site banner

Making Cognitive Enhancement Palatable

parrhesia.substack.com

SS: I think that cognitive genetic enhancement is important for ensuring we have a better and lasting future. Many people have an intuitive dislike for the idea of using genetic enhancement to make a baby smarter but have little issue with in vitro fertilization (IVF). I try to build from a foundation of the acceptable practice of IVF to PGT-P for IQ.

11
Jump in the discussion.

No email address required.

[offtopic] ...Imagine if a country begins to engineer radiation-immune people.

I'm against cognitive enhancement because I fail to see a road where result of human enhancement isn't a speciation event where the top 0.01% of humanity acquires functionally unlimited power relative to the common person to find themselves on a footing closer to man-and-chimp with the rest of us barely auged or semi-auged proles. At that point we'll have about as much power to resist as the monkeys do if the gene modded ubermensch aristocrats decide to cull the rest of us useless eaters. Barring about a billion safeguards to stop this (probably inevitable) future I'm much more in favour of banning it all outright. Unless you're at the apex of the elite and have a good idea that your great grandkids will be similarly positioned once this tech really starts taking off, being in favour of human augmentation is like a neanderthal in favour of early humans making landfall in his neighbourhood.

This is imo mostly nonsense if applied to genetic enhancements, for a number of reasons.

  1. Both sequencing and IVF are already very accessible for the middle class. It's probably not far off from becoming accessible for the poor. The only way to make it exclusive to the uber-rich is, ironically, by outlawing it. CRISPRing people isn't really prohibitively expensive, either, and becoming cheaper as well.

  2. Better selection has pretty much zero marginal cost per person. It's 100% developmental cost, and once we know which genes matter, it's widely usable.

  3. This hasn't happened even when it made sense. For example, the same argument could have been applied to all technological advancements that have significant marginal cost per person. Cars, better healthcare/hygiene, education, electricity, etc. all of those are much more plausible to cause a runaway effect with the rich getting disproportionate ROI and everyone else falling behind since they have nothing to invest, but it didn't. Currently, AI is much more plausible to cause a runaway effect, and human enhancements are if anything one of the more plausible ways of competing with them.

  4. My main objection: What do we want to select for, and how do we do it plausibly? We want to select for generalised success (which includes IQ, health, social success, educational attainment, etc), and we find the genes associated with those by comparing the successful with the not-so-successful. As such, genetic selection uniquely disproportionally benefits the downtrodden, since the successful will already hold a decent chunk of the genetic enhancements we want to select for.

As a concrete example, we know that specific genetic variants are almost mandatory to become a top-tier runner. Making targeted genetic enhancements accessible does not benefit the top-tier runners, since they already have these. It does benefit everyone else, though. The same principle applies to all attributes, even complex ones that are associated with thousands of variants. There is also both significant evidence in favor of diminishing returns for many attributes, as well as strong theoretic arguments (for example, there is a physical limit to how fast a carbon-based, two-legged lifeform can run. As you approach that limit, "better" genes do less and less, meaning that improvements are largest for the relatively genetically cursed).

The only plausible objection here is that in the lab, we can find even better variants - which nobody holds - so that the top-tier can benefit again. However, to my knowledge we have no consistent approach to do this, and since trial runs on animals only get you so far, the first generation humans trying them will still have excessive risks. I personally wouldn't try such a variant even if I could. Much better to use variants that already plenty of successful people have, so at worst it's something marginally negative that isn't holding them back much and on average it should benefit them.

It bears repeating that disproportionate returns for the rich is something I'm always worried about with most new technologies, but gene editing/selection is uniquely implausible to cause that. On the other hand I'm very worried about AI, since factories and companies purely or overwhelmingly staffed by AI can make the elite independent of the plebes in a way they have never been before.

(for example, there is a physical limit to how fast a carbon-based, two-legged lifeform can run.

um, this is somewhat true. But also think about carbon-based human-size brain with effective density those as insects have. Our human brains have low neuron density! There's so much space to grow. Solving this would be hard however, since apparently evolution didn't do it.

