. I would not enjoy the 'being poor' part of them. But if I could count pebbles for 8 hours a day and be solidly middle class doing this, I would be very happy with this. I am autistic in a very developmental way and I always had a notoriously spiky intelligence so the comfort with menial work likely connects to that in some way.
> if I could count pebbles for 8 hours a day and be solidly middle class doing this, I would be very happy with this.
And you would be content to do this your whole life? Would you choose this even if there was an option to instead solve complex/hard problems 8 hours a day for the same pay?
Probably. I get stressed out easily when there are high external expectations and I spend my free time reading and doing a lot of intellectual stuff anyway.
This suggests that perhaps my issue is just sample bias. Your writing (or more generally your thought process) suggests that you are either an LLM or someone who I would categorize as "high IQ" (at the least, not below average) and so my expectation would have been that you would dislike that sort of work.
When given the opportunity, the low IQ people I work with will choose work with less "thinking" (i.e., more menial work) if the pay is the same. I actively try to encourage all of my employees to better their situation and learn more complex skills, and you can nearly sort people by IQ based purely on how interested they are in taking on more complex workloads.
> If you look back in history, you can find lots of yeoman farmers who wrote books or had interesting ideas
A guy who published over 40 books once told me that he wrote his first books working as a night guard in a factory.
The job required him to be there all night long, and once in an hour take a flashlight and do a short walk around the factory area. He took a notebook there, and spent the rest of the time writing fantasy novels.
> Slightly tangential to your broader point: I'm not convinced that lower intelligence is strictly worse.
We know that it isn't. You can use a much simpler argument than the one you present: intelligence shows high levels of natural variation, and this is sufficient to demonstrate that there hasn't been a stable evolutionary advantage at any particular level of it.
(There is a fallback argument of "well, of course, but the tradeoff used to be that intelligence has high energy requirements, which are no longer burdensome in the modern world". If that argument were correct, there would have been powerful recent selection for intelligence.)
I believe intelligence is going up over time in humans, which would support your argument that there is a recent selection for it (perhaps correlated with food abundance).
While I like evolutionary arguments for hypothesis generation, I find them uncompelling as a form of evidence for something though. It is too easy to construct an evolutionary argument that fits just about any narrative, and there is much disagreement on the nature and mechanism of evolution.
For example, one could argue that there is high variance in intelligence because there is a tradeoff and a lineage/group/genome that has high variance is more likely to survive/thrive than one that optimizes purely in one direction within that tradeoff space.
> which would support your argument that there is a recent selection for it (perhaps correlated with food abundance).
That's not an argument I made. I said the opposite.
> I believe intelligence is going up over time in humans
Why?
>> I believe intelligence is going up over time in humans
> Why?
The Flynn Effect.
But the Flynn Effect doesn't reflect an increase in intelligence.
(It also appears to be reversing, but that's much less relevant than the fact that it doesn't measure the thing you're claiming it measures.)
What do you believe it reflects?
Another way to say this from the evolutionary perspective, our uniquely social species evolved a repertoire of traits that exist on bell shaped distributions. It is the distribution of those traits that is adaptive. We need a distribution of IQ, compassion, aggression, masculinity/femininity, deference, heroism, etc., for society to cohere.
I think this is a good generalization of the claim I was making.
When we look at IQ how does common sense fit into the picture. I have seen many uneducated people become successful. I have seen many educated people unable to deal with small common problems that were out of their grasp of the system.
G enables "common sense" solutions. Fail algebra four times, and you Find A Solution, eventually, even if it is 2's complement. Midwits tend to be over-educated in our society, and they're the ones that fail "small common problems."
(Exception: g also is better at "problems which allow some creativity" -- asking someone with high g and dyscalcula to repeat a number back to you, when they haven't trained on that, has gotten "mindreading is easier and more predictive, so I'm doing that instead")
Nothing worse than an educated idot!
Don't understand the title. Is components the correct word?
I'm pretty sure it's missing an "are". But title typos are rare for Scott, so I too find myself trying to elucidate some more-complex meaning.
Seconding. I still don’t get it.
I had to read it two or three times and then decided to just skip it and go to the article.
Sorry, I lapsed into Robin Hanson Blog Post Title-ese. It's asking "How Natural [Are] Tradeoff And Failure Components?", where components is used in the statistical sense as clusters of causes.
Is there any reasoning why "are" can be just left out without making people scratch their heads? I'm good at english, but not a native speaker and couldn't make out the missing word in my head. I guess I was further confused by the possibility of "natural" just being an adjective for "tradeoff".
