site banner
PaperclipPerfector  1mo ago (text post)   28186 thread views

Culture War Roundup for the week of October 14, 2024

This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.

Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.

We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:

  • Shaming.

  • Attempting to 'build consensus' or enforce ideological conformity.

  • Making sweeping generalizations to vilify a group you dislike.

  • Recruiting for a cause.

  • Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.

In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:

  • Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.

  • Be as precise and charitable as you can. Don't paraphrase unflatteringly.

  • Don't imply that someone said something they did not say, even if you think it follows from what they said.

  • Write like everyone is reading and you want them to be included in the discussion.

On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.

3
Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

Building off the embryo selection discussion below:

What do you think IQ is exactly?

I’ve always thought about a general factor of intelligence as very similar to a general factor of athleticism. In this context, IQ is a measure of the former much like a triathlon time can be a measure of the latter.

In every sport, triathlon time is going to be positively correlated with ability across the whole population. However, the absolute best performers on specific tasks will not be the ones that do the best in triathlon, because each task has room for optimization that has negative tradeoffs for triathlon performance ("no free lunch"). If you single-mindedly select for triathlon performance, you’ll get a generally more athletic population. On the other hand, you’ll funnel away from getting a Bolt, a Phelps, a Messi, a Jordan, a Federer, etc. Contributions to athleticism aren’t necessarily linear. Individually sub-optimal parameters can align just right to produce optimal results.

There are potential unforeseen consequences of restricting available gene-space by widespread adoption of IQ optimization. Traits are notoriously polygenic (each trait is affected by many genes), and virtually every gene is pleiotropic (each gene affects many traits). Our understanding of both intelligence and genetics is rife with unknown unknowns. Would we still get von Neumann, Einstein, etc.? Supposing the technology became widely available and affordable, is that a fence you’d be willing to tear down?

Edit: It seems I didn't communicate my main concern particularly well. There are two issues with a myopic optimization on IQ: one is negative health effects due to pleiotropy of the associated genes. The other, which I am more concerned with here, is the potential for "lost opportunities". This is what I was trying to illustrate with the triathlon analogy. You can get a narrowing of the variations in intelligence types and a potential restriction on the very upper end of ability. We don't know if Newton, Gauss, Einstein, von Neumann, Ramanujan, and Tao all had a similar combination of traits that led to their exceptional abilities, or if they all had different pieces that fit together in unique ways to produce a unique form of genius (what I meant by "not summing linearly"). Analogous to the way that Phelps, Bolt, and Messi have very different body compositions that produce their unique athletic excellence. A population of excellent triathletes would be more athletic, much like a population of people with 115 IQ would be more intelligent, but that kind of optimization may come at the expense of the variation needed to produce those truly exceptional at related but slightly orthogonal tasks.

because each task has room for optimization that has negative tradeoffs for triathlon performance ("no free lunch").

"No Free Lunch" is cope outside of the context of a competitive environment of evolutionary adaptation. This isn't D&D character creation, you don't have a set number of points to spread around. Some people are just strictly better, and others are just strictly worse. While at some point optimizing for one thing might preclude other things, we're a long way from that frontier. The Marathon was contested for a considerable period of time before runners routinely broke the mark that the best Ironman triathletes have set today.

There's no free lunch in genetics in the sense that if there were something that was simply better for the organism in terms of survival and reproductive success, over enough iterations it would have happened already. But, in our case, we aren't really dealing with a competitive evolutionary environment, and a lot of what would have been evolutionary tradeoffs that would have made an adaptation a dead-end until the last hundred years are now trivially unimportant. Tradeoffs like 'burns 2x calories' or 'takes an extra year to mature' might be fatal in the Great Rift Valley and literally meaningless in Berkley.

That said, I share your concern that IQ might be an imperfect measurement. One of the things that frustrates me about IQ debates is that we're rarely limited to talking about actual IQ scores from an actual IQ test, instead dealing with layers of people using proxies like profession or "sounding like" a high IQ person, then correlating that back to IQ, then correlating IQ to that indicator. It's a weird kind of autocorrelation: we know Einstein must have had a massive IQ because he did a bunch of things that indicate intelligence, and because Einstein had a massive iQ we know you need a massive IQ to do things that indicate intelligence.

Since you mentioned athlecisism, IQ is a score that takes several activities - imagine running, lifting weights, foot-rope drill and maybe throwing a stone. Then average it to some score hoping you captured something like athleticism.

For IQ test the aim is to measure this underlying intellectual ability, the famous psychometric g score/factor of "general intelligence". Individual tests like matrices or vocabulary or object assembly have correlation between 0.6-0.8 with g, but neither of them is perfect correlate. In fact it is harder to say what this general IQ is supposed to be, similarly if let's say professional weightlifter is more athletic than a professional marathon runner.

IQ captures the ability to learn and recognize patterns. Nothing more and nothing less. This is, however, pretty fundamental to the construction of tools, complex societies and delayed gratification, which are necessary prerequisites for civilization.

Don't confuse the map for the territory. IQ tests are a construct that is inherently less reliable at the extremes. In order for Einstein to be Einstein, its was necessary but not sufficient he have a high IQ. All of his collogues likely had similar IQs (von Neumann is the only person I've read about where this might not apply, though he did die somewhat tragically young, which is fertile ground for mythmaking).