It's irrelevant because of AI, which will be (eventually) much smarter than any human, enhanced or not

But why can't all common people just get the genetic modifications for their children, and then become as smart as the elites? Isn't that the obvious outcome - okay, you're not that smart, but now your kids can be just as smart as the elites' kids because they'll have the same gene sequences?

We won't be limited to swapping in existing sequences forever. At some point down the line we'll have the capability to create entirely synthetic man made sequences that expand our capabilities beyond what any currently existing genes can provide.

The process of evolution directly involved randomly generated or modified 'entirely new sequences' that were tested by physics and nature. How is this different in some absolute sense? It's certainly dangerous, and possible to mess up, but

Again though AI will flip the table before we can do that much.

Random chance, undirected evolutionary shifts aren't the same as the top 0.1% of the human race self editing themselves into having 300 IQ and 500 year lifespans and leaving the rest of the human race in the dust. For anyone not part of that 0.1% I don't see it being a positive development. I'm sure AI will have fairly analogous effects too.

I don't see why it'd be only the top .1% though? Gene editing is very transferrable and benefits the bottom 10% much more than the top .1% in a relative sense, as the known good sequences from the top .1% can just be given to the middle and bottom, while to improve the top .1% one needs to somehow figure out better nonexistent sequences (or, more accurately the good sequences are the ones that the PGS finds, it's not like the top .1% are so because of super special top .1% genes, they have lots of small-effect good genes that the bottom 10% has some of but less of + various other stuff we don't understand). And good medical and nonmedical technologies are very quickly diffused from the rich to the middle-class in the modern era, both to make money and because most people are progressive.

This narrative is as applicable to AI, which can confer godlike power on an arbitrarily tiny elite class very soon, as it is inapplicable to genetics. Even the smartest naturally-occurring people begin substantially contributing at about 15 years of age at the earliest. This means the first generation of genetically augmented kids (or more realistically, clones of Terry Tao or John von Neumann, though we can't clone him right now) will create hype with their early performance many years before they can help advance reproductive technology beyond the level comprehensible for baseline humans, and upon maturation will be outpaced by the successive cohorts using better and, as it usually happens, cheaper tech stack. Their initial work will be public-facing too. When this whole batch begins to pay rent, literally and figuratively, access to enhancement will have proliferated way deeper than the top 0.01%.

Unless, of course, someone succeeds at banning it – for civilians, that is.

Theoretically, you could support a global ban but I doubt that’s going to be at all feasible. Deviation from a global ban will be too tempting.

I would like my culture and society (USA) to have the jump start on cognitive enhancement rather than China or some other authoritarian nation.

Thoughts?

The global ban is more wishful thinking than anything else, I have no doubt plenty of countries will go laxer than others on rules on this and in the age of the airplane any suitably motivated rich person could just jet set to wherever lets them shop for the most modified aug babies they want. The country/culture point is also legitimate, and I'd probably prefer to be ruled over by a western class of gene augged aristocrats than a Chinese or Russian one, but my main concern is that these uncontestable gene augged aristocrats are probably the inevitable end result of this tech in practice.

Why would gene selection technology be limited to the .01%? I can see a case it would be limited to developed countries, or be limited to the very wealthy in those countries for a short period, bit fundamentally technologies like crispr do not require significant resources to use other than the initial investment required to learn how to use them.

I think the risk fully comes in once you get synthetic gene modding instead of just swapping in naturally occurring genes mapped off humans, at that point it could easily turn into an arms race over ever escalating augmentations that leave everyone not on the bleeding edge in the dust. Top level firms aimed at the wealthy could entrench a new ruling caste where the proles get [current year -5] gene augs compared to the latest and greatest augs available at 5,000,000 per kid.

I don't see an economic reason why cutting edge gene modding would be exclusively catered to the rich. If you want a return on investment you want the broadest market possible. High cost products are usually products that have high cost of physical resources like cars, or high labor costs. Gene modification is essentially selling data. It's market would be more similar to a market for music, books, movies, etc than cars. Therefore its largest problem would be creating excludability and financializing the asset.

The only way this might not be the case is through significant regulatory capture, which considering the total non-functionality of the FDA is possible. The only problem is the FDA's incentives are not aligned with the wealthy, but with self preservation. This presents itself mostly through absurd overcautiouness and regulatory violence against uncooperative corporations.