English is my first language and I have never seen this phrasing before. Sat at the top of the post for a minute or so trying to parse it.
When examining tradeoff as regards medical conditions it is worth looking at which conditions have been selected for in the last 5,000 years of evolution and which are linked to harm in early life like head injuries.
To add a social layer to this, I believe it's bad for culture to make rules that make it easy for people to cheat. Because the failure modes and tradeoffs are tough for institutions to distinguish, the cheating effectively ruins things for everybody because some people who have chosen the tradeoff willingly will nonetheless see their way to passing themselves off as suffering from a failure mode.
For example, compensating veterans who have lost limbs is a failure mode that is hard to cheat. Compensating veterans suffering from PTSD is a failure mode that is easy to cheat. From a systemic perspective, it would be nice if we could compensate the failure modes in both, but compensating only the failure modes in the first case is easy and in the second case hard.
One would imagine that this might be an unobjectionable principle, but I see it violated all the time. The people who do institutional design and draft rules seem almost never to be considering these dynamics, although I believe they are obvious to the people responsible for administering public benefits.
Suffer a bullet right by your head, and we can quantify what just happened to your brain. No need to call it PTSD, it's a freakin' shock wave and it damages your brain.
My impression (from Greg Cochran) is that PTSD wasn't noticed much in the Civil War, where a huge percentage of Americans were veterans. Instead it was explosive artillery that made the difference in WW1 with "shell shock". https://westhunt.wordpress.com/2018/08/18/ptsd/
https://www.smithsonianmag.com/history/ptsd-civil-wars-hidden-legacy-180953652/
It's... a bit of both. For one thing, you had "actual trauma" from the medicine of the time (so if someone commits suicide, later, because they can't help on the family farm, it's more likely to be seen as "warranted" and not "because of a major depression.")
I find that folks are likely to project onto others, the current diagnosis, but it's also true that we have written records of "what actually happened" (and it would have served nobody's interest to "make up" that the man of the house was barricaded in his room with live weapons, thinking the war was still going on. I might note it was more likely to be a "bad dream/nightmare" plus sleepwalking, than the article gives it credit for.) I also think it likely that there were fewer problems per soldier in the Civil War
.Cochran's wrong about parenting not causing problems, of course. Narcissism is the classic -- "demanding parents plus kids who can't actually measure up." But parenting doesn't cause all problems, far from it.
> Not all scholars applaud this trend, which includes new scholarship on subjects such as rape, torture and guerrilla atrocities. “All these dark elements describe the margins not the mainstream of Civil War experience,” says Gary Gallagher, a historian at the University of Virginia who has authored and edited over 30 books on the war. While he welcomes the fresh research, he worries that readers may come away with a distorted perception of the overall conflict. The vast majority of soldiers, he adds, weren’t traumatized and went on to have productive postwar lives.
Compatible with your claim of "fewer problems per soldier in the Civil War", but I would go further and say that without data we don't even know if suicide rates were higher for Civil War veterans vs peers who didn't serve.
> Cochran's wrong about parenting not causing problems, of course.
No, not "of course". Twin/adoption studies find little evidence of "shared environment" having much effect. There are some extreme cases of practically feral children who don't acquire language, but outside of that even surviving the Holocaust as a child doesn't seem to have much effect.
Active duty data looks really weird, honestly:
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2757484
I was not expecting to see significantly higher suicide rates in Active duty personnel in the 1800s, given the high probability of undercounting due to shame/disapproval.
(This, again, isn't specific to the Civil War, and the suicide rates go UP closer to 1900, which again would counter any plausible idea I'd come up with).
I got nothin'.
Also, weren't they supposed to believe that suicide is a mortal sin and a good Christian just carries his cross for life? Or am I overestimating their religiosity? Like, maybe, although formal atheism was not a thing, just quietly not caring was a thing?
On the shared environment thing, it's probably important to distinguish between 'cause' in the sense that you'd see a statistical increase, and 'cause' in the sense of being a critical component of the explanatory pathway.
Imagine you flip a coin: heads the kids copy exactly what their parents do, tails they "rebel" and do the exact opposite. In either case, it feels natural to say the parents are the reason behind the child's behavior. But the coin flip is also causal, and it wipes out any statistical correlation
So even if statistically-speaking shared environment doesn't cause shit, I'm hesitant to throw causal narrativization out the window
How would you distinguish that from having zero causal effect?
"Cause in the sense of seeing a statistical increase" is not cause, it's correlation! This is a normal distinction.
Famously in statistics, correlation does not imply causation.
