[The Memes] Maybe the "Yes" people are buying "Yes" for the lulz. It's kinda fun to tell people that you bet that Jesus Christ would return this year!
Maybe you mean this expansively, but it doesn't ring true to me as stated because my main guess for many / most bettors would be more serious: they're "manifesting" or "hyperstitioning"--or more generally, they think there's some good effect from nudging other peoples's sense of what is other peoples's investment in 2nd-coming worlds.
Downvoting because it seems like you've barely read anything I wrote and also don't know anything about genetics or intelligence, and are now posting AI slop, but I will upvote a thoughtful post making an argument using information and logic that address why people think it might work.
Yeah this seems like an important question. I'm not sure what to think. Ideally someone with more background in medical ethics could address this. E.g. I'm not sure how to navigate what would happen if, for example, law enforcement claimed it needed access to some info (e.g. to enforce regulations about germline engineering, or to use in forensic investigation of a crime); or if there were a malpractice suit about a germline engineering clinic, or something. I'm also not sure what is standardly done, and what the good and bad results are, in situations where a child might have an interest in their parents not sharing some info about them.
But certainly, in a list of innovation-positive ethical guidelines for scientists and clinicians regarding germline engineering, some sort of strong protection of privacy would have to be included. This is a good point, thanks.
Yeah I'm not, like, trying to sneak this in as a law or something. It's a proposed policy principle, i.e. a proposed piece of culture.
My main motive here is just to figure out what a good world with germline engineering could/would look like, and a little bit to start promoting that vision as something to care about and work towards. I agree that practical technology will push the issue, but I think it's good to think about how to make the world with this technology good, rather than just deferring that. Besides the first-order thing where you're just supposed to try to make technology end up going well, it's also good to think about the question for cooperative reasons. For one thing, pushing technology ahead without thinking about whether or how it will turn out well is reckless / defecty, and separately it looks reckless / defecty. That would justify people pushing against accelerating the technology, and would give people reason to feel skittish about the area (because it contains people being reckless / defecty). For another thing, having a vision of a good world seems like it ought to be motivating to scientists and technologists.
near-evolutionarily-optimal range. That has not happened with intelligence,
What makes you think this? As I said, it's not clear to me that there's been much selection pressure for intelligence in the past few thousand years.
Also, the "evolutionary optimum" can change. E.g. calories are not much of a problem in the developed world, but that's recent.
Also, there's always an influx of de novo mutations, and evolution has limited selection power. I'm not clear on the math here exactly, and I think kman has suggested that mutational load isn't the main source of IQ-associated SNPs, but it demonstrates that it's far from ironclad logic to infer from evolutionary pressure on a trait that the trait should be near optimum in linear variants. The brain is one of the organs with the most diverse gene expression profile (I mean, more genes are expressed in the brain than in most other tissues); and IIRC most genes are expressed in the brain (not confident of this, maybe it's more like 1/3 or 1/2. But anyway, there's a lot of genes potentially relevant to brain function, so there's a lot of surface area for mutational load to drag things down a bit.
genes are not simply choosing a level of intelligence.
I don't know what you mean by this. Are you talking about pleiotropy? Between what and what? I mean of course genes do lots of things, but IIUC so far as we've observed, the correlations between most measured traits are pretty small (and usually positive between traits most people would judge desirable, e.g. lower risk of mental illness and higher intelligence).
I'll repeat that I'm not very learned about genetics, so if you want to convince even me in particular, the best way is to respond to the strongest case, which I can't present. But ok:
First I'll say that an empirical set of facts I'd quite like to have for many traits (disease, mental disease, IQ, personality) would be validation of tails. E.g. if you look at the bottom 1% on a disease PRS, what's the probability of disease? Similarly for IQ.
or beyond those?
I rarely make claims about going much beyond natural results; generally I think it's pretty plausible there's some meaningful thing we could feasibly do that's like +6 -- +8 SDs on intelligence, but I'm much less confident about the +8 SD claim, and not super confident of the +6 SD. Like, I think the default expectation ought to be that we can meaningfully get to +6 SDs; this seems like the straightforward conclusion. (I'm just restating the intuition / impression.)
Why do you expect that only a very small fraction of natural SNP differences are needed to get the extremes of natural results
Assuming linearity, the math is fairly straightforward. In the simplest model, with 10,000 fair +1/-1 coins (representing all the variance in a trait, so some coins are environmental), an SD is 50 coins and the average is 5,000. So there's 100 SDs of variance available. Obviously this is mostly meaningless in terms of the trait, as linearity would not remotely hold, but my point is that the issue isn't the math of additive selection. See here for more (e.g. about if the coins are biased https://tsvibt.blogspot.com/2022/08/the-power-of-selection.html#7-the-limits-of-selection ).