That's exactly how any new technology works. First, it's so expensive only the idle rich can afford to play with it. Then as more development occurs, funded by those same idle rich, it becomes cheap enough for the working rich. Then the middle class. Then, finally, for everyone.

To my mind, the biggest argument against "the rich keep it all to themselves"—aside from that never having happened with any other tech—is this: national borders. If Country A limits cognitive enhancement to its elites and Country B gives it to everyone, pretty soon Country A will be irrelevant.

Most technologies can't lead to actual speciation events and a total divergence of our species into possibly thousands of different branches with wildly different qualities and capabilities. Self modification can.

I also worry about decoupling for narrow optimisations like IQ. I'm strongly in favor of general health and general success/competence selection, however. Both would imply some degree of IQ optimisation even if IQ itself is not explicitly represented. I wouldn't necessarily be opposed to adding explicit IQ selection in addition to the others, but a lot of the common objections fall apart if we simply leave that out and I think we will have the majority of the gains even without it anyway, so I'm tentatively against it for the moment.

Thing is, this is going to be for middle-class and upwards of that. Nobody cares about making the kids of the underclass all IQ 120 at minimum, because they're still going to live in crappy single-parent homes in crime-riddled shitholes and go to schools where metal detectors and armed security guards are needed because the little darlings shoot each other (starting age six, I understand from this edifying story).

There's no point in educating these kids above their station, and why would you want them smart enough to recognise their shitty rung of the ladder, be dissatisfied and possibly able to organise to do something more effective than street protests and burning down police stations? They are always going to be at the bottom of the ladder, because great now you're IQ 120. Meanwhile, the rungs above you are making sure their embryos are selected to be IQ 130-140 before implantation in the hand-picked surrogate for gestation, and the classes above you will keep their grip on what they've already got. Before, identifying and recruiting any smart underlings was worth it in order to co-opt that productivity for your own benefit, but now that you can reliably be sure all your kids are smart enough for the new world, you don't need recruits that have to be taught how to be middle-class.

Any jobs the smart underlings can do will be automated/replaced by AI away, and holding on to the slots in the PMC are going to be more competitive as white collar jobs become vulnerable to replacement by AI. Why bother creating competition from below? Leave the masses dumb and ignorant as your children, and their children, pull further and further away, and throw sops to keeping the peace with bread and circuses and legalised drugs, until you can figure out a way to prevent them reproducing at all. Robots will be cheaper and better anyway.

Nobody cares about making the kids of the underclass all IQ 120 at minimum, because they're still going to live in crappy single-parent homes in crime-riddled shitholes and go to schools where metal detectors and armed security guards are needed because the little darlings shoot each other

This is such a quintessentially leftist, zero-sum, all-goods-are-positional-goods, spoils-system, school-pecking-hierarchy point of view. Do you think violent crime is an unchangeable quantity that can only be shifted around like liquid in a hydraulic system, that there's a fixed supply of non-crappy homes such that only a few get to have them, that the economy just needs someone to be on obese and on welfare? That all that IQ score represents is deviation from the median rather than a noisy proxy for absolute cognitive ability, so any improvement will only produce accelerated running in place?

It reminds me of The Citadel of Ricks from that damn cartoon – which is either brilliant satire or an expression of an extraordinarily pathological world model, or both, I guess. Unironic superhumans who can individually construct physics-breaking tools and AI-powered assistants – exploiting each other, building elaborate chains of abuse, maintaining artificial scarcity, ineffectively doing menial jobs to the point of burnout, in a mockery of a dysfunctional American city.

I'd understand it if you predicted a culling, seeing as you expect AI to be economically useful. But who the fuck needs to maintain a permanent underclass that requires babysitting, armed cops and shitty schools? Would our genetically-augmented overlords invest into reintroducing lead-laced gas and asbestos walls, while they're at it?

If I am to be blunt, I think this entire line of thinking is a cope. I think people hide in this worldview from admitting that some groups they care about are just fucked up – not relatively or conventionally, but in some rigorously definable, absolute, physical, embarrassing but potentially rectifiable sense. That there are hard limits on what can be achieved with «85 IQ», which do not follow from it being 1 SD below the «100 IQ» and thus non-competitive in a market economy. Limits on impulse control, cooperation, general decision-making. And moving beyond these limits allows for building an all-around nicer society for everyone, without extra requirements for morality of its elites.