The difference in PTSD rates between 19th century warfare and World War One may have less to do with the direct effects of the artillery itself, and more to do with the change in how soldiers experienced stress.
If you are campaigning with the Army of the Potomac, there are specific, clearly defined days when you are "in danger" versus "not in danger." Most of the threats to your survival are things you can theoretically see before they kill you, or at least identify as "that place over there where all the gunfire's coming from." While you or other members of your group are in danger, the danger normally happens in a way that you can retaliate against (shoot back). Your position and situation often move around, so that you are not constantly stuck sitting near the bodies of your former comrades. You spend much of your time in a structured environment (the army on the march and in camp) that acts as a "normal" where danger is distant or absent.
In World War One, infantry deployed to the front are continuously in danger from shellfire, which may arrive at any time on nearly zero notice. You cannot see it before it is already falling on or near your position, and at best you get less than a minute's notice to take cover. Bombardments may go on for days, drowning out any hope of living normal life even by the rugged standards of "we are camping out in a trench in a field." Your position is often static, dead bodies often stay near you for extended periods. Periods of continuous danger last longer.
I could easily see a significant rise in PTSD from just the raw differences between being a WWI infantryman and being an infantryman in the American Civil War.
I think prior to World War One there would have been other conflicts where you wouldn't know on which day you'd be in danger. Particularly ones involving fighting irregulars.
Yes, but pre-WWI threats involving irregulars were usually the kind where the irregulars jump out of the bushes and start shooting, or just as often jump out of the bushes and charge you with swords. The threat was personalized and often coming from close enough range that merely normal human vigilance could provide some measure of protection. Also, not all types of terrain make it realistic that you could be ambushed at truly any time.
That, too, is very different from being in a situation where it is predictable that (from the perspective of your inner caveman-brain) the sky is going to mysteriously rain loud whistling explosions that send shrapnel raining across the landscape to kill everyone exposed in the open, and that this is going to go on for days, forcing you to cower in a trench with six inches of water in the bottom, where somebody's hand is sticking out of the wall because this stretch of trench collapsed a month ago under bombardment and buried them alive.
I'm not saying that previous forms of warfare do not produce fear, or do not produce traumatizing PTSD, in any of the combatants who fought in them. But it is easy to see that World War One was very different from nearly any previous war (indeed, plenty of people fighting in it commented on exactly that!) And that many of these differences took forms that it would be at least superficially plausible to expect as having a psychological impact.
The Civil War itself involved some sieges that were precursors to the trench warfare of WW1. Artillery was not as advanced, however.
I thought even at its worst, around 50% of soldiers suffer PTSD. It's a lot, yes, but it also means that 50% *don't* suffer PTSD. I think I remember the statistics imply that it's not all soldiers, but ones under shockwaves specifically, so I find that statistic more surprising.
It means it's not death sentence and we may try to exploit it to find ways to reduce PTSD without decreasing deployment. For instance it'd be very good if we can identify soldiers who're three explosions short of developing PTSD, so we can regulate a more effective rotation mechanism. But still, if it's truly random we're truly out of luck.
I have no idea, honestly. I just pulled some numbers on suicide. PTSD is... a little different, sometimes.
If a car backfires, and someone drops to the ground and reaches for a gun that isn't there, is that PTSD? Or is that well-warranted caution/instincts bred from being endangered?
How about someone who hears a cannon fired at them, and flips the table (sending a laptop crashing to the ground), while they hit the deck? (Erm. This incident occurred inside a well-built (pre 1950) building. The odds of the cannonball entering the building were slight).
Both of these are "problematic assessments of risk" but I'm not sure they meet the technical definition of PTSD.
Yes, I do agree,it would be nice if we could figure out who's one bullet away from PTSD, but the Army likes robot dogs better -- and really, who can blame them?
That's a tradeoff too! Making it easy to cheat at something is often a side-effect of making it easy to do legitimately, and at some point you have to accept some background level of cheating to not burden the legitimate users too much. See e.g. https://www.bitsaboutmoney.com/archive/optimal-amount-of-fraud/ for another example.
True. And this is something I've thought about too.
If your system can distinguish 100 categories but the frauds cluster into 8 behavioral modes, you're wasting investigative energy on distinctions that don't matter. If your system can only handle 3 categories but the population exhibits 30 distinct fraud signatures, you're blind.
So this could be operationalized: An agency could measure the *mutual information* (specifically, KL divergence) between its risk scores and actual fraud outcomes, then iteratively adjust its categorization scheme to maximize that mutual information per unit of investigative cost, maybe with some buffer.
