I get unnerved reading speculation about a Gattaca-like future. The biggest issue is that this technology will likely be much more accessible to the wealthier strata among us, allowing the gap between richer and poorer families to magnify significantly.
The problem with technology like this that tends towards eugenic aims is that there’s no certainty that those who choose tactically advantageous traits (height, intelligence, attractiveness) for their children will also choose eusocial traits (humility, honesty, kindness). There could be an arms race among amoral billionaires who could father leagues of superhuman, psychopathic children.
Even further, in all likelihood the prospect of chromosomal selection will be obsolete compared to direct modification of the embryo’s DNA, maximizing all possible traits with minimal relatedness to the parent. If your goal is perfect children, why care if they’re statistically your spawn by virtue of genetic relatedness? Or if you care so much, you could preserve similarity in superficial traits like certain neutral facial features but maximize every other aspect.
The whole “intelligent psychopaths” idea is an interesting one. I’ve spent time thinking about such a possibility.
I think there is a legitimate risk that if you had a high-trust, kind society with a lot of selfless people, such an environment would be ripe for exploitation by a smart psychopath.
But I’d be surprised if intelligent psychopaths have an advantage over intelligent non-psychopaths in today’s world. Reciprocal cooperative tendencies evolved in humans because groups of humans are extremely powerful, and cooperation is required to make such groups function.
Even further, in all likelihood the prospect of chromosomal selection will be obsolete compared to direct modification of the embryo’s DNA, maximizing all possible traits with minimal relatedness to the parent. If your goal is perfect children, why care if they’re statistically your spawn by virtue of genetic relatedness? Or if you care so much, you could preserve similarity in superficial traits like certain neutral facial features but maximize every other aspect.
I don’t think you can push traits more than 4-5 standard deviations in any direction without serious risks. I don’t have a good way to estimate the exact safe levels, but my guess is if you push more than 2 or 3 standard deviations out from the most extreme humans you’re going to find negative pleiotropy between the traits you’re selecting for and some other trait that are important but weren’t included in your index.
You can see examples of this in animal husbandry and agriculture. When Normal Borlaug made the first strain of hyper-productive dwarf wheat, the heads were so heavy with grains that the stalks bent and broke in the wind, causing large crop failures. He had to breed the wheat to reinforce the stalks before the crop became a viable alternative to existing varieties.
Likewise, I think if you literally just flipped every IQ-affecting allele into the position expected to maximize IQ, there would almost certainly be some unanticipated negative side-effects.
So even though the theoretical gains from genome synthesis or iterated CRISPR editing are larger, I don’t expect them to provide that much of a practical advantage due to the limitations with how far we can push traits.
Edward O. Wilson, in his book “The Meaning of Human Existence”, posits that there exists a natural tendency for humans to act selflessly when the in-group is endangered by an out-group, but selfishly when there is no danger, as a selfish action that negatively impact’s the group’s fitness would not harm his or her survival.
This implies that human societies have nursed a dual-cooperation strategy, which selects for traits that create fertile, productive, high trust societies to guard against threats, but creates in those societies fertile ground for exploitation by selfish behaviors, the genes coding for which are passed on more and thus become more common in successive generations.
Cooperation is powerful, but even the smartest, most cooperatively adept groups are susceptible to exploitation. There are things that non-psychopathic people just couldn’t stomach doing that psychopaths could easily do. The magnitude for material, reproductive and social success is too high for a highly productive, amoral agent to ignore the benefits. The only way for a cooperative society to guard against it is to be, by default, zero-trust.
Your point about negative pleiotropy makes sense and will likely be a huge hurdle for genetic modification that seeks specific metrics. However there are just too many low-risk low-hanging fruit in the human gene pool to make it necessary for genetic modification to venture into the more extreme territories to get the benefits it needs. There already exist natural born humans 5 standard deviations above any given average trait: the success of serendipitously fit humans by pure natural chance is enough of a basis for artificial traits to be easily pushed to a reasonably higher level. Winning the genetic lottery isn’t about hitting one giant bullseye, it’s about avoiding 1,000,000 micro bullets.
Also, regarding this common chain, my comment got a very strong immediate negative reaction, but I’m unsure what people’s main dismissal of it stems from. The belief that I’m veering towards false alarmism?
I think that a relative weakness of psychopaths is that they have to be rare in the population, because (a) an interaction between two psychopaths is likely detrimental for one or both of them, and (b) the psychopaths lose the element of surprise as people become more familiar with them.
So if we genetically engineer superintelligent children, some of them psychopathic, some of them not, I suppose the interactions between them will lead to some stable ratio.