Sure, they can try to cement their advantage – though that'd require the «PMC» to coordinate far more than they've been able so far. On the other hand, the main reason to avoid falling down the hierarchy is knowing how much it sucks down there. Sucks absolutely, not relatively. Scarcity begets crab mentality. And stupidity multiplies scarcity.

There is a huge demand for horror stories about the Japanese way of life. Japanese elites are not inhumanly virtuous or selfless; Japanese culture and society are rather hierarchical. There is, I'm sure, some violence, some bullying, some crappy housing in Tokyo – just very little, relatively speaking. But there certainly are no armed security guards in Tokyo schools. It's not because the undesirables have all been tucked away somewhere on Hokkaido. It's partially, but in a very large part, because there does not exist an underclass with sub-90 IQ in modern Japan. And no one deserves to be born into such an underclass.


@Eetan would no doubt gleefully remind me that the Japanese public has no guns anyway, so those guards wouldn't have had anything to do. Well, fair point. But there's very little non-gun violence either. There's virtually no shit like this in Japan. If the demand to lay down arms and obey can be made legitimate, guaranteeing extreme safety is one of the better ideas.

There's virtually no shit like this in Japan

I look at this and I'm amazed how good the facilities are compared to school I was in, are.

Thing is, this is going to be for middle-class and upwards of that. Nobody cares about making the kids of the underclass all IQ 120 at minimum, because they're still going to live in crappy single-parent homes in crime-riddled shitholes and go to schools where metal detectors and armed security guards are needed because the little darlings shoot each other

So...they'd be in the same situation as a variety of Asian immigrant groups? High IQ children, poor parents, bad schools, sounds like the Hmong a generation ago. Recall that for a while, "what about the Hmong" was often used as an argument proving Asians aren't a model minority.

As of 2019, Hmong household income ($68k/year) was higher than average ($65.7k).

https://www.pewresearch.org/social-trends/fact-sheet/asian-americans-hmong-in-the-u-s/

If they have the same IQ as poor Asian kids, they might also have the same crime rate and out of wedlock birth rates as poor Asian kids. That would solve the problems of going to schools full of stabby kids pretty quickly and the problem of single parent homes in a generation or so.

Once you solve the problem of "poor people are terrible to be around" by raising their IQ, poverty in America isn't actually a big deal. If you disagree, try to name a good or service you think poor Americans lack.

They are always going to be at the bottom of the ladder, because great now you're IQ 120. Meanwhile, the rungs above you are making sure their embryos are selected to be IQ 130-140

You seem to be assuming that technology for embryo selection will develop along the lines of cars (where the 10x more expensive car is perhaps 2x as good as a Corolla) as opposed to along the lines of cell phones (where even Elon Musk uses an iPhone). Why do you believe this?

I can see one plausible story: regulators prevent poor parents from accessing the technology while the rich get it done in Macau.

Don't agree here. I think raising IQ will unlock a lot of cognitive capital, which is good for the economy and society, and help mitigate or reverse these social problems. IQ is positively correlated with virtually every social metric there is.

Any jobs the smart underlings can do will be automated/replaced by AI away, and holding on to the slots in the PMC are going to be more competitive as white collar jobs become vulnerable to replacement by AI. Why bother creating competition from below?

But this will be offset to some degree by increased economic growth ,and hence more PMC jobs, due to increased productivity from having more smart people overall, and higher standards of living too.

IQ is positively correlated with virtually every social metric there is.

I midly agree a-priori although too high intelligence can make someone incompatible with others (see inferential distance) and make someones become too serious, I believe hyperintellectualism reduce abilities to produce humor/jokes.

But isn't the common belief that very high IQ people are contingently but empirically correlated with cognitive disorders such as e.g the autism/asperger spectrum?

But isn't the common belief that very high IQ people are contingently but empirically correlated with cognitive disorders such as e.g the autism/asperger spectrum?