My guess would be that enough variation is preserved in the short term. There would be a new norm with the same standard deviation. But that's just a guess.

As an analogy, I think of IQ a bit like horsepower in a car. You can measure power a few different ways, they're all correlated but slightly different, and bigger numbers don't always translate to more actual speed on the road. A lamborghini has a lot of horsepower, but so does a digger.

Strictly speaking, IQ predicts educational capacity. It's correlated imperfectly with a bunch of other positive mental attributes, but bigger numbers don't always translate to more intelligence in the real world, and at the extremes the statistical selection effects are strong.

At the high and low ends, IQ is dysfunctional. Above a certain high threshold, more IQ has negative real-world effects, many of the "smartest" people in the world can't manage their own lives, stay employed or be understood by normies. Even our distant patron Scott looks and sounds either insane or stupid in person. Very bright guy, but a total weirdo IRL, and I'm pretty weird myself.

A society with an average IQ of 120 or 130 would have incredible human capital, a society with an average of 170 would collapse in about six minutes. To bring it back around to my analogy, most of us don't actually want a thousand-horsepower supercar to drive. Roads full of Bugattis would be a nightmare. A bit more speed than average is fine, but as you get faster, there's fewer and fewer places you can drive, and fewer and fewer uses until you get to something like a drag racer, which is fast as hell, and totally useless.

At the high and low ends, IQ is dysfunctional. Above a certain high threshold, more IQ has negative real-world effects, many of the "smartest" people in the world can't manage their own lives, stay employed or be understood by normies. Even our distant patron Scott looks and sounds either insane or stupid in person. Very bright guy, but a total weirdo IRL, and I'm pretty weird myself.

I'm a bit odd, but tradesmen tend to be so it's not obviously due to brains. Is this actually true? I'd thought that Chris Langdon and Ted Kaczynski were exceptions that test the rule- sure, very high IQ individuals are more likely to be oddballs that make strange decisions due to tail effects, but Terrence Tao and Bill Gates were more typical examples.

Let me put it this way, everyone sat in a circle in a room, and Scott moved his swivel chair out into the hallway and spun in a circle while he talked, so all you saw of him was his knees swing past the doorway every three seconds. Plus, that scintillating mind in print is entirely incomprehensible in conversation. It was interesting, but if I didn't already know from his writing that he was smart, I'd have thought he was seriously mentally ill, or perhaps autistically retarded.

Which it is my assertion that very high IQ basically is.

I am not sure I buy into this logic that embryo selection for IQ will not lead to some intelligence diversity. For some higher IQ individuals their intelligence might be more related to memory, or others will bemore related with processing speed, and then some others who are more autistic but still not full autists. I recall reading of research that some brains are more accuracy oriented and others more speed oriented.

The current default is a reduction of IQ from generation to generation. I can see real potential problems with genetic engineering but I am not convinced IQ selection in embryos in particular will be a problem.

What is your plan? Because it is easy to support the status quo that includes a lot of destroyed fences, and then paint embryo selection as the scary alternative. But are you sure what you would have us do is the best option?

The current default is a reduction of IQ from generation to generation

Proof? The flynn effect strongly suggests otherwise.

This post provides a collection of evidence for intelligence decline.

https://emilkirkegaard.dk/en/2021/02/recent-evidence-on-dysgenic-trends-february-2021/

Interesting. Though-- it looks like the big takeaway from that post (assuming it's true) is actually rather the opposite of the standard "dumber people breed more" line. Rather, it implies that the cultural value we've placed on phenotypic intelligence has increased to the point where convergence on intelligent behaviors is reducing the fitness of being genetically predisposed towards intelligence-- everyone trying to act smart makes it less efficient to actually be smart.

I say this because the other thing this paper points out as declining is a genetic predisposition towards being thin and tall, despite the fact that our culture obviously favors and provides reproductive success to people who embody those things. That would be in line with non-genetic causes of height and thinness (i.e., enough money and leisure time to be healthy and fashionable) becoming decisive over genetic causes (and therefore tradeoffs, like heart disease chances for height or digestive issues for thinness)

My position is a rather unsatisfying "there are enough unknowns that I have no idea what the best option is", but I was curious to discuss others' thoughts.

If a young Englishman scores highly on IQ tests they will likely be admitted to a three year degree program and complete it successfully. Sitting their final exams, they will be recalling material learned at the beginning of their course, three years earlier: success depends on having a retentive memory. Retentive for years. But the IQ test only took an hour or two; no chance to probe multi-year memory retention. How does that work?

One idea is that IQ tests are probing for a healthy brain and a low mutational load. It is still a little unclear why the genes that help with rapidly solving little puzzles should be the same ones that boost memory.

I could imagine that sustained selection based on IQ test will eventually break the correlation of test performance and long term memory. In a dozen generations, say 2323 or 2384, there may be a crisis in University admissions. Too many students are really sharp mentally, but they forget their course work after 6 months and end up failing.

The other, which I am more concerned with here, is the potential for "lost opportunities".

