Couldn’t GPU restrictions still make them more expensive? Like let’s say tomorrow that we impose a tax on all new hardware that can be used to train neural networks such that any improvements in performance will be cancelled out by additional taxes. Wouldn’t that also slow down or even stop the growth of smaller training runs?
GeneSmith
Karma: 1,740
No, the option to select against all diseases in proportion to their impact on quality-adjusted lifespan is the default. But parents can re-do the calculation to like take age of onset into account if they want. Or they could add other non-disease traits to their selection criteria (like intelligence, as estimated by some third party service).
I agree, it’s very sub-optimal for parents to have to do all this themselves.
It’s worth noting that most of the major news orgs passed on this story despite being offered the opportunity to cover it. We don’t know why they did it yet, but given that various orgs have covered the Snowden documents and other whistleblowers that the government very much didn’t like, my guess is they did it for reasons related to the quality of the story rather than any conversations with government officials who encouraged them not to cover it.
My priors against us having discovered alien tech are very high, though not literally infinite.
But I still don’t have a clear story for exactly what’s going on. Most of the videos of UFOs look pretty similar: silvery orbs flying around at very high speed. I haven’t yet heard an explanation of how this could be explained by camera artefacts, weather phenomena, or anything else.
Other videos like this one released by the Navy show non-spherical objects that even rotate while moving. I struggle to think of what could be causing this.
I’m too lazy to look into it right now, but at the very least there’s a scientific mystery here. Whether or not the explanation turns out to be interesting remains to be seen. There seems to be a big stigma against reporting UAP in the military, which some NASA officials think is hindering our understanding; with fewer recorded phenomena, it’s hard to figure out what’s going on.
I think people underestimate the degree to which hardware improvements enable software improvements. If you look at AlphaGo, the DeepMind team tried something like 17 different configurations during training runs before finally getting something to work. If each one of those had been twice as expensive, they might not have even conducted the experiment.
I do think it’s true that if we wait long enough, hardware restrictions will not be enough.
I know Genomic Prediction at least has used “Quality Adjusted Lifespan” in their past papers, so I think they’re used interchangeably.
Both they and Orchid provide the expected absolute lifetime risk of each disease in their reports, so parents can re-prioritize embryo implantation according to their preferences. Doing this methodically is pretty tricky though; every disease has a different age of onset distribution and a different impact on life expectancy and quality of life.
My hope is that one or both of them create more granular tools to help patients pick embryos according to their values and preferences. But so far the best you can do is just looking at the raw numbers and googling stuff about average age of onset etc.
Choosing a sexual partner IS an example of genomes being steered by parental choice.
In the process of this development, these networks assign themselves a physiological form gender; intersex people get a mix of attributes at this stage, but for most people, even for most trans people, this stage almost entirely selects one profile of sexual dimorphism; typically for people with XX chromosomes, this stage selects female, and for people with XY, this stage selects male. However, it’s well known to science and can be looked up that sometimes people can be apparently entirely one body-form and have no desire or urge to transition, and yet have opposite chromosomes from their body’s layout-presentation.
This sounds like you’re talking about the SRY cascade without explicitly naming it. But that process starts at 6-8 weeks. Embryos are screened after 3-7 days of development. There’s no way to see what is happening with the SRY cascade at that stage.
Maybe there are some genes that influence the course it is likely to take? I understand your concern.
But I also think some of this tech could HELP acceptance of trans people. I suspect that many of the people who are anti-trans are at some level worried about its effects on family formation. Without freezing gametes, many trans people will not be able to have biological children after medically transitioning. Gametogenesis could give anyone the ability to have kids if they want them, regardless of their gender identity or transition status. I think that would at least do a little to quell the hate against transgender people (though I understand most of the people who get upset about transgender people are not very rational about their views).
Is tendency to wish to transition one’s gender genetic? I recall reading that the number of people self-identifying as trans has tripled in the last decade. That would suggest that the trait is probably not very heritable and would be hard to select against.
Parents have skin in the game when it comes to their children.
To me it seems unlikely that the nanotech necessary to do the massive somatic editing needed to replicate the effects of germline editing would not come far ahead of the tech necessary to just upload someone’s brain into a computer. And if you have that, why would you even bother with the huge limitations of biological intelligence?
I suppose it’s possible we might have some global ban on further improvements to computer intelligence, but in that case why wouldn’t you just genetically engineer people capable of solving alignment? Why hang around at the proteins & genes level of tech long enough to make massive somatic editing work?