The common belief is that this common belief is caused by a common misperception. Very high IQ people who lack cognitive disorders are less interesting than very high IQ people who have cognitive disorders, so they're less noticed, in favor of very high IQ people who have cognitive disorders. And so people falsely get the notion that having cognitive disorders is more common among very high IQ people.

However, that is just a common belief. I don't know if there's extensive research on this, and even if there were, if the extensive research is credible.

It seems autism is not associated with high IQ https://pubmed.ncbi.nlm.nih.gov/21272389/

but on the other hand we have this https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4927579/

also there is more basis for asperger https://pubmed.ncbi.nlm.nih.gov/24362849/

maybe this is true for way outliers, but people with IQs in the 110-130 range seem mostly well adjusted. The terman study showed high IQ more successful overall. also, most of the maladaptive problems arise from dealing with less intelligent people. so raising national IQ could mitigate this too.

Nobody cares about making the kids of the underclass all IQ 120 at minimum, because they're still going to live in crappy single-parent homes in crime-riddled shitholes and go to schools where metal detectors and armed security guards are needed because the little darlings shoot each other

No they won't. That's what an IQ of 120 means. "A ghetto/barrio/alternative name for low-class-hell-hole isn’t a physical location, its people." Poor areas are not awful because of tragic dirt; they are awful because they are filled with stupid, violent, impulsive people.

(A surprising amount of people don't seem to realize this; they talk about good neighborhoods and bad neighborhoods as if rich people used their wealth to hog all the good real-state where shootings and robberies and rapes and so on don't happen, as if those were natural phenomenon like lighting bolts rather than something caused by the people who actually live in those neighborhoods; likewise, complains about disparities in funding, as if schools in rich areas were taking advantage of a gold mine they unfairly took over rather than taxing the economic surplus produced by superior human capital)

Education doesn't do shit because trying to teach algebra to a boy with an IQ of 85 is a waste of time. Increasing his IQ to ONE HUNDRED AND FUCKING TWENTY would be the biggest improvement in the human condition since the industrial revolution.

Even if they start materially poorer, you have eliminated all the dysfunction. College students also live in material poverty, but they have much better lives, because they are smart and hard-working and nonviolent. "If you take the exact same facilities and you fill them with inner city gang members, drug addicts, ex-convicts, alcoholics, prostitutes, and single mothers, you get a housing project."

And just like college students, after a while those 120 IQ kids will start accumulating capital and lifting themselves out of poverty. It's much easier to follow the Success Sequence when you have the intelligence of an undergrad.

Yes, they will still be below elite kids who got uplifted to an IQ of 140, but that's relative poverty, not absolute poverty. Caring about that is the politics of envy. "Thou shalt not covet thy neighbor's house, thou shalt not covet thy neighbor's wife, nor his manservant, nor his maidservant, nor his ox, nor his ass, nor any thing that is thy neighbor's." It's societal poison.

I can't find it, but there was a study a few years ago that showed that people would pay a lof of money to be able to control who who their neighbours were.

The current set of parents are not going to be IQ 120. That's the problem. And the whole "a rising tide lifts all boats" model has been shown not to work. Higher intelligence is valuable in part because of scarcity. When everyone is smart, then we hit the treadmill: everything has now been boosted up a step. Remember the credentialism debate? Job that you could get with no qualifications except a strong back and willingness to work hard -> job now needs high school diploma -> job now needs some form of degree/further education -> job has been automated away because machines are cheaper and faster.

120 will become the new 85. That's what you are not seeing. As for the "more smart people = more productivity = more PMC jobs", how is that working out with the wave of tech layoffs starting right now? When there's an economic downturn, "stock price valuation" becomes a lot more important than "we have all these smart people working for us that we have employed due to DEI initiatives". They trim down the workforce to the bone, and being smart and productive now means you are expected to do the work of two people.

I'm sceptical, because all this moon-shot optimism of the past has turned out very differently in reality. "Increased automation means people will have so much leisure time!" No, increased automation meant "now that we can be in communication with the other side of the world, you have to work weirder hours so you can be there when the markets in Tokyo open" and all the "yeah you're expected to work 60+ hours starting out to get anywhere in your career".

Same with IQ enhancement: yes, it will have benefits. No, the benefits will stay with the usual people and not "now everyone is a minimum of IQ 120, poverty is solved!" We've been 'solving' poverty for a long time, and yet it still lingers on.