Would this outweigh the advantage of having more mid-level high-IQ people as engineers, researchers and entrepeneurs, increasing the amount of wealth flowing around in the system as a whole?

If we try, I'm sure we could construct arguments about the downsides of building roads. It uses a lot of cement, damages ecosystems, induces car fatalities, increases urban temperatures... There might be five or ten moderately bad things about building roads, more if we delve into hypotheticals. But there is one really good thing about building roads! You can travel quickly and cheaply on them! And that is more than enough to outweigh the costs and the negatives. The downsides can be addressed with related measures, skilled planning and implementation.

I suspect it is the same with IQ. Smarter children is a good thing. There is always the option of taking the genetics of elite scientists and cloning them exactly, if it's the genetics that make geniuses. If it's some combination of innate capabilities and some esoteric twisting of the brain via childhood experiences or plain luck, then the boon of having many moderately smarter people will be helpful in producing the luck and opportunities needed.

I'm imagining a hypothetical scenario where we can raise the average but reduce the variance. In such a case, it comes down to whether you think it's more important to have smarter mid-level engineers, doctors, etc. vs having more and smarter geniuses. I don't have a strong opinion one way or another, but was curious what others thought.

A reduced standard deviation in IQ already exists- women. Very few worldchanging geniuses, far fewer retards.

I leave what can be gleaned from this experiment to your interpretation.

What do you think IQ is exactly

IMO IQ is the malleability of animal pattern recognition, which allows for human intelligence dissociated from pure animal instinct (sex, aggression, status). The cognition which would otherwise be dedicated to learning what is sexually attractive and dominant can instead be allocated toward abstract, socially-mediated goals. This explains why high IQ is associated with higher rates of virginity and delayed age of first sexual experience (the instinct is less powerful; otherwise, the high intelligence would lead to earlier and easier sexual experiences). It also explains why having a high IQ doesn’t appear to be related to skill in music composition, as an element of animal instinct is essential. And it explains the popular notion that there is something socially “strange” about very smart people.

It also explains why having a high IQ doesn’t appear to be related to skill in music composition, as an element of animal instinct is essential.

Do you think that music is distinct from other forms of art in this regard (painting, literature, etc)?

I think that great artists tend to be above average in IQ, at least.

It also explains why having a high IQ doesn’t appear to be related to skill in music composition

What's your source here? I've certainly read that musical ability is positively correlated with IQ, but it seems that maybe there's something you know about music composition specifically?

What do you think IQ is exactly?

I'd put it as "Generalized ability to efficiently process increasing levels of complexity."

Now, its fair to say that efficiently processing some areas of complexity won't translate automatically to others, I think we can take autistic-savants and similar cases as evidence.

But that's really the sum total of what it seems to 'represent' about a person. If you moved them from Tic-Tac-Toe, to Connect-Four, to Checkers, to Chess, at which point would they genuinely start struggling?

Someone who works mostly with 2-dimensional concepts or in constrained workspaces probably demands lower IQ than someone who works in 3-D (or 4-D!) concepts in very open-ended environments. The former, for example could be a NASCAR driver who just has to be aware of his immediate surroundings and only has to navigate a closed circuit, and the latter would be an airline pilot or, perhaps, the technician who fixes the airplane, where there are a lot more variables at play, to say the least.

Reality can be 'infinitely' complex in theory, but someone who is comfortable with higher levels of complexity and can deduce certain patterns or cause-effect relationships is, almost certainly, going to be better at navigating the world. I read some research a while back, which I haven't been able to find again, suggesting that there's a strong negative correlation between reported IQ and the number of auto accidents someone experiences in their life.

Makes intuitive sense to me. The ability to think ahead and grasp possible consequences of an action "if I do X, then Y could possibly happen, and I might be injured or killed." and to notice when others are behaving in a way that might likewise cause an issue will help avoid negative outcomes by simply avoiding situations that could lead to such outcomes.

Now, high IQ can be hobbled by intense OCD, or high anxiety, or a lack of executive function, and I think that is mostly what will explain the divergence between IQ test results and real world success and status. Being socially inept can also be a major impediment. The slight 'paradox' is that an IQ test is a very constrained environment with minimal distractions and all the problems are 'legible' so even somebody with a crippling mental illness can probably perform well if they have the mental horsepower.

But I do think that, especially when measured across broad populations, IQ differences are the main reason some places are able to create and maintain complex civilizations with bridges that stay up, computers, and airplanes and others just revert to the simplest techs they can operate despite tons of outside assistance pouring in.

Our understanding of both intelligence and genetics is rife with unknown unknowns. Would we still get Von Neumann, Einstein, etc.? Supposing the technology became widely available and affordable, is that a fence you’d be willing to tear down?

I think so. The space of all possible designs for human minds is large, and contains Einstein and Jeffrey Dahmer and Hitler and Mister Rogers, so we would certainly not want to move more into the space where there are more sociopaths than 'normals,' but the space is still constrained and thus its highly unlikely we accidentally produce a few MEGAHITLERS by accident.

The risk of creating a bunch of Jeffrey Dahmers (IQ of 145, allegedly) instead of more Einsteins and Von Neumanns is pretty minimal, and probably wouldn't kill us off, and on net I think we see improvement in everybody's standard of living. And probably faster than we would have 'normally.'