I think the reason that people use the same term for both is that historically, the lines between enabling, encouraging, recommending, de facto enforcing, legally enforcing and plain violently enforcing genetic changes have been rather fluid. It never started with enforcing it, but it tended to end up there.
Was there even a way to do voluntary eugenics before embryo selection? I guess maybe paying people to have more kids or fewer kids might count. But the options were pretty limited.
In your own writing, you had somebody comment, saying it should actually be a moral obligation to perform embryo selection for all who had the financial means. Your response was not “that would be a different evil thing, no”. It was “I’m not quite sure I would agree with this yet, though I can see the case being made for it.”
There is a difference between a moral obligation and a legal one. I think people have a moral obligation to donate money to cost-effective charities, but I also support their right not to do that.
The difference is understanding there ought to be a limit on how the state can be used to enforce other people to comply with your moral standards.
Also, as I already said I don’t think this technology is yet cost-effective enough to warrant such a moral obligation.
Yet that story ended in real life with kids with disabilities being gassed to death.
You seem to be implying that there is somehow a direct causal link between films arguing for a right to die and the holocaust. I don’t think that slope is as slippery as you portray it. There are plenty of people nowadays advocating for the right to die who don’t believe in killing disabled people.
And again, that’s not even what I’m arguing for. I’m merely arguing that parents should have the right to give their children better genetics so long as those changes have at least some benefit to the well-being of the child.
There are of course always going to be some traits that are on the borderline of what we might consider ethically permissible. For example, should deaf parents be allowed to purposefully have a deaf child? I lean towards yes, particularly if the child can choose to reverse the condition with some kind of implant or medical procedure later in life.
If you want a different word for this, please also state how you intent to ensure, in the long run, that the freedom not to use this tech is maintained. Not just as a theoretical legal right, but in practice. That this won’t end with us standing in a classroom, and the teacher saying, in disgust, I cannot believe this child was not deselected, so I wouldn’t have to deal with this neurodivergent mess.
To be frank, I have thought less about these kinds of issues because I am so worried about AI. In my mind, the greatest benefit of this technology is it might provide the human species with individuals capable of guiding us through the incredible technologically-induced upheavals we are likely to see over the next century. My greatest fear is that we simply are not going to have enough time for these children to grow up. Next to that, worries about discrimination in the classroom or workplace have seemed relatively minor.
Still, it deserves to be addressed. I don’t think you’re going to see the effect you’ve described without something stronger than simple embryo selection. By itself, embryo selection maxes out at about 1 standard deviation of gain on IQ. That will have a very noticeable impact, but given you’ll be selecting on other traits besides just IQ, children born with its benefits will just appear to be unusually talented. The difference won’t be big enough to mark them as like fundamentally different.
Even if you get in-vitro gametogenesis working, you’d max out at maybe 22 IQ points.
Of course the goal is to eventually get iterated CRISPR or chromosome selection or some other advanced techniques to work that would be truly transformative. But it will take quite some time for this tech to be scaled out to the point where selected children become the majority of new births. I would guess at least 40 years and possibly longer.
The only way I can see that happening is if we manage to put a global moratorium put in place on AI research and the improvement of computer hardware. But if we did manage to coordinate on such a wise proposal, I suspect it would be lifted when we figure out how to make an aligned ASI that acts in the bests interests of humans in general. It seems likely that we will solve that with a bunch of genetically enhanced geniuses around to work on the problem.
For the short window during which this might be a problem though, there are a couple of ways to deal with it:
Have different schools for kids of roughly similar abilities. This is already standard practice in many countries such as Vietnam.
Introduce universal basic income to ensure that no one ends up truly destitute due to genetic predispositions that are no fault of their own
Once embryo selection or whatever technique we’re talking about is cost-effective enough, subsidize access with government funds so that parents will never be denied access due to financial constraints. The same logic that compels us to fund public schools would also compel us to subsidize access to this tech if it is cost-effective enough; society as a whole has a strong interest in ensuring the next generation is healthy and productive.
I’m a little too tired to read through the study you sent me. I suppose it’s possible she either got alcohol wrong or that new studies have come out since the book was published which should change the conclusions.
I would note that it’s very hard to assess the impacts of drinking in the current environment because women who adhere to health advice about not drinking are probably going to pass on genes to their children that make them less prone to the very neurodevelopmental disorders the studies purport to measure.
You could do a study where you find women who you think are at high risk of drinking during pregnancy and then perform an intervention to encourage some of them not to drink. But given that ERBs adhere to a strict Copenhagen interpretation of ethics I doubt such a study design would ever be allowed (even though conducting it would improve children’s health if alcohol does indeed have a negative impact).