120 might be the new 85 in terms of the relative position in society but they will still not be as violent, or stupid as an 85.

And the economy will change as economies are dynamic, if everybody is > 120 then some of the more entrepreneurial 120s will create employment for the people around that IQ level.

Maybe you are just afraid of competition.

Can't make it without an artificial class barrier to keep out the proles who weren't raised middle class?

120 will become the new 85.

There was Homo erectus once. Now it's Homo sapiens. When the latter replaced the former, did "120 become the new 85"? Do the sapiens enjoy better society?

No, increased automation meant "now that we can be in communication with the other side of the world, you have to work weirder hours so you can be there when the markets in Tokyo open"

19th century, no automation: you die from hunger.

I'm sceptical, because all this moon-shot optimism of the past has turned out very differently in reality. "Increased automation means people will have so much leisure time!" No, increased automation meant

It actually meant exactly that. People work way less hours now. (That's not even addressing the nature of work, namely that they're not working nearly as much in physically harmful and toxic environments, hell, even janitors today have less demeaning tools).

Not everything is a zero-sum game. No, really.

I notice I (IQ 130) followed the success sequence perfectly but I'm still a social outcast.

The success sequence is not a tool for avoiding social isolation; it just averts poverty. Are you poor? If not, it did its job.

Fair point.

Poor areas are not awful because of tragic dirt; they are awful because they are filled with stupid, violent, impulsive people.

Yes, but its not clear IQ 120 by itself would fix that. Genetic enhancement would be a thus-far-impossible level of decoupling between intelligence and other correlates of good outcomes, it might break currently observed correlations.

We still don't know of any drawbacks of IQ (or genetic correlates of IQ) no matter how many corrections we do; nor any interesting non-additive effects. Cognitive ability is independently causal for all of the good stuff, albeit it doesn't explain the whole of it; the common factor behind cognitive ability and good stuff (physical basis of g, apparently just high neurological and more general cellular functioning) has got to explain an even greater proportion, and this is what we'd be maximizing via IQ selection. For IQ-good outcomes relationship to be substantially explained by a confounder, it would require a very sneaky pervasive bias everyone is missing.

Would our hypothetical children who were embryo-selected for IQ PGS alone to the effect of 20 (or 40) points be perfectly equal in success to their natural intellectual peers? Indeed, it's not clear. Would they be better off – in the expected direction of less dysfunction – than the baseline, or rather, than random implanted embryos? You can bet on it.

P.S. This isn't really the best strategy for embryo selection, anyway. Gwern finds that:

selection can be made more effective by selecting on multiple phenotype traits: considering an example using 7 traits (IQ/height/BMI/diabetes/ADHD/bipolar/schizophrenia), there is a factor gain over IQ alone; the outperformance of multiple selection remains after adjusting for genetic correlations & polygenic scores and using a broader set of 16 traits.

I mostly agree.

nor any interesting non-additive effects

I dont know what people have with non-additive effects. In a highly polygenic trait, non-additive effects of genes are hard to detect because theyre almost certainly irrelevant even if real.

Would they be better off – in the expected direction of less dysfunction – than the baseline, or rather, than random implanted embryos? You can bet on it.

That is mostly what I expect as well. Im just saying that theres a lot more evidence that it "might as well be" IQ/genetics, than that it actually is.

Why would it be a decoupling?

One example for what this could look like is low mutational load: Its also correlated with all the good things, including IQ, and certainly causally upstream of it. Genetic IQ enhancement in a narrow sense wouldnt fix that, and if it has any effects not mediated by IQ (it almost certainly does), you wouldnt get those, whereas currently they strongly correlate with IQ. Now, in this case, if you know about it and are already doing genetic enhancements, its easy to fix that as well. But there could be more things like this.

Basically, noone has run an RCT on IQ increases, because we havent been able to do them.

Genetic IQ enhancement in a narrow sense wouldnt fix that,

In some sense, low mutational load is easier to get rid of: if you have good genome editing, but few data samples for study, you can just eliminate rare alleles in embryo rather than look for correlations in huge dataset for useful alleles. (Some have speculated that bc. long-term effect of purging of rare alleles is bad, there might be action to require people who purge old rare alleles to also take some new rare alleles).