People have adaptations against exploitation, such as trusting people you know for a long time or someone you trust knows them for a long time, or getting drunk together (it is difficult to fake non-existing emotions when drunk, and the next day people will remember weird behavior). In a high-trust society we often stop relying on them, but as the trust reduces, they can become more popular again.
I get unnerved reading speculation about a Gattaca-like future. The biggest issue is that this technology will likely be much more accessible to the wealthier strata among us, allowing the gap between richer and poorer families to magnify significantly.
The problem with technology like this that tends towards eugenic aims is that there’s no certainty that those who choose tactically advantageous traits (height, intelligence, attractiveness) for their children will also choose eusocial traits (humility, honesty, kindness). There could be an arms race among amoral billionaires who could father leagues of superhuman, psychopathic children.
Even further, in all likelihood the prospect of chromosomal selection will be obsolete compared to direct modification of the embryo’s DNA, maximizing all possible traits with minimal relatedness to the parent. If your goal is perfect children, why care if they’re statistically your spawn by virtue of genetic relatedness? Or if you care so much, you could preserve similarity in superficial traits like certain neutral facial features but maximize every other aspect.
The whole “intelligent psychopaths” idea is an interesting one. I’ve spent time thinking about such a possibility.
I think there is a legitimate risk that if you had a high-trust, kind society with a lot of selfless people, such an environment would be ripe for exploitation by a smart psychopath.
But I’d be surprised if intelligent psychopaths have an advantage over intelligent non-psychopaths in today’s world. Reciprocal cooperative tendencies evolved in humans because groups of humans are extremely powerful, and cooperation is required to make such groups function.
I don’t think you can push traits more than 4-5 standard deviations in any direction without serious risks. I don’t have a good way to estimate the exact safe levels, but my guess is if you push more than 2 or 3 standard deviations out from the most extreme humans you’re going to find negative pleiotropy between the traits you’re selecting for and some other trait that are important but weren’t included in your index.
You can see examples of this in animal husbandry and agriculture. When Normal Borlaug made the first strain of hyper-productive dwarf wheat, the heads were so heavy with grains that the stalks bent and broke in the wind, causing large crop failures. He had to breed the wheat to reinforce the stalks before the crop became a viable alternative to existing varieties.
Likewise, I think if you literally just flipped every IQ-affecting allele into the position expected to maximize IQ, there would almost certainly be some unanticipated negative side-effects.
So even though the theoretical gains from genome synthesis or iterated CRISPR editing are larger, I don’t expect them to provide that much of a practical advantage due to the limitations with how far we can push traits.
Edward O. Wilson, in his book “The Meaning of Human Existence”, posits that there exists a natural tendency for humans to act selflessly when the in-group is endangered by an out-group, but selfishly when there is no danger, as a selfish action that negatively impact’s the group’s fitness would not harm his or her survival.
This implies that human societies have nursed a dual-cooperation strategy, which selects for traits that create fertile, productive, high trust societies to guard against threats, but creates in those societies fertile ground for exploitation by selfish behaviors, the genes coding for which are passed on more and thus become more common in successive generations.
Cooperation is powerful, but even the smartest, most cooperatively adept groups are susceptible to exploitation. There are things that non-psychopathic people just couldn’t stomach doing that psychopaths could easily do. The magnitude for material, reproductive and social success is too high for a highly productive, amoral agent to ignore the benefits. The only way for a cooperative society to guard against it is to be, by default, zero-trust.
Your point about negative pleiotropy makes sense and will likely be a huge hurdle for genetic modification that seeks specific metrics. However there are just too many low-risk low-hanging fruit in the human gene pool to make it necessary for genetic modification to venture into the more extreme territories to get the benefits it needs. There already exist natural born humans 5 standard deviations above any given average trait: the success of serendipitously fit humans by pure natural chance is enough of a basis for artificial traits to be easily pushed to a reasonably higher level. Winning the genetic lottery isn’t about hitting one giant bullseye, it’s about avoiding 1,000,000 micro bullets.
Also, regarding this common chain, my comment got a very strong immediate negative reaction, but I’m unsure what people’s main dismissal of it stems from. The belief that I’m veering towards false alarmism?
I think that a relative weakness of psychopaths is that they have to be rare in the population, because (a) an interaction between two psychopaths is likely detrimental for one or both of them, and (b) the psychopaths lose the element of surprise as people become more familiar with them.
So if we genetically engineer superintelligent children, some of them psychopathic, some of them not, I suppose the interactions between them will lead to some stable ratio.
People have adaptations against exploitation, such as trusting people you know for a long time or someone you trust knows them for a long time, or getting drunk together (it is difficult to fake non-existing emotions when drunk, and the next day people will remember weird behavior). In a high-trust society we often stop relying on them, but as the trust reduces, they can become more popular again.