But there is a slippery slope towards a scenario where people select for sex, skin (and hair and eye) colour, not being queer, not being neurodivergent… and I do find that a dystopian scenario where we would lose something valuable that enriches our world.
Yes, this is actually a fairly common critique of embryo selection. One useful intuition pump I’ve found helps me think about it is the reversal test; should we make people sicker or more mentally distraught to enrich the world? It’s a bit odd to imagine that evolution somehow put us at the perfect equilibrium where any increment or decrement in mental illness rates would result in a worse society. It’s especially odd to think that since evolution doesn’t care at all about either of those things except insofar as they affect inclusive reproductive fitness.
Also, my experience so far just talking to people makes me think parents are going to have different priorities regarding the traits they select for.
But I assure you, if my parents had gotten to pick, they would also have made sure that I am not gay. Or tall. Or enby. Or have ADHD and autism. Things that were challenges, to be sure, but that I do not see as a net negative. I also know that both my parents found it disagreeable that I am highly gifted, and smarter than them; my father explicitly considers it a disease state and unpleasant complication, and would certainly have selected against it. I am very glad that he could not.
I’m sorry about your parents. That sounds like an unpleasant experience.
I don’t think the thought experiment of “erasing” someone like you from existence is really a very good test of the morality of embryo selection. You are a person with decades of memories and ties to the community of people around you. In my view the morality of “erasing” you feels a lot different than making a choice between two embryos. Unless you believe in souls or something, an embryo is almost pure genetic potential. It has no internal organs, let alone a brain. Even the placenta hasn’t formed yet.
We also already have a scenario where it is incredibly difficult for a poor person to have proper upwards social mobility. If their competitors are literally superior from birth, prior to also getting their personal tutoring and private schools and trust funds and brain implants… at some point, no amount of hard work will make you competitive anymore, the different classes will become insurmountably separated and fixed. Not because the rich parents want this—they just want the best for their children, who wouldn’t. But if access to this tech is not fair, and its legitimate usage is not carefully reflected and set, the changes could be dystopian indeed.
I would point out that all the dynamics you described are already true to some degree; there are some people born with such extreme genetic disadvantages (through a combination of parentage and bad luck) that there are some paths in life simply closed to them.
Of course embryo selection will increase variance, so your point is still well taken. I’ve spent a fair amount of time thinking about this and the obvious solution here is just to work very hard to make this technology cheaper and better. If we make enough progress on that front then we can just have the government subsidize the technology and give free access to anyone that wants it.
Inequality WILL still increase in the meantime, but there are some dynamics that I think help us here:
There is a ceiling on improvements through embryo selection or editing. That limit is determined by the amount of variance in the human gene pool. The ceiling is very high, but its existence makes it plausible that some people will get there first and others will catch up
To push beyond that ceiling you’ll either need to generate new genetic variants and test them out in people. This will require the cooperation of a very large number of people. To make good predictors today, you need literally a million people or more in a database. To a first approximation I would guess you’ll need that many if you want to test out a ton of new genetic variants and have enough statistical power to distinguish true positives from false positives.
The main way rich people can get an advantage in embryo selection is by harvesting more embryos or getting access to better predictors. The predictors are all made from huge databases, most of which are public. So it’s unlikely rich people could maintain a monopoly on the best predictors. Also, it’s hard for them to get a big advantage by selecting from a lot more embryos. You can of course pay to use a better clinic, and you can pay to go through more egg retrievals to harvest more embryos. But there are steeply diminishing returns; you’re still sampling from a normal distribution. The expected maximum value of N samples from a normal distribution is sqrt(ln(N)). That’s an INCREDIBLY slow growing function. If you go from 10 embryos to 1000, the benefit only increases by 70%.
Will rich people still have an advantage? Yes. But genetic enhancement does not have the same runaway “intelligence explosion” dynamics that AI does.
I think this is probably true at the moment given the current efficacy and the prices of IVF.
The only reason I hesitate is because I think the tendencies and aptitudes of people at the top have a gigantic impact on the rest of society and that improving those tendencies would have a very large impact on the world (especially if parents of children likely to end up in positions of power select for traits like kindness and pro-social tendencies).
But if you ignore that for a moment, then you’re of course correct that embryo selection can come nowhere close to the efficacy of distributing bed nets or supplementing iodine-deficient populations.
However, it’s going to take some time to bring down the price of embryo selection and future technologies for genetic improvement, so I think it’s probably good to start on that now.