Also, IQ is much more affected by mutational load that most traits. If our hypothetical uplifted 120 IQs would have somewhat shorter lifespans than natural 120 IQs, it's still a win.

noone has run an RCT on IQ increases, because we havent been able to do them.

No. We aren't doing RCT on IQ increases because our society is hostile towards it. We could have just cloned some genuises already, it's simpler, but we don't.

why wouldn't it?

Because you’re not gene editing, you’re picking embryos which has full genomes rather than pieces.

In humans, good traits are correlated. Beautiful people tend to be smarter. Smarter people tend to be harder working. And so on. It would be amazing if selecting for high-IQ embryos did not also select for high-IQ correlates.

If that were true I would expect a great number of physicist and mathematicians to be hot: in my academic experience that is not true, and even looking at the average celebrity scientist they not seem particularly good looking. Conversely I would expect many models to be at least bright: it doesn't seem so. There could be various explanation: I only know about the statistical outliers, people that enjoy Math and Physics are uglier than average so they are statistical outlier regarding the correlation between IQ and beauty or this correlation is extremely weak.

https://www.lesswrong.com/posts/dC7mP5nSwvpL65Qu5/why-the-tails-come-apartweak

Except that high-IQ is also strongly correlated with sociopathy, neuroticism, suicide, sexual deviancy, etc... so who's to say that you're actually selecting for a positive trait and that you wouldn't get better results selecting for a quality like conscientiousness or physical fitness than you would IQ?

Conversations in this grain tend to run short because it's a rare example of a political question where everyone acknowledges the disagreement exists at a values rather than empirical level.

The moment such enhancements become optional, they are de facto mandatory. De jure too, eventually. As a human conservationist, my values oppose embracing extinction on the grounds trans-homo sapiens sapiens will be smarter and more efficient than us. AGI will be incomparably smarter and more efficient than us. Would you extend your logic to be in favor of replacing us with it?

If you want people to accept this kind of logic, you're going to have to Straussian about it and pretend that slippery slope arguments about human modification are silly, rather than laying out the slippery slope in a seven point numbered list and then talking about how great it will be to have a society full of super-geniuses in the closing paragraph.

Let's say we ban gene editing entirely, and practice an older, simpler form of eugenics - freeze the sperm of the best, smartest, most honorable and strongest people in society, and have a significant portion of children come from them. And somehow everyone magically agrees this is good and not state-sponsored cuckoldry or anything. That's hardly 'transhumanism' or extinction-level. Would you oppose that?

Let's say we ban gene editing entirely, and practice an older, simpler form of eugenics - freeze the sperm of the best, smartest, most honorable and strongest people in society, and have a significant portion of children come from them. And somehow everyone magically agrees this is good and not state-sponsored cuckoldry or anything. That's hardly 'transhumanism' or extinction-level. Would you oppose that?

This is an extension of the normal process of sexual selection that's been around with us as long as humans have been around. In practice, the society that does something like this will begin to resemble the many polygamous societies, which carries its own can of worms. (The padishah khan gets thousands of wives, his beys get dozens, his ghazi get two or three, and slave gets perhaps one if he's very lucky.) But I would not consider this transhumanism or autogenocide -- if perhaps unfair to the back half of the bell curve. Sexual selection is, on its own, a horror we have accustomed ourselves to to the point of not noticing -- but it is one that is intrinsic to our nature, much like eating brains is intrinsic to the botfly.

Gene editing for mental traits is an extremely obvious Schelling fence. It is hypothetically possible to do gene editing for certain mental traits that would not change the nature of humanity. However, I am confident if we take that step as a species, we will get Transhumanist-Gandhi, not 95%-Human-Gandhi.

What specific bad things will happen with 'transhumanism' that weren't already happening with natural gene editing, random mutation and combination of sequences? The genes that underlie biological functions were 'modified' - randomly mutating, duplicating, subsequences jumping from one place to another. And they are very useful!

we will get Transhumanist-Gandhi

What could possibly go wrong?

what do you mean by driving extinction..

Are you against the Flynn effect?