If I was presented with a button that, when pushed, instantly raised every living person's IQ by 5 points (as measured on tests), but changed nothing else, I would happily push it, I think it would substantially improve things in the near term and would have few negative side effects even across the long term.

What this tech sort of promises to do is achieve that same outcome, but across a longer timescale.

Jeffery Dahmer may have been very smart, but the vast majority of serial killers are of average or below-average intelligence.

I'll add the caveat that the selection of serial killers who have been caught might not reflect the entire population of serial killers. The smarter ones might have avoided detection entirely.

But I picked Dahmer because his whole thing was he was particularly intelligent and completely sociopathic and depraved... so we do NOT want more of them running around if we start selecting for more intelligence.

It would take quite a few additional Dahmers to make up for the loss of the predominantly low-IQ people committing 15,000 homicides per year.

Or just a couple of them, but they have access to bio-engineered diseases.

But enough about Fauci.

*Peter Daszak.

I think of intelligence like I think of processing power in a computer. Now below a certain level, if you don’t have enough, it’s going to be nearly impossible to do anything useful. I think there are several types:, linguistics, mathematics, art, social. These can’t be used interchangeably— meaning I can’t use artistic intelligence to understand math or language, nor can I use mathematical intelligence to learn to write poetry. To my mind these sit atop a more general CPU that is needed for any type of thinking. And I further think that we’re dealing with multiple genes in multiple places which to my mind would complicate any sort of simple correlation to ethnicity. Until we know which genes exist in which population it’s impossible to tell for sure.

I think of intelligence like I think of processing power in a computer. Now below a certain level, if you don’t have enough, it’s going to be nearly impossible to do anything useful.

The reason I can't quite use this analogy is that even if you have a slow computer, as long as it is Turing-Complete, it CAN complete any given task you put before it, even if it takes literal centuries.

So being faster or slower to complete tasks is not quite the same as being able to handle more complex tasks. I sincerely believe there are problems that 150+ IQs can handle that are utterly beyond a 100 IQer, even if you gave the 100 specific, detailed instructions on how to complete it and gave them years to work on it without interference. MAYBE if you stuck a team of cooperative 100s who are at least capable of delegating tasks and getting along.

So there's other bottlenecks. "Working Memory" is probably the big one. I think extremely high IQ people are also defined by being able to fit a LOT more information in their working memory and thus can can bring all those mental resources to bear at once, rather than having to painstakingly write everything out and do each individual mental calculation one at a time.

So perhaps add in RAM to the equation. If you can't fit the majority of the problem in your head, at least big enough chunks of it to make progress, then you'll find yourself unable to ever solve it.

Side note, this is often how I feel most constrained when faced with complex problems. I can't actually 'visualize' the problem in my head because trying to load all the details in ends up pushing some parts out, and I can compensate by writing out bits of info, but this always slows me down substantially.

even if you have a slow computer, as long as it is Turing-Complete

Here's the thing, though. No real world computer is Turing-complete. They all have finite storage and thus fail the infinite tape requirement. For an obvious example, try running eg. Stable Diffusion on an early 90s PC - you simply can't because they don't have enough storage for the model and results, even if you allowed infinite time.

Reminds me of the assertion that a 2004-2006 research supercomputer could probably have been capable of training GPT-3.

I would be careful associating working memory with the brain's ability to actively model complex problems. The latter is a conscious process, while the former is unconscious. An 80 IQ person can, with pen and paper, rotate any shape or model any system given enough time, or calculate out a 6-move chess sequence that Magnus Carlsen could perform in seconds mentally, but he could never have the spontaneous causal associations in his mind that naturally occur to more intelligent people. The lack of this faculty, and exclusively this, is what precludes low IQ people from complex things. This is why the computer analogy is weak. And why low IQ civilizations just can't get it together. If it were only a matter of processing power, nothing would stop them from busting out the compasses and graph paper. But intelligence is really a phenomenon of the subconscious, of the brain noticing a pattern and showing this to the conscious mind. For that reason it can never be taught or compensated for.

While I agree with you, I think "busting out the compasses and graph paper" is what science is. We've reached the limits of what we can do in our minds, so now we do mathematics on paper. This allows us to calculate things that we cannot wrap our heads around (try visualizing infinite-dimensional spaces for instance)

One thing I have noticed that less intelligent people do is solving the same problem over and over again. Even culture wars are like this. "X people are discriminated against, and it's totally not their fault, so we need to give special rights to X group to prevent this, as it's only fair to escalate their power/position in society". Even as a teenager I generalized this problem to every related problem which could exist, but somehow society still sees a sense of novelty in "We are X, we are victims, give us power or you're bad and support bad things"

Edit: My point is that, even with pens and stacks of paper, stupid people cannot generalize or reach levels of abstractions which gives them the advantages of space. Space is really powerful though, even more than Time (which is probably why PSPACE and EXPSPACE are bigger than PTIME and EXPTIME respectively. Not that I actually know complexity theory)

Not sure if I'm agreeing or disagreeing, but consider that writing something out is not the equivalent of having it in your working memory. Although human language is very rich, if we consider writing out a problem to be the equivalent of forcing some million-parameter vector in latent space into a sentence of unicode text, then there's likely to be a huge loss of information/nuance that we can't perceive consciously. It may be that the ability to hold slightly large/more concepts in your mind is responsible for the spontaneous causal associations you describe.