Granted, this is all kind of ignoring the 800 pound gorilla in the room which is AI. For embryo selection to really matter, there needs to be time for these children to grow up. If we make ASI in the next 20 years, there’s probably not much point. I’m still hopeful that we can get a pause in place on AI development. But without one this whole endeavor has pretty low odds of achieving anything IMO.
Yeah. Wait until they’re grown, then do runtime modification. Stem cells define genomes, genomes plus lifetimes define people, people define consent.
If I thought this was actually feasible with anywhere near our current level of technology then perhaps I would agree. But it is simply impossible without god-like biotech. Half the genes that show up in intelligence GWAS (or any other trait for that matter) are only active during development, particularly during early development in the womb and early childhood. The other half affect cells that literally last a lifetime (namely neurons).
Modifying one gene that is active in one particular stem cell population can currently be done for like cancer immunotherapy, but it costs like $500k and you have to radiate the person to kill their non-modified stem cells.
If we’re going to assume that someone creates aligned ASI I think this whole conversation is somewhat moot. If by some miracle we manage to solve the alignment problem before the mad scientists at OpenAI create Cthulhu, one of the first things people are going to use it for is to upload their own brains to a computer. Like why would you stick with spongy meat if you can just run a digital copy of yourself a million times faster?
I think there’s a pretty low chance that ASI goes well. Even if we manage to align the interests of ASI with its creators, whoever gets control of it is going to rule the world. This fantasy dream where everyone benefits equally seems very unrealistic to me. It’s way more likely that we have one or perhaps a few quasi-omnicient immortal dictators.
In other words, this technology as it is now should not be used, and we should wait for the drastically much more difficult version that allows editing genes at runtime before we deploy it at scale. In the meantime, I strongly support the coalition against doing offspring editing.
I don’t follow this argument. You’re saying we shouldn’t edit the germline or select embryos because the child has not provided consent?
If so we are prioritizing something that’s impossible to get (consent from a ball of stem cells) over the expected future wellbeing of a person.
The argument that parents need to get consent from their children before doing things that the parents believe are in the interest of the child also undermines the basic rights of parents. It is very common for adults to do things children don’t want or don’t like because they believe it is in the best interest of the child.
Undoubtedly some of those things are in fact not good for the child and are only done in the name of serving the child’s best interest when in fact they serve some needs of the adult. We have rules around what parents are allowed to do to their children that reduce some of the worst abuses of this power. But the general principle of adult control over children seems pretty important to how society functions.
If this service can’t be made free, it should be destroyed on deployment, to prevent moloch.
How many of the important technological advancements that have been made in the last 10,000 years could hold up to this standard? If every technology had to be free and universally available on launch, we would have no technology. There would not be medicine, nor manufacturing, nor clothing, nor cooked food. This is an untenable standard.
If a couple chose to pick the embryo they expected to suffer the most in life, is it epilogenetics?
If someone asks me I would of course say no because I would like the term to encompass traits that we can reasonably expect the child in question to be glad we gave them.
But I would note that even if use of the term triggers discussion of the question you posed, I still think that is shifting the conversation in the right direction; away from “is embryo selection bad because it reminds me of this other thing I don’t like” and towards “should parents be allowed to select for traits that their children probably wouldn’t like?”
Interesting viewpoint. I think your point about the morality of having children despite the high natural miscarriage rate is a good one.
My basic view is that human moral value develops throughout pregnancy (and indeed continues to develop after birth). I don’t think there’s a simple binary switch from “no value” to “value”. I’d treat it more like a gradual ramp-up beginning with brain development during pregnancy.
I’m curious how you feel about culturing of naive embryonic stem cells. It’s possible to culture cells from a very early embryo and maintain their epigenomic state. One might then perform some editing on each, then grow each into a colony of perhaps 100 cells before destructively sequencing some of the stem cells and then performing subsequent edits on the stem cells in which the edits successfully took place.
If done correctly, the process would result in an embryo with much better prospects for a healthy and happy life. One embryo goes in and one embryo goes out. But the sequencing in the interim steps would require the destruction of naive embryonic stem cells.
Would you consider such a process morally permissible?
Wow. Ok, I guess my odds that this is actually an alien spacecraft went up a little bit.
It’s interesting that at the end they quote a NASA official who stated that they haven’t found any evidence of extraterrestrial life yet, directly contradicting the whistleblower. That means either the evidence the whistleblower has isn’t sufficient to convince the scientists at NASA or the DOD isn’t sharing it with them.