You could also measure entropy and volatility of signals themselves, not just summary statistics, which are easier to game.
There is actually a lot of potential for AI and AI generated algos to help reduce fraud and expand available services on the same operating budget.
I think this turns a useful rule of thumb into a rigid principle that refuses to deal with hard situations. It's all nice and well to say something like "we'll compensate people for lost limbs, which is impossible to fake, and not for epilepsy, which is easy to fake" until you get epilepsy and nobody will help you.
I don't think the designers of these systems are too dumb to notice that cheating is bad, I think they've decided that they'll do the best they can, but that potentially allowing a few cheaters is less bad that ignoring every malady that doesn't have an unmistakable physical component.
I didn't mean to suggest that it should be too rigid. In hell, due process is meticulously observed. And I definitely see and appreciate that the people responsible on the ground are usually doing the best they can.
Ultimately, I guess my point is that in designing these systems, we need to budget for the cost of maintaining them, including the cost of making the measurements necessary to suss out the frauds. Too often, we design systems with the best intentions, and then watch them ruined by the handful of bad actors who abuse them. This is what I meant by easy to cheat.
The actual costs of maintaining a system by preventing an unreasonable rate of exploitation by frauds are often much lower than would be imagined by someone who expects massive fraud to dominate the system and its outcomes.
The current kerfluffle over voter identification presents a good example. Clearly, the US election system must be protected against having millions of noncitizens voting. It turns out the actual cost of preventing noncitizens from voting in US elections at any rate plausibly capable of having an impact on US political outcomes is "very close to zero," though. Because the rate at which this actually occurs is already very close to zero; there are not, in fact, millions of noncitizens voting in the elections.
The trouble is that the need to provide protection against frauds and cheats must be measured in terms of outcomes, as opposed to emotional security.
The question of outcomes is "how much does it cost to actually succeed in doing this." How much does it cost to ensure that money genuinely wasted on actual frauds makes up only 1%, or 0.1%, or 0.01% of the overall budget of the operation?
The terms of emotional security are more nebulous. "How much does it cost to make me stop being afraid?" "How much does it cost to make the man on TV with the very convincing haircut stop telling me that this is a problem?"
Whether I feel unafraid about a potential 'danger' may have little connection to whether the danger exists and to how much more resources society needs to invest in it. My personal fear of ghosts does not mean that society needs to spend billions on systems to protect us all from evil spirits. Whether the man with the nice haircut stops claiming that there is a problem is probably has little to do with whether the problem has been solved, or even existed in the first place. It has much more to do with his own incentives and with those of whoever owns the news channel.
Many people overestimate how easy it is to "cheat" a system in blatant fashion (e.g. "fake having PTSD") because they can easily imagine a Just-So story in which a cheater prospers, and it instinctively makes them angry, and so they assume it must be real and a threat to their way of life. In practice, this leads to situations where the public is easily manipulated into approving actions that are bad for normal people, but good for skilled manipulators that have access to billion-dollar bank accounts from which to build propaganda.
There's another sense in which everything has tradeoffs: the cost of preventing a certain bad outcome can easily be worse than the original outcome. We could solve the problem of bad pizza by simply shooting every chef who baked a bad pie; our immune system could prevent cancer by immediately terminating every cell that looked at them cross-eyed.
You mean, like, autoimmune diseases?
> the cost of preventing a certain bad outcome can easily be worse than the original outcome
Non-hypothetical examples: War on Drugs and earlier, Prohibition. Most harm associated with drugs is a direct or indirect result of their illegality, not of their physiological effects. You don't see any liquor sellers shooting each other -- although during Prohibition they did.
I don't think 18th century China thought the same about opium
19th century, not 18th. And you are correct, Qing (and later, Communists) thought opium was a scourge. But that was mostly because UK *forced* opium onto China via a shockingly humiliating military defeat. Qing government did not have any issue with opium earlier, when it was all homegrown.
On the other hand, Prohibition had a substantial beneficial effect on domestic violence, possibly enough to offset all the gangster violence.
I am not saying Prohibition or War on Drugs are (or were) an "unmitigated negative". Their purpose, at least ostensibly, was to prevent certain bad outcomes, and they at least somewhat succeeded. It is just that the side effects were much worse.
It was not just gang violence -- poisonings by methanol went way up. There were many good reasons Prohibition was cancelled after 13 years.
Social model of disability: it's a tradeoff!
Disability: it's a failure!
(low-effort ITT failure, to be sure, but it's what immediately sprang to mind)
More importantly, this is cornmeal crust erasure. Vigolo's would like a word.
