Maybe this is where "AI" will come from, particularly military AI. If the genome were synthesized, then some people could convince themselves that what grew wasn't human and could therefore be shaped and enslaved.
Even if the technique for synthesizing a human genome was available literally tomorrow, it would still take a few decades for the first genetically synthesized human to be born, to grow up and to get an education before it becomes useful to anything.
That would leave lots of time for non-biological AI to catch up.
Interesting ethical issues arise when talking about "writing genomes" and essentially creating a designer human synthetically. Now let's say the scientific community decides against doing this based on ethical reasons. What's to stop rogue scientists from going ahead with the research and going around the global community? Would they need the protection of a nation state to move forward? What could be done to stop them? We're getting into cloners from Kamino territory here...
* One thing that can stop them is the law; it could be made illegal, which does stop a lot of behavior. International treaties often work too.
* Also, if it's considered out-of-bounds by the scientific community then it could ruin the scientists' career; that's probably not a risk that a talented scientist with a promising career would take. They couldn't even publish a paper, publish in arXiv, get feedback on their work, or participate in any online discussion. They couldn't go to a conference (there wouldn't be one in their field) or even tell anyone what they are working on. What a life to choose.
* Significant progress likely requires more resources than could be mustered covertly, including financial, technological (where do you buy equipment and supplies?), and especially labor: Not only do they need talented PhD students and postdocs, but progress in most fields requires a whole community of scientists who build on each other's work, not one who discovers it all themself.
The technology advances when it can advance, no matter what. The raise of surveillance would have happened even without NSA. The Deep Learning advancements happened because of faster computers, even though humans have discovered some of the best ideas 20+ years ago.
<rant>
There is potential for achieving superhuman intelligence by genetic selection. By scanning millions of people's DNA and their personal traits (predicting phenotype characteristics from genotype), they want to build a classifier that would rate a DNA before it got developed. Then they would be able to select embryos before IVF. This is done today too, but only to filter out some simple genetic defects.
A selection of 1:2 gives a 5 IQ point boost. A selection of 1:1000 embryos gives a boost of 40 IQ points. If we repeat this over a few generations, humanity will evolve much faster.
This iterative improvement by AI rating of genomes could be sped up by creating sperm and ova directly from embryo stem cells. This way we don't need to wait for a generation to pass to have the kids become parents. In a few years many DNA selection -> embryo iterations could be tested and selected and evolution would be sped up even more.
In the short run, 40 to 80 years, we face the appearance of AGI and human-computer interfaces. We will evolve together with it, and this kind of DNA synthesis would be an essential tool for the future. Even if a few people have been enhanced, when they have kids with non-enhanced people the new genetic advancements could be passed to the whole population in a few generations. The final aim is to extend life, intelligence and perhaps gain immortality and become much more than we were before.
Even if we don't do AI based DNA enhancement, we still need to prepare for the possible catastrophe at the point of singularity. We need to create databases of human embryos so we could create humanity back in case it is destroyed. I'd also put in these databases a copy of all the human produced data and culture, just in case they need to start from scratch. These databases would have to be hidden and remain incommunicado for long enough to make sure we didn't get wiped out. They could even be cared for by robots for millions of years before opening up, as an insurance ticket that we survive.
</rant>
My point is that we don't really control the evolution of these technologies. They will be developed because they can be developed. In the past they weren't developed because it was impossible. We, humans, created even the a-bomb immediately after we understood nuclear physics. We spied on people as soon as the internet (and even telephone) was available. Why would we be able to stop from making synthetic DNA? The temptation to prolong life or to enhance future children exists, the means also exist.
> The technology advances when it can advance, no matter what.
I know it's a popular theory but I don't know that it stands up to scrutiny. Nuclear weapons technology hasn't advanced in many years, for example; the few who could do the R&D don't seem interested. Technologies involving cannibis and LSD have not advanced in large part, AFAICT, because they are illegal. How do you overcome those and other obstacles I raised above? Another: Powerful interests freeze development in many industries.
Here's another way of thinking of it, more theoretically: There are an effectively infinite number of technologies that could be developed; only a few are chosen. Almost all tech is ignored.
I think that calling it inevitable is a way to duck difficult questions and obligations, similar to the way that some blame the free market and self-interest for their choices (as if those are gods who forced the mortal's hands).