IQ seems to have thousands of small contributions from different regions. 10% of the variance is therefore in the ballpark of 10 trait SDs. Again, I'm not saying you can get to 250 IQ; what I'm saying is that the math of selection and variance isn't the problem. Lee et al. state "In the WLS, the MTAG score predicts 9.7% of the variance in cognitive performance[...]"; this was in 2018, I would bet we can do substantially better now.
Lee, James J., Robbee Wedow, Aysu Okbay, Edward Kong, Omeed Maghzian, Meghan Zacher, Tuan Anh Nguyen-Viet, et al. ‘Gene Discovery and Polygenic Prediction from a 1.1-Million-Person GWAS of Educational Attainment’. Nature Genetics 50, no. 8 (August 2018): 1112–21. https://doi.org/10.1038/s41588-018-0147-3.
Why do you expect that effects would be linear?
TBC I certainly don't expect the effects to be literally linear even in the typical human range; it's more that expect them to be fairly linear. Like if the answer is that a trait-mean couple that selects their child's genome to have a (carefully, accurately as best anyone can) predicted IQ of mean 170 actually tends to have a child with mean IQ 155, I'd shrug and be like "huh, that's weird and surprising, let's investigate and make sure to communicate this fact to parents"; and I think this possibility is nontrivially strategically relevant; and it means we should accurately describe this plausible outcome, in order to not overhype etc.; but I wouldn't be totally shocked. If the child tends to have a mean IQ of 125, I would be shocked, yeah. (The 170 vs. 155 thing would be hard to notice for a while because testing intelligence at that range is barely feasible, but just saying for illustration.)
Also, certainly I would expect strong nonlinearities at the extreme tails at some point. I'd certainly strongly advice against, maybe even condemn, pushing noticeably outside the regime of adaptedness.
Why do I expect the effects to be fairly linear in the human envelop, to the point where increasing a bunch of causal variants increases the trait?
Some of my impressions come from here: https://arxiv.org/pdf/1408.3421 "On the genetic architecture of intelligence and other quantitative traits", Stephen D.H. Hsu, 2014
and from https://gwern.net/embryo-selection
(TBC, I'm not saying these sources present the strongest arguments; I'm just saying where my impressions historically come from.)
I'm basing this off of selection, not editing. I haven't looked into the genetics stuff very much, because the bottleneck is biotech, not polygenic scores.
Would look forward to your rebuttal! I just hope you'll respond to the strongest arguments, not the weakest. In particular, if you want to argue against the potential effectiveness of selection methods, I think you'd want to either argue that PGSes aren't picking up causal variants at all (I mean, that there's a large amount of correlation that isn't causation); or that the causality would top out / have strongly diminishing returns on the trait. Selection methods would capture approximately all of the causal stuff that the PGS is picking up, even if it's not even due to SNPs but rather rarer SNVs. (However, this would not apply to population stratification or something; then I'd think you'd want to argue that this is much / most of what PGSes are picking up, and there'd be already-made counterarguments to this that you should respond to in order to be convincing.)
What I meant was smart people having more kids.
I mean, I'd encourage most people in general to choose to have kids, but yeah, trying to influence other people's reproductive choices on the level of persons is creepy and eugenicsy; it's much more practically and motivationally contiguous with even creepier and eugenicsier stuff such as racist immigation policies, etc.
but in what sense is that true that's not true of, say, your smarter relative having the kids instead?
In a huge quantitative sense. You and I differ at very roughly 4 million SNVs (maybe more like 5-10, but 4 is a convenient number). My sibling and I differ at roughly half that, 2 million SNVs. If I had a child without germline engineering, I'd pass on 2 million SNVs. If it's my sibling's child, it's 1 million SNVs.
If I use selection, I'm still passing on 2 million SNVs, though in a way selected for association with some traits--but the selection is pretty weak in the scale of SNVs; it'd correspond, morally, to something like hundreds or thousands of edits, I think (haven't gone through this carefully). In other words, O(0.1%), even if you count the alteration as being totally unrelated DNA. Similarly for editing. You could have a 180 IQ super-healthy kid with a natural lifespan of 110 years, with a genome that would be nearly indistinguishable from being your non-GE'd child. (Unless you whole genome sequence them and find a surprising coincidence of health and IQ alleles drawn from your own genome / a couple dozen SNPs that weren't in either parent.)