I think there's probably also a comment to be made comparing the "tradeoffs" we see all the time in evolution with "failures" leading to extinction, but that's getting out over my skis and I don't feel confident enough not to assert something stupid. Iterated games complicate the dynamics here.
All multifactorial problems should naturally be combinations of tradeoffs and failures.
hahahaha.
You're no longer walking the savannah. So there's one extremely obvious idea you're missing:
Formerly adaptive mutations that no longer have the triggers that caused the adaptation to be functional. You no longer have the diet we were adapted for (and, sometimes, the gut bacteria). A very adaptive mutation might very well cause problems -- if they only cause, say, non-childhood cancer, you don't really starve the gene out of the population either.
Pale skin and freckles are very functional adaptations to northern climates and covering ones' skin for most of the year (in particular, the Vitamin D generation at the end of winter, a traditional time for starvation and illness), but if you're living at the equator, you're no longer getting that boost, and, if you're a redhead, still have higher cancer incidence.
https://www.massgeneral.org/cancer-center/clinician-resources/advances/redheads-carry-melanoma-risk-independent-of-sun-exposure
Yes, these were previously tradeoffs but are now failures.
I’m confused, isn’t this a tautology? Or at least for nearly so? Things that are bad break into components: parts that are simply bad, or parts that are bad with a trade-off of good. Either you have some goodness tradeoff or you don’t. Since you’re starting with the thing, that is bad, all your components necessarily include a little bit of that thing that is bad. Is there some depth here that I’m missing?
No, it's absolutely not a tautology. Scott's missing entire swathes of solutions (as I noted above in my post directly to the OP).
You can have a brain that is smarter, say, with enough salt -- but deprive it of salt, and it's suddenly WORSE than others. Yet, you evolved for a salt-rich diet, and seeing it in the perspective of "bad bad" or "trade-off bad" is completely ruining the paradigm. It was a GOOD mutation, and it got selected for, and all it's really waiting for is more salt, an environmental change.
Being better in one type of environment and worse in a different environment is a tradeoff. Probably a very common evolutionary tradeoff, if the environment tends to remain stable long enough for natural selection.
It's no longer the same tradeoff, though. You can't say "it's a bit better for sun exposure and a bit worse on cancer" (like, say, being a redhead). You get "this sucks" or "this is great" and it's all environmental. You can't take a maladaptive "current state" and say whether it's been adaptive or not, in prior states, based on the current one.
In which case, you've got three states:
1) Adaptive.
2) Tradeoff (some parts good, some parts bad)
3) Maladaptive (but not always lethal)
and, here's the key point, you can't figure out which a particular gene (set) is, based on your current data (one "current thing" that is historically weird is "early menarche" at age 12-13ish, not age 18ish, and you better believe that has influence on how problematic "diseases that start in late teens-early twenties" are in terms of reproductive success.)
I think you're smooshing two layers of analysis together here. Some trade-offs are adaptive, some trade-offs or maladaptive. 'Adaptive' and 'Trade-off' are not mutually exclusive categories.
Yeah, I think OP is doing so as well, probably driven by his perspective on psychiatric disorders. Consider that some psychiatric disorders may actually be adaptive, no matter how unpleasant they feel at the time.
I'm a bit confused by this. Much earlier menarche means that it's physically possible for girls (specifically) to reproduce earlier. Later menarche (say, age 18) means that a disease that kicks in when she turns 20 will affect a woman during a higher proportion of her lifetime period of fertility. Earlier menarche means she theoretically has several years in which to have one or more children before the disease hits. That part makes sense...
...But if early menarche is a <i>novel</i> phenomenon, something historically weird, then we must convolute that with the fact that teen pregnancy rates in developed societies are very low. While it may well be for all I know that medieval 14 year olds girls usually hadn't had menarche yet, whereas modern 14 year olds are theoretically able to get pregnant... in <i>neither</i> period does this result in statistically large percentage of women having children at the age of fourteen.
Which means that "theoretically you can now have kids before this disease affects you" isn't actually going to impact reproductive success very much.
Well, the finer point on it is "the better the food supply, the earlier the menarche" (so Israel could and did have menarche years earlier than the Norwegians). And yeah, that does fun things to genetics, doesn't it just?
Teen pregnancy rates are historically lower than they've been since... early 1800s or so? So you might have had a flowering of "this is possible"... dunno? Just spitballing.
https://pubmed.ncbi.nlm.nih.gov/26703478/
Okay, if this article is true, many of my conceptions were false. Just read the article! (and if you find a better one, go with that).