Working memory is a passive process, it's not what we use to consciously model things. Not sure what we'd call the modeling area of the brain, I've heard sensorium used.

It may be that the ability to hold slightly large/more concepts in your mind is responsible for the spontaneous causal associations you describe.

It's a bit of a mystery really. All we know for sure is, working memory/modeling ability/intelligence are strongly correlated. When you and I say modeling ability we're probably thinking about shape rotation or figures and so on, but I believe each form of intelligence has its own type of modeling ability, which is accompanied by a strong working memory (at least in that field). So I suppose there's no knowing which is the 'essential' component, the two always occur simultaneously.

I'm sure you've seen the Gwern essay on embryo selection where these lines of argument are touched on (https://gwern.net/embryo-selection). The whole thing is ofc great, and on this point the TLDR is: yes absolutely selecting myopically on a single trait can go wrong (especially over long timescales and/or small populations) but given the size of our population and existing genetic variation it's not a pressing worry at all. Further it seems that presently IQ correlates are other "good things" like overall health, so the tradeoff does not even arise.

Selecting only on one trait means that almost all of the available genotype information is being ignored; at best, this is a lost opportunity, and at worst, in some cases it is harmful—in the long run (dozens of generations), selection only on one trait, particularly in a very small breeding population like often used in agriculture (albeit irrelevant to humans), will have “unintended consequences” like greater disease rates, shorter lifespans, etc (see Falconer 1960’s Introduction to Quantitative Genetics, Ch. 19 “Correlated Characters”, & Lynch & Walsh 1998’sCh. 21 “Correlations Between Characters” on genetic correlations). When breeding is done out of ignorance or with regard only to a few traits or on tiny founding populations, one may wind up with problematic breeds like some purebred dog breeds which have serious health issues due to inbreeding, small founding populations, no selection against negative mutations popping up, and variants which increase the selected trait at the expense of another trait.^28 (This is not an immediate concern for humans as we have an enormous population, only weak selection methods, low levels of historical selection, and high heritabilities & much standing variance, but it is a concern for very long-term programs or hypothetical future selection methods like iterated embryo selection.)

The footnote (no. 28) continues the argument:

Although unintended side effects due to bad genetic correlations is frequently raised as an objection to selecting on intelligence, it is not an issue, as the genetic correlations of other traits are so uniformly in desirable directions (with the practically-unimportant exception of myopia, and the dubious exception of autism spectrum disorder symptom checklist scores). In any case, we can note by their behavior that people do not genuinely believe this objection, as they are, among countless other things, not upset by having intelligent children, do not regard successes like childhood vaccination or iodization or lead remediation or improved nutrition or the Flynn Effect as humanitarian disasters, interpret reports of (hollow) gains on intelligence due to formal schooling as reasons to increase funding for formal schooling, are worried by brain-damaging diseases like Zika (thus passing the double reversal test) and would be horrified by a proposal to starve or drop babies on their heads to spare them the supposed terrible unintended side-effects of being above-average.

The footnote (no. 28) continues the argument:

I support embryo selection but I don't buy the argumentation Gwern puts forward in this footnote. People may be fine with rolling the dice naturally, but still take issue at monumental, population wide changes due to engineering.

"Yes, I'd love my kid to be selected from a better distribution, but I fear long-term effects if this policy is implemented at mass-scale because I am not a geneticist and it seems risky" is a perfectly consistent position.

The point of the argument in the footnote is to show that, once the "genetic engineering" boo lights are removed, everyone's revealed preferences favor the same outcome as the world in which we select embryos for higher intelligence, harmful comorbidities included (real or imagined). If people somehow think that rolling the dice with nature is less likely than embryo selection to unintentionally couple higher intelligence with undesirable traits, to the extent that it's preferable to accept "natural" outcomes orders of magnitude worse than their preferred outcome to mitigate the risks unique to embryo selection, they either have a dismally wrong understanding of embryo selection (which, reminder, is just rolling the dice a bunch and picking the best-looking result) or they're not reasoning consistently.

At its core, objecting to the reasoning of the geneticist with a shrug and an "I dunno man, sounds risky" isn't actually an argument about the risks (surely the geneticist has deeply considered them, and our objector is already on-record as lacking the qualifications to do so!) - it's an expression of distrust.

I suspect people's primary objection, regardless of whether they clearly understand and express it as such, has nothing to do with the long-term risk of embryo selection at a genetic level, and is instead based on the same obvious ethical and political concern for any eugenics proposal - that it will be applied unfairly by some groups to gain power over others. This includes geneticists and their employers miscalculating, misrepresenting, or lying about risks, evading liability for accidental harms or unsatisfactory outcomes, and charging enormous sums of money for extremely modest benefits; and rich parents granting their children an effortless comparative advantage over the majority of children whose parents don't (or can't) pony up to rig the game for themselves.