I mean, if you don't care about that, God bless you. I think most people do care about it though.
We already know how to get smarter kids
What are you refering to? It's true that there are companies currently offering embryo screening based on polygenic scores, in at least one case including IQ. These methods are fairly weak though. (I mean they're cool, and could have significant impact on diseases because for diseases the initial genomic vectoring is the most impactful on absolute disease risk. But they won't e.g. make tens of thousands of world-class intellects. See https://www.lesswrong.com/posts/2w6hjptanQ3cDyDw7/methods-for-strong-human-germline-engineering#Method__Simple_embryo_selection .)
People also do mate choice, but this is, while not exactly zero-sum, kinda zero-sum, and still doesn't get you super-healthy long-lived world-class intellects with high probability. Or do you mean more broadly education and stuff?
Do you expect that to change? Why?
Generally the basic reason I expect it to change is that the technology will have big benefits, so people will think about how to do it without also genociding people.
By default, I expect it to change, though it could happen pretty slowly. Or rather, it could be quite delayed. That seems to already be the case; I think if we'd wanted to make this technology, and had wanted that for, say, 20 years, we could probably have already had it.
I somewhat expect that once the first quite noticeable germline engineering is proven out, people will want to use the technology themselves. E.g. you have a bunch of 17-year-olds who are already quite impressive intellectually, and not in a kinda-cringe prodigy way but in an actual way; and you have 10000 people age 0-20 who have a noticeably miniscule rate of death due to disease (of course they'd still die from accidents and such).
Also, it's worth noting that public sentiment about germline genomic engineering is much better described as "quite mixed" rather than "anti". I want to do a more thorough review later, but the few polls I've seen get favorability numbers for germline engineering between 20% and 70% (depending on the question, e.g. about preventing disease vs. increasing intelligence or other capacities; and depending on the country). The few debates I've watched are similarly, coarsely, half and half.
I hope to help this process along at least a bit, though of course I probably can't do much. Partly it's just communicating what the tech will do, and how we know that, etc. Partly it's communicating about misconceptions. Partly it's making the tech actually be beneficial; a big thing here is making it strong (hence effective and greatly beneficial to individuals; and also cheap, and therefore widely accessible; see https://www.lesswrong.com/posts/2w6hjptanQ3cDyDw7/methods-for-strong-human-germline-engineering#Strong_GV_and_why_it_matters ). Partly it's thinking about to set up society to avoid possible socially emergent downsides. (Yes, I understand I contribute at most a very small amount to "setting up society"; the idea is to contribute to figuring out what that could look like, to be beneficial.)
current societies have comprehensively decided that eugenics programs are bad
I'm still unsure, and curious to learn, what people actually object to. (My current list is here, though it's trying to be complete rather than emphasizing what people most care about: https://berkeleygenomics.org/articles/Potential_perils_of_germline_genomic_engineering.html .) The original, and most abhorrent, things that are called eugenics, are actually bad. But they are in most ways diametrically opposed to genomic emancipation. There is overlap, which people could be rightly worried about--namely, generally caring about genes enough to somehow intervene on some genes. But the logic is fundamentally opposed: eugenics says:
Everyone should have children with Good genes, so that We/I don't have to live with the negative externalities of Your bad genes (/race).
whereas genomic emancipation (in its pure, and therefore not completely workable, form) says
Everyone should be empowered to have as much (well-informed, high-precision, skillful) influence over their future child's genome, so that I can give My child a good life by My lights, and You can give Your child a good life by Your lights, which will probably be good because empowered pursuit of diverse human goods is probably good in aggregate.
I don't know what they have in mind, and I agree the first obvious thing to do is just get more data and try linear models. But there's plenty of reason to expect gains from nonlinear models, since broadsense heritability is higher than narrowsense, and due to missing heritability (though maybe it ought to be missing given our datasets), and due to theoretical reasons (quite plausibly there's multiple factors, as straw!tailcalled has described, e.g. in an OR of ANDs circuit; and generally nonlinearities, e.g. U-shaped responses in latents like "how many neurons to grow").
My guess, without knowing much, is that one of the first sorts of things to try is small circuits. A deep neural net (i.e. differentiable circuit) is a big circuit; it has many many hidden nodes (latent variables). A linear model is a tiny circuit: it has, say, one latent (the linear thing), maybe with a nonlinearity applied to that latent. (We're not counting the input nodes.)
What about small but not tiny circuits? You could have, for example, a sum of ten lognormals, or a product of ten sums. At a guess, maybe this sort of thing