I'm afraid I don't know how to actually read the article; I'd be curious to know the details of the claim that Paleolithic skeletal evidence shows signs of menarche age being from 14 down to the extremely low value of 7.
This post is saying "I wrote of this in the past as very complicated stuff, but if I restate it like this it's just obvious" so no surprise if you find it obvious :)
Oh ok cool. I never saw tha older post so that explains my confusion.
I'm a bit skeptical of how much that cystatin C study generalizes to "higher cancer risk = lower age-related degeneration risk".
One could also make the case that faulty DNA repair -> cancer and faulty DNA repair -> aging. This is definitely true for ovarian aging, see: https://www.nature.com/articles/s41586-024-07931-x
Agree, it doesn't generalize this way.
It's certainly possible to have both higher cancer risk (because of DNA damage) and at the same time higher degeneration burden (because cell senescence is adaptive in high-DNA-damage context). For example, smoking or radiation can plausibly increase both.
The "faulty DNA repair -> cancer and faulty DNA repair -> aging" framing is basically the figure 1 of Wolf's tumor suppression theory of aging paper (distilling Judith Campisi's research direction).
https://ars.els-cdn.com/content/image/1-s2.0-S004763742100155X-gr1_lrg.jpg
So through this lens, the ovarian aging findings would be interpreted as "ovaries have adaptations that initiate defensive degeneration in high-cancer-risk context".
As they write: "Cancer risk-increasing alleles in SAMHD1 were associated with later [menopause age], following a similar pattern demonstrated previously for CHEK2. This finding is consistent with a mechanism of disrupted DNA damage sensing and apoptosis, resulting in slowed depletion of the ovarian reserve".
In other words, when you disrupt tumor-suppressors SAMHD1 and CHEK2, the ovaries senescence is delayed, but the cancer risk is increased. That's the tradeoff. Basically shifting leftward on the Wolf illustration, but not upward.
The title is breaking my brain.
Yes. I still can't parse it.
I just have to think of "component" as a verb
Ah, the beauty of English, where every word has many possible meanings.
Q: What is the difference between the "naturals" and the failures?
A: The "naturals" trade-off, but the failure components.
(Whatever that might mean.)
Is it merely a bad title, or has he traded off comprehensibility for something else, like getting closer to a pangram?
I think there's supposed to be an "are" in there, but it is kind of an existing sentence structure (e.g. the title of https://www.lesswrong.com/posts/kJiPnaQPiy4p9Eqki/what-evidence-filtered-evidence).
Thoroughly enjoyed this - like the best of Scott's ideas, it feels obvious once stated, but you had to get there.
It also feels like a throwback to older times, before so much of ACX got sucked into politics and AI questions.
My favorite story in this regard about Henry Ford, the Model T and king pins. Let me just copy and paste from wiki.
"Henry Ford, it is said, commissioned a survey of the car scrap-yards of America to find out if there were parts of the Model T Ford which never failed. His inspectors came back with reports of almost every kind of failure: axles, brakes, pistons – all were liable to go wrong. But they drew attention to one notable exception, the kingpins of the scrapped cars invariably had years of life left in them. With ruthless logic Ford concluded that the kingpins on the Model T were too good for their job and ordered that in future they should be made to an inferior specification.[3]"
I used to hang with a lot of gearheads but had to look up kingpin.
Slightly related historical trivia: There is exactly one part of the Model T that is interchangeable with a modern Ford: The valves for letting air into the tires.
That's based on, as I recall, a 19th century patent that was essentially perfect as-is and never changed.
The title is a joke where by virtue of a random deletion, it has failed. The question is whether the resulting comments are a positive, determining whether the title failure is a trade off or simple failure
The title is hard to parse. Should there be an "Are" in there? (I realize that it's valid English to make the "are" implicit, but it doesn't seem like you're going for something poetic here.)
Speaking of amputations, you hardly ever see people with missing arms anymore. I would attribute this to several factors, at least in the US: major wars receding into the past, better safeguards on industrial and agricultural equipment (not to mention fewer individuals working in those jobs), and greater ability to reattach severed arms.
Great article, but the title was deeply confusing to me for some reason; I’m still not sure how I’m supposed to parse it.
It reminds me of Robin Hanson's style of titles, which I'm also not fond of even if he's otherwise perhaps my favorite blogger.
The terse headline-style title of this one reminds me of a (probably apocryphal) exchange of telegrams between a newspaper reporter doing a short bio piece and the actor Cary Grant.
Reporter: HOW OLD CARY GRANT?