That's a good point, but there are many reasons to think it will be much harder to contain genomics than atomics:
1. You can sense atomic weapons materials at a distance.
2. It takes significant capital equipment to make atomic weapons.
3. It takes precision equipment to make atomic weapons.
4. While theoretical knowledge of atomic weapons is readily disseminated, it is easy to find and target experienced experts in the making of atomic weapons.
I can't find corresponding "choke points" that could control genomics, even on the scale of a synthetic human genome.
Genes don't encode all the necessary information for growing up. There are time windows for activation of high level functions. (Cats get blind if their time window for visual learning is not used). So you need to take into account the time window and the other interactions before trying to speed up intelligence.
> This iterative improvement by AI rating of genomes could be sped up by creating sperm and ova directly from embryo stem cells. This way we don't need to wait for a generation to pass to have the kids become parents.
You don't need to create sperm or ova to create a genome. That's adding layers of indirection. Bostrom/Sandberg should study biology instead of handwaving like that. It's much faster to forego some of those steps completely. They are so busy being excited about "you don't need to wait an entire generation" that they don't realize "you don't need a sperm nor ova to rate a genome".
You'd have to find a country that cares about blue-sky health research and not bioethics, which sounds tricky. Despite the hype, this sort of technology wouldn't have any useful immediate applications- even if we could make a human with a complete custom genome, we don't have a good idea a what genes to give him/her in order to produce useful qualities like intelligence, strength, etc.
While the US debates such efforts in fear what could happen, other nations who don't see it as a problem will do this. It has always been a matter of time.
This foolish move could seriously damage legitimate genetics research and applications, and other science, in the public eye, affecting funding and even legality. The reality may be different than first glance, but we live in a world of first impressions and the people invovled should know that.
The participants must have some sense of where the lines are and also respect that the public has a right to make decisions for their own society.
I'm just curious, why is it terrifying? Would a super intelligent human being produced by science be any more terrifying then one that occurred naturally? Couldn't we be creating a being that could push us forward collectively? Think about the contribution to modern life Einstein, John Von Neumann, and Bohr had. If we had a way to make very intelligent designer humans, there could be a chance they could push us collectively forward in a similar way. I imagine that an Einstein born now, with access to the whole worlds information could be much more impactful (or maybe he would just succumb to liking cat videos.) All humans designer or not can be terrifying, beautiful, impactful, and other worldly. You don't have to be unnatural to be an evil bastard, and maybe an engineered human could be like Gandhi. Don't prejudge, just because its unnatural doesn't mean it is not good.
Just like GMO doesn't mean bad (I eat organic food by the way), a lot of this goes to intent. If science wants to push us forward and make us better. I'm all for it. If they want to engineer us to be resistant to industrial chemicals or pollution so that companies can pollute more, or make super soldiers that is when you have to be terrified . Not just because we have something artificial.
I don't know. A lot of people agree with you, that creating superintelligent humans is necessary.
I'm scared of being replaced. I'm also worried such humans would be so different than us, they wouldn't identify with us. Maybe even come to despise us. They would form their own ingroup and view us as enemies even.
I've met someone that had high IQ, and identified with high IQ like some people identify with their country. He legitimately hated low IQ people, was ok with exterminating them, and said he would prefer to die or commit suicide than lose his intelligence. It was such an alien way of looking at the world to me.
>I'm scared of being replaced. I'm also worried such humans would be so different than us, they wouldn't identify with us. Maybe even come to despise us. They would form their own ingroup and view us as enemies even.
You're on Hacker News and you're afraid of nerds? Come on. We're right here. Ask us what we think of you.
>I've met someone that had high IQ, and identified with high IQ like some people identify with their country. He legitimately hated low IQ people, was ok with exterminating them, and said he would prefer to die or commit suicide than lose his intelligence. It was such an alien way of looking at the world to me.
To be honest, he probably went through a lot of pain and maltreatment to acquire that kind of point-of-view. He's not an alien! He's a person. Just because his cognition has less noise in it and he needs fewer samples to reach the same inferences doesn't make him some fundamentally different kind of creature: all the algorithms are the same, only their free parameters are different.