So far, the only one tradeoff we know for sure for IQ is 'doesn't want children'. Suppose we made genetic score for triathlon. What makes you think sports celebrities you listed wouldn't be above average in these scores? What if we don't get Einstein but get 10x more geniuses that are better than Einstein?

An interesting question is why IQ suppresses fertility rates. We know, incidentally, that high IQ's are strongly correlated with poor sexual/romantic success. Maybe it's not a coincidence that ultra-high IQ groups like Ashkenazim and Tambrams use arranged marriages(doesn't Japan widely use don't-call-it-arranged marriages?). It also seems like, in the US, groups with higher out-of-wedlock birth rates are the ones which see the most dysgenic IQ selection.

There's only one recorded case of a virgin birth in human history. Kinda hard to have kids without a partner. I suspect that within marriage fertility rates are not particularly correlated with IQ.

Poor romantic success can't be the only explanation https://www.emilkirkegaard.com/p/not-just-dysgenic-outcomes-but-dysgenic [intentions too]

My understanding is that, in the developed world, IQ (or at least, earnings and education as IQ proxies) are positively correlated with male fertility but negatively correlated for female fertility.

Basically, smart men earn lots of money and so can attract a wife more easily, while being able to afford to house more children.

Smart women spend their most fertile years in education and 'greedy' careers, leaving little time for babymaking.

High fertility among low IQ people was previously driven by teenage pregnancy, but that has mostly disappeared in the past couple of decades

My understanding is that the correlation of male IQ with fertility is limited to men with conservative attitudes about gender- liberal men don't get a fertility boost from higher IQs(presumably this is because men with more conservative attitudes about gender are more willing to 'marry down' on average- the surgeon will go for a cute nurse where infectious disease docs are mostly interested in ObGyns and pediatricians with the usual fertility penalty from decade long higher ed tracks), but that conversely sudden spikes in male wealth(eg male lottery winners) lead to fertility increases regardless of IQ so I think you've probably gotten the mechanism right.

Out of marriage fertility is declining, including among low-IQ people, but it's still much higher among the poor and dumb than among wealthy, educated, smart people, and differences in developed world TFR tend to be driven by out of wedlock childbearing because married couples tend have similar TFRs in western societies.

Isn't intelligence correlated with mental illness?

That seems like a big tradeoff.

The opposite. Higher IQ = better mental health.

https://pmc.ncbi.nlm.nih.gov/articles/PMC5014225/

I mean, it makes sense right. Higher IQ leads to better life outcomes which leads to better mental health.

Perhaps you could "correct" for life outcomes and get a neutral result, but that wouldn't make a lot of sense as high IQ causes good life outcomes.

Yep, an entire class of low intelligence alcoholics & criminals aren't considered mentally ill because their affliction is socially acceptable.

Instead of manic genius, you get alcoholic loser.

Hmm. It seems this is an area of at least some some contention, with some research (e.g. a Mensa survey, which obviously is based on a self-selected sample) suggesting a higher correlation, but most research suggesting IQ is protective.

There's also some evidence suggesting a genetic correlation between autism and high intelligence.

My guess is that (to the degree that IQ is genetic) is that it's probably possible to "overselect" for it to the detriment of other good things (although IIRC we also know that e.g. autism is probably at a minimum correlated with other factors, such as older parents and maternal fever during pregnancy).

Higher IQ leads to better life outcomes which leads to better mental health.

In my personal experience I have observed that the connection between high IQ, good life outcomes, and mental health is not strictly linear. But that's anecdotal and a very small sample size.

I think Mensa selects for people whose IQ test score itself is their highest "achievement", i.e. the lowest performers at any given tier of IQ. So it's very possible that Mensa members could have on average unusually poor mental health.

They looked at phenotypes. Imagine looking at two nearly identical twins, both with genetic predisposition to schizophrenia but one was lucky to avoid it, and the other got it along with IQ drop that goes with it. This would inflate positive correlation between IQ and mental health. There should be looked something like Scott wrote about: https://www.astralcodexten.com/p/non-cognitive-skills-for-educational

This is well put. What we don't understand as well, when it comes to different forms of intelligence, is what the trade offs are. Whereas it's very easy to see that the weight reduction that helps you cycle faster can hinder your sumo wrestling prowess.

What do you think the trade-offs are? I think you mentioned below how spending time with Cambridge math students led you to believe that high-IQ people are defective in other ways.

I have seen statistics showing that high IQ is highly correlated with many other factors of success, such as marriage stability, high incomes, mental health, physical health, and lifespan. People who have high IQs do not go senile as early.

Furthermore, it seems to me people with above-average IQs have more friends than those with below-average IQs.

But, clearly the stereotype of the Poindexter math student exists for a reason. So it may be that there is a threshold beyond which negative characteristics emerge. Or it may be that it's not so much high IQ, as it is an autistic level of fascination with math, that leads to the typical phenotype of the Cambridge math student.

In any case, there seems like a clear and obvious benefit to going from an IQ of 100 to one of 115.

The experience of the IQ 150 math student is so strange that it doesn't really generalize to the greater population. Although, even here, society will massively benefit from higher IQs even if the individual doesn't.