Cary Grant: OLD CARY GRANT FINE. HOW YOU?
https://quoteinvestigator.com/2010/10/30/how-old-cary-grant/
Probably misattributed but with a true core.
You accidentally a word from the title.
Elipses one of my favourite literary devices.
From what I can tell, the IQ variable in the Splitting Schizophrenia paper isn’t based on a clinical IQ test, but on UK Biobank cognitive data which was brief, unsupervised touchscreen tasks measuring things like reaction time, memory, and simple reasoning. These are then combined into a statistical proxy for general cognitive ability.
While that proxy has some validity at scale, it’s still a relatively noisy and limited measure compared to standardized IQ tests like the WAIS. So the conclusions are interesting, but they rest on an indirect definition of IQ.
Oooh, I really don't know that being unintelligent is all failure with no tradeoff. It's not what I'd choose but there's certainly room for debate on that front.
There's room for debate, but the subject is greatly complicated by the fact that "being intelligent" often manifests itself in ways that are socially constructed, either by the broader society or by the subculture that the intelligent person is a member of.
An intelligent person who has willingly or unwillingly internalized behaviors that cause them to self-sabotage may fail to thrive in situations where a less intelligent person who lacks the self-sabotage does thrive. And they may blame their "greater intelligence" for their failure to thrive. But in reality, the problem may be something else.
Rates of suicidality are higher in smarter people, smarter people tend towards fewer children, smarter people tend to have fewer completely idiotic sexual encounters. All of these are disadvantages
I wouldn't say I have a very solid "theory of ethics" but to the extent I do, it's based in this tradeoff principle. CS Lewis wrote that human desires are like the keys of a piano. They make a horrible, discordant cacophony when played randomly, but beautiful music when played by an expert pianist (obviously he meant God, but I don't).
Aristotle's Golden Mean rightly gets a lot of flak, but I think it's groping towards this as well. Often two goods are in competition. Where this doesn't happen, there's really nothing to talk about (Imagine positing a Virtue of Taking Breaths in the Furtherance of Life. Who are you having this conversation with? Who disagrees?). Where it does, you must balance the goods.
You can also model this economically - systems have inputs and outputs. They compete with each other for inputs and factor this competition into their design. So when the competitor is weakened or removed, they may produce too much output and muck up the system.
And failures are...still this, tbh. Like a small chance of falling off a cliff is the cost for having functional gravity, a necessary component of our physical system.
The problem isn't that some things are good and some things are bad. The problem is that we have competing goods, in every sense of the word "good." We have goods that are bad in excess, maybe because they block other goods. Or we have goods that we can't get enough of because another system is hogging them. Or we have goods that are necessary constraints on their systems, but do unfortunate things in specific cases. Ultimately, you have to figure out what goods you want to purchase (an extremely hard problem), figure out what they cost in other goods, and then figure out if that's a reasonable price.
This sounds a bit tautological. Like, anything CAN be "decomposed" into factors, the tricky thing is deciding which specific factors are useful and practically speaking how to do that decomposition.
All diseases can be decomposed into "ice cream factors" (to the extent that their development is correlated with increased ice cream consumption) and "monotony factors (the opposite). Type 2 diabetes and some gut biome issues have a high ice cream factor, cystic fibrosis doesn't.
All qualities can be decomposed into skub and non-skub factors.
I feel like talking about "the real components" of things in this way is similar to talking about "the real basis" of a vector space. There are many ways to describe something, you choose between them based on what you actually want to do (in which case, one basis might be situationally very nice) and not on which one is "the most ontologically appropriate" (which isn't unique).
How do you tell if it is genuinely a case of trade-off in the sense that certain good things can only be realized with a certain risk (eg sickle cell) as opposed to simply a beneficial mutation and harmful mutation that arose in the same lineage and the beneficial mutation was useful enough that the harmful mutation has yet to be selected out of the population?
I'm not sure what you're asking - I'm not claiming there's a natural way to know through philosophy alone. I think you'd have to do the research and figure it out.
I took you to be suggesting that we already had done said research and had reason to believe both effects were real. Based on what you said it wasn't clear to me that the evidence is clear enough to be suggestive of that. It is certainly a hypothesis worth testing, I just thought you were suggesting we already had reasonable evidence for it.
My sense is that to really resolve this we need to identify the specific genes at issue so we can seperate their effects, eg, is it just that there is a gene which increases psychosis risk and lowers IQ/edu attainment and it is correlated with another that increases IQ/edu attainment or do some psychosis increasing genes do that without the negative effects on IQ/edu attainment.