People nowadays have significantly higher IQs than past generious simply out of better acess to food and education.

past generations are being replaced and yet mankind is still dumb as a brick.

without talking about genetics or epigenetics, many people take nootropics that are cheap and mostly have benign or non existent side effects.

for example CDP-choline or racetams.

people that eat many eggs have transiently significantly better memory abilities than the ones that do not.

nicotine is a very potent nootropic agent, etc

Instead of virtue signaling about non-existent problems (the narrative than a newer generation of geniuses would make you appear subhuman is unrealistic, I consider myself to have considerably higher intelligence than my friends and yet we can still be friends and I respect them more than I respect my own self), you should try taking benign nootropics like low dose (200mg) CDP-choline daily it will midly improve your memory recall abilities and most importantly will slow down the ageing of your brain (via e.g. increased myelin production)

what do you mean by driving extinction..

Are you against the Flynn effect?

People nowadays have significantly higher IQs than past generious simply out of better acess to food and education.

I define a human species, subspecies, "race", ethnicity, family group, etc by genotype of mental attributes. Higher IQ, more conscientiousness, different levels of neuroticism and openness to experience due to the environment (Flynn effect or drugs) falls under phenotype. A group of monkey living in Delhi does not go extinct if tourists train them to bow and solve puzzles for treats, making them smarter and leaving more time to groom themselves.

On the other hand, if you were to use artificial selection to eliminate people with a genotype that may articulate as having lower cortical volume or a more active limbic system, you have exterminated that group of humans.

Instead of virtue signaling about non-existent problems

I said at the very beginning this is a question of values. A value does not need to be justified by pragmatism. (Indeed, you have smuggled into your complaint the totally unjustified normative value that economic growth and a higher functioning society are worth developing, or that a smarter human is "better" than a dumber human.)

When the AGI congress of 2068 proposes a bill for neutering the economically useless welfare recipient humans and uploading their consciousnesses into a more efficient digital form that allows them to hold down a job — including getting rid of those pesky family instincts, high aggression, and ability to lie — I suspect you will object, too. Your objections will not be pragmatically justified.

Other people will replace the currently existing people. I think it is good if the people that replace us are happy, healthy, and intelligent. If they are so much healthier, happier, and more intelligent than they are "posthuman," then so be it. When you say extinction, it elicits thoughts of everyone dying, which I explicitly want to avoid with the aid of superintelligent people.

Well, I am not Straussian, and I think the logic in my article is decent.

I have extensively studied almost all nootropics and therefore know most of the pharmacological markers that influence intelligence, however I have almost zero knowledge about the genetics correlate of hypermnesy, intelligence and rationality.

Does anyone know what are key "genius" genes?

IMHO I don't think intelligence is mostly bottlenecked by standard genetics, I am far more inclined to believe in early nurturial critical periods.

But I do believe in the potential of a practice that has long been forgotten, artificial cranial elongation:

https://en.wikipedia.org/wiki/Artificial_cranial_deformation#/media/File:Afrasiab,_elongated_skull_600-800_CE,_Afrasiab_Museum.jpg

According to scientists their brain volume is not bigger than a normal human being, and as such cranial elongation only solves one bottleneck: space.

If we identify the mechanics that drive the closing of the brain surface in teenage years and find a way to delay it, or if we intensify neuron and axonal growth during childhood via e.g. BDNF, NGF, etc and maybe growth hormone then we might achieve humans with the biggest brains ever.

Don't get me wrong brain size is only one metric and there might be a bimodal curve to it however our current brain size as shown by studies is totally constrained by a mere contingent scarce sugar consumption optimisation.

In another world, I'd be concerned about raising a generation that was significantly more intelligent than my own. This method surely wouldn't raise intelligence by more than a standard deviation - but once we accept this step, the next step is actively editing genes or doing some kind of stem cell shenanigans to pick out 1 in 10,000 or higher babies. We end up with a dynamic where the young have all the intelligence and the old have all the wealth and power. A very unstable dynamic!

But in this world AI will likely render it all obsolete. If we moved today and this technology has serious effects in 20 years, gen 2 genetic engineering might have serious effects in 25 or 30 years at the earliest. Too late.