Neurotypical people with high IQs learn how and when to pretend to be stupider than they are very early - I can see my 4-year-old son (who has a diagnosis, but is noticeably less autistic than his brother or either parent) doing it already. So if you did meet a 160-IQ neurotypical, they would come across as being as smart as necessary under the circumstances, not as smart as they actually are.

You would identify that person because they have enjoyed improbable success in multiple different g-loaded activities, not because they are scarily bright in person - scaring the normies is, after all, stupid.

In any case, there seems like a clear and obvious benefit to going from an IQ of 100 to one of 115.

It's currently beneficial to be 115 iq rather than 100 iq but it could certainly be possible that such a big jump in iq in a single generation risks various other deleterious effects when selecting so strongly for only one thing.

And what if we expand this to include people who have a median familial iq of 85 or even lower. Should the target still be 115? How big a rate of increase is really healthy? How hard should we really select?

So this is impossible without gene editing. The reality of this technology is much different than perception. The average woman with IVF has 7 embryos to choose from. Many of those will be unviable or not ideal for other reasons. Maybe, maybe, you can select between 2 or 3 on average.

You're not getting IQ of 115 from a median family IQ of 85, not that those people do IVF anyway.

I didn't say this was currently possible, I responded to your example of going from 100->115.

Let's say the industry booms and technology advances so we can have a 100 embryos to choose from in 10-20 years, what then?

I think the embryo case below probably is going to be about the tail end of the curve because of the Silicon Valley demographics of people who'd use an embryo selection service for IQ. They are people with high IQ, who've been successful because of their IQ, and are in a culture that is a bit obsessed with IQ, so they especially care about that of their kids and may be trying to terminate the IQ130 embryo in favour of his 140 brother!

I find it highly likely that at the lower, mainstream part of the curve, IQ is indeed more simply correlated with many good things – a less alloyed good if you will.

But as for trade-offs to do with types of intelligence in general, it just seems pretty intuitive to me that if you take a bunch of archetypal impressive people, successful in different ways, you will find many whose brilliance could have been compromised by being too good at logic and not distracted enough from logic puzzles by other parts of life.

Sprint speed or vertical leap is a much better measure of general athleticism. In an endurance sports, the trained athlete always beats the untrained one. But you would never beat Usain Bolt in sprinting even if you trained your entire life and he just did normal childhood activities.

But obviously, there's a lot of training involved in IQ as well. IQ tests rely on the fact that people aren't training for them, and quickly become tainted when people do. I've seen data showing that, among IQ test subskills, the one which is most correlated with g is actually vocabulary. This is, of course, easily trainable. But training speed varies by intelligence. And, of course, smart people will develop large vocabularies in their day-to-day life. Other skills, such as digit span, may feel less trainable, but there are tricks you can learn in only a few hours to improve your ability considerably.

In any case, to answer your question, IQ is g in so much as we are able to measure it. And what is g? It is the general factor of intelligence determined by analyzing correlations between different skills. There are clusters of skills which are highly correlated and relate to things that we might consider mental abilities. But there are other skills (like running fast) which are uncorrelated and do not pertain to general intelligence. So, mathematically, a single factor g is determined by its correlation to other skills which we know to be g-loaded. A person with high g will likely have a good digit span, but we won't be able to assume anything about their running speed. (I'm a little out of my depth here on the mathematical rigor, so perhaps someone can explain this better.)

In any case, I think IQ tests measure g well enough up to at least 2 standard deviations above the norm (assuming no intensive preparation).

How would you make an IQ test better to measure g more appropriately?

My point was less about the specifics of measuring intelligence or athleticism and more about asking if given the degree of uncertainty, is it really a good idea to run headfirst into embryo IQ selection. There are almost certainly aspects of intelligence not captured by IQ tests that help with mathematics, physics, music, writing, etc. By optimizing so narrowly for IQ we don't know if we could be excluding the regions of gene space that might generate a brain that performs best at those tasks, much like focusing only on triathlon (or vertical leap or sprint) performance would exclude the musculoskeletal parameters that make Messi or Phelps so perfect for their chosen tasks.

Sure, there are aspects, but 1. what makes you think they would be negatively correlated with IQ 2. we can't measure them for now or it is too time consuming

Right now, most genetic studies on intelligence do not even use IQ but use EA (educational attainment) as a proxy for IQ. Obscurantists laugh that genes explain small fraction of EA, but what do you want if these studies do not use full genome sequences and mediocre proxy of IQ?

I don't think they would be negatively correlated, I just think there are enough unknowns that the possibility is far from negligible. My priors are that the "no free lunch" theorem applies to intelligence, so if I had to guess I would expect some degree of tradeoff, particularly at the upper ends of performance (whether in intelligence or athleticism). Hence why we never see people elite in both running and swimming events despite both being strongly, positively correlated with "general athleticism".

There is no reason to believe the "no free lunch" theorem holds and plenty of reasons to believe it doesn't. There are lots of genes with purely deleterious effects (those which are fatal to embroys, for instance. Or cystic fibrosis).

Even if the tails come apart -- e.g. you cannot simultaneously MAXIMIZE running and swimming ability -- does not mean that there's no genetic free lunch. Starting from a genotype that wasn't near human capacity in either, you could increase both.