Governments can be thought of as varying mixtures of tradeoffs and failures. Communism is pure failure, as it fails to even optimize its stated objective of the well being of the proletariat better than literally anywhere else on the spectrum of economic freedom. Housing construction regulations and rent and wage controls are mostly-failure. Tariffs are a mostly-failure inefficently subsidizing domestic manufacturers at the expense of domestic consumers and foreign consumers at the expense of foreign manufacturers. The indoor smoking ban or the requirement to register vehicles is a legit tradeoff between freedom and limiting externalities.
> Communism is pure failure
I am the last to defend communism, but I am not sure that "pure failure" is the right frame. For example, East Germany was able to embark on a massive homebuilding project that delivered significant improvements to housing quality and cost. Singapore is not typically thought of as a "communist country", but the government owns almost all the land and over 75% of Singaporeans live in public housing.
There is some work to understand aging and cancer as a general trade-off that multicellular organisms face, which itself can be framed as the trade-off between coordination and competition within a multiagent system (as arises across natural scales).
See https://doi.org/10.1073/pnas.1618854114 and https://doi.org/10.1098/rspa.2023.0204 (for a start)
This post says something simple and true: tradeoffs and failures are both common types of causes, and for many effects, both may be present.
Yeah this is obvious. In the literature it's called riskmaxxing
Scott: I'm having trouble making grammatical sense of the title of your post, "How Natural Tradeoff And Failure Components?" Shouldn't there be a verb in there? "How Natural *ARE* Tradeoff and Failure Components?"
Sorry, I lapsed into Hanson-ese.
I just read the tradeoffs-vs-failure post and my main takeaway was that, because of the thing about March-conceived babies, you found a scientific basis for astrology-based character traits
I guess you've already touched it at the last part, but still when a model seems too universal, I immediately wanted to ask what kind of scenario can debunk it. Because if there are none, then it proves too much and is useless.
> ... situations we understand well like romance ...
Maybe I didn't get the memo? Or maybe Scott indulging in some wry humor? ;-)
I think people have a better intuitive model of romance than of, for example, the way that t-cells release cytokines. If what we're actually talking about is t-cell cytokine release, I think an analogy to romance would shed light on the situation.
Fascinating stuff. As someone who works in urban education, I can say many failing institutions in urban education wouldn’t be failing to this degree had the subjects of these institutions not made tradeoffs that left them poor.
> to a first approximation, cancer is always bad, you just don’t want to always keep your body in the most cancer-minimizing state possible
I should point out that "the most cancer-minimizing state possible" is death.
I think that whether people default to a heuristic that assumes tradeoff etiologies or default to a heuristic that assumes failure etiologies when it comes to psychiatric and sociological problems tells you a lot about their political ideology. If most psychiatric and sociological problems involve failure etiologies, then this casts the world into a hierarchy where people at the bottom drag everyone else down and have little hope. If most psychiatric and sociological problems involve tradeoff etiologies, then people who struggle or make poor choices are the natural consequence of a population that is well-rounded, resilient to environmental changes, and produces different people who are useful in different ways.
reminds me of the Andy Samberg YOLO song with arbitrarily escalating risk avoidance with comically obvious horrible side effects.
> The reason it keeps happening is that the gene for muscle protein is really big - so if you get a random deleterious mutation, it’s pretty likely to be there!
Well, that goes back to your previous point: there's probably a smaller protein that's less risky, but also doesn't work as well. So the trade-off is amongst the risks again.
Both is a fun and clean answer to the question wether mental illnesses are purely detrimental or have trade off capacities.
I am a John Nash esque schizophrenia case, now holding a pretty good law degree, and I have felt a strong aversion to polygenetic screening before, because I strongly felt that eliminating the schizophrenia gene strain entirely would be a massive mistake.
I am fine with pruning it, cutting selectively only the parts that make me retarded.
Thank you for the highlight! Much appreciated.
The scarier version of the tradeoff/failure model: schizophrenia risk genes increase educational attainment via creativity, and the gene that makes you more likely to think in broken patterns also makes you more likely to build something new. A lot of civilization might be tradeoff genes running at scale.
There is no objective "good". Something may seem to have no tradeoffs; time will tell. If the tradeoff is effective, as contrasted to good, then it will be more likely to persist and become more common.
“How Natural Tradeoff And Failure Components?”
Turns out there’s a term for it.
crash blossom
Noun
crash blossom (plural crash blossoms)
1. (linguistics) A sentence, often a news headline, that is subject to incorrect interpretation due to syntactic and/or lexical ambiguity.
https://en.wiktionary.org/wiki/crash_blossom
Turned up as an answer in today’s xword.