If I were to refine my thoughts, I would say I believe not in an absolute, universal no free lunch, but rather a weaker, "not much free lunch beyond a particular threshold". As you put it, where the tails might tend to come apart.

My concern lies there, at least to the extent that human progress depends on the abilities of the individuals at those tails.

Wouldn't trade-offs being everywhere in evolutionary biology count? I'm not an expert, but as I understand it, if intelligence were a 'free lunch' wouldn't we expect to see far less variation in intelligence in humans? Natural selection having already optimised it?

The most significant drawback for IQ is 'doesn't want to have children' and longer reproduction cycles. Also, a fraction of variation in intelligence is non-linear (heterozygosity) which simple selection cannot maximize and this is why agriculture uses f1 seeds, albeit aiming other traits.

There's lots of traits that haven't reached fixation, but this doesn't mean there's no free lunch. Evolution is slow and the environment changes much more quickly; we are not at equilibrium. The "no free lunch" theorem is equivalent to "all organisms are equally fit", which is clearly false.

Natural selection was still in the process of optimizing our genes when they hit the point where their phenotypes could invent and spread much faster-optimizing memes and then (in an instant, geologically) start wondering whether faster optimization was possible for genes too. Even brain size growth, perhaps the cruelest obvious tradeoff, doesn't show any obvious signs of having leveled off.

Okay. I wouldn't worry too much about these things, since in any embryo selection other factors will outweigh IQ. In IVF, embryos are already screened for chromosomal defects. They are also graded based on a cell development and uniformity.

If (lucky you!) you had multiple embryos with no defects and high grades, then you might do additional screening which is already available from Orchid. This might tell you if your child had a high risk of chronic disease or mental illness. If you still had multiple candidate embryos after all that, you might choose one based on IQ. But consider the uncertainty of this test, and also how siblings already exhibit a high degree of similarity with IQ. The difference in expected IQ might be 1 or 2 points with huge error bars.

So I wouldn't worry that we're going to be breeding people like pugs with huge defects caused by over selecting for obscure characteristics.

The biggest risk to our genetic health is that people are having children later in life and intelligent people have too few children.

But consider the uncertainty of this test, and also how siblings already exhibit a high degree of similarity with IQ. The difference in expected IQ might be 1 or 2 points with huge error bars.

Where's you getting this from? Per Jensen, average difference between sibs is 12-13 IQ points. If your PGS captures r^2=25% variance, then expected difference is 12.5*sqrt(0.25) IQ points. (That is about 10-20% difference in income when adult). That's for two embryos only. Steve Hsu gave more data - how how far you go with more embryos - but i can't remember where it was

Nobody knows how much of the variance is captured. I assume a low amount as this is unproven technology.

I'll grant that I assumed the average difference between sibling was less than 12-13 points, so if that's true I'll admit the difference in IQ can be more than 1 or 2 points.

(Caveat: even though IQ is mostly genetic, it's still partially environmental. So if the mean difference is 12-13 points, the mean genetic difference is less.)

Nevertheless, until we radically change how IVF works we won't be selecting from dozens of embryos. I still rate this intervention as incredibly minor without further advancements.

Sasha Gusev lists IQ PGS as explaining 41% variance at population level and 14% between sibs, since he tries to give as much deflated numbers as possible without lying, true must be higher.

the mean genetic difference is less.

sibs share environment.

ROI of improvement in possible child income by IVF PGD expenses is high

What is this place?

This website is a place for people who want to move past shady thinking and test their ideas in a court of people who don't all share the same biases. Our goal is to optimize for light, not heat; this is a group effort, and all commentators are asked to do their part.

The weekly Culture War threads host the most controversial topics and are the most visible aspect of The Motte. However, many other topics are appropriate here. We encourage people to post anything related to science, politics, or philosophy; if in doubt, post!

Check out The Vault for an archive of old quality posts. You are encouraged to crosspost these elsewhere.


Why are you called The Motte?

A motte is a stone keep on a raised earthwork common in early medieval fortifications. More pertinently, it's an element in a rhetorical move called a "Motte-and-Bailey", originally identified by philosopher Nicholas Shackel. It describes the tendency in discourse for people to move from a controversial but high value claim to a defensible but less exciting one upon any resistance to the former. He likens this to the medieval fortification, where a desirable land (the bailey) is abandoned when in danger for the more easily defended motte. In Shackel's words, "The Motte represents the defensible but undesired propositions to which one retreats when hard pressed."

On The Motte, always attempt to remain inside your defensible territory, even if you are not being pressed.


New post guidelines

If you're posting something that isn't related to the culture war, we encourage you to post a thread for it. A submission statement is highly appreciated, but isn't necessary for text posts or links to largely-text posts such as blogs or news articles; if we're unsure of the value of your post, we might remove it until you add a submission statement. A submission statement is required for non-text sources (videos, podcasts, images).

Culture war posts go in the culture war thread; all links must either include a submission statement or significant commentary. Bare links without those will be removed.

If in doubt, please post it!


Rules


Recommended Posts And Communities

Recommended Realtime Chats