Smart people have empathy and feelings too, and in fact, most of the time the most genuinely hurtful thing we experience (speaking as someone who is too often told that he's the "smart guy" in the group, even among nerdy, intelligent people) is precisely the Straw Vulcan stereotype about intelligence. We're not cold-hearted, we're not aliens, we're just people. Putting us in gifted programs as kids is a great idea. Forcing us to compose prodigal concertos at age 12 is probably a bad idea. Treating us with respect and fellowship works best.
No I'm not "afraid of nerds". I am a nerd. These wouldn't be nerds, they would be to Einstein, what Einstein was to an average person. The distance would probably be far greater actually. They could plausibly be a different species than us.
What happened to this guy I was talking about, was high IQ defined his identity, what separated him from others. Fortunately most high IQ people do integrate fine into society, we all still have a lot on common.
But it's likely the superintelligent babies would not integrate well. They would only be able to communicate and find things in common with other supers, and not idenitfy with us normies. We'd be subhumans to them.
I find it terrifying because regardless of their IQ in their nature they would still be humans that are shaped as much by their environment as any other toddler child tween teen adult.
The day we're able to make batman a reality someone will also make The Joker a reality.
Unless they would immediately build a spaceship and leave for a better place, it is obvious they would look at us as under-race that is heavily polluting and destroying the same planet they are force to inhabit.
While some might think, "oh well they would just guide us on how to treat Earth better and we all gonna live happily ever after", I would rather bet there would be no good reason for such investment on their part; you don't invest time and energy into training monkeys not to leave the Zoo perimeter; you just put high fences and plug them to electricity.
Honestly, and selfishly, the better race of human is just that - a better version of humans. Sadly I personally would accept their superiority.
don't we have a moral obligation to improve our children as much as possible? the universe is tough enough without holding ourselves back from something that may be possible.
Perhaps, and perhaps our children will use "Homo Sapiens" to describe a sub-human too stupid to live, the way we do "Neanderthal", once there's nothing left of us but bones.
"By some estimates" is reference to the logic that one professor scored humans on a log scale 1-5 and Einstein was a 0.5, therefore some human could be a 0.1
By that same logic, there are 0.0001 or 10000. Its completely baseless.
What are you talking about? It's based on IQ, which is a statistical measure of intelligence, normalized to fit a normal distribution. Every 15 IQ points represents exactly 1 standard deviation. Meaning more intelligent than some percentage of humans alive.
So a 1,000 IQ would mean that the probability someone had been naturally born before of equal or greater intelligence would be more than 1 in a trillion trillion trillion...
This is possible because genes are mixed up randomly every generation. The probability of getting the exact optimal combination of genes, out of a thousand genes which affect intelligence, is like flipping a coin and getting heads a thousand times in a row. It will never happen unless you intelligently select every gene artificially, intentionally place every coin down on heads.
That's only because existing IQ tests aren't normalized past 160. IQ is defined by a statistical distribution, and that requires sampling the general population. People with more than 160 IQ are so incredibly rare, there's no samples to define that end of the distribution.
But that's irrelevant to this. Because we are dealing with the actual genes that determine intelligence, we can estimate the exact distribution of intelligence, and predict where a set of genes would fall on that distribution. See my comment below.
The person that would be the result of such modifications would most likely have serious psychological issues and probably other problems that we cant even imagine.
We are like kids playing with loaded machine guns, currently we are working on removing that fiddly thing called a safety.
And we're like how hard can i be, our parent mother nature does it all the time.
Once we are ready to do this, there would need to be many studies done to correlate genetics with diseases and abilities. We would know exactly what set of genes causes various mental diseases and could avoid them. High IQ people do not generally have serious psychological issues, and are on average better than the general population. But yes it's a possibility.
People are already different in terms of capabilities. Yet they are equals in terms of rights, and there is no reason that should change, even with this kind of things.
That change of equality is more or less inevitable. Imagine a line of humans as far ahead of us as we are of dogs. They're certainly going to have their own ideas on how to use the resources of the world and will almost certainly think they know what is best for the rest of us. (Hell, there are plenty of modern day humans who already think that about their peers.)
As you said that's a mentality that exists already among many people. It's more of a political issue than a technological one.
I personally believe that smart people don't feel any need to get privileges of any kind, since they are by definition more capable to attain their goals anyway. You may not give a dog lots of rights, but you don't see it as a threat or a competitor either.
> They're certainly going to have their own ideas on how to use the resources of the world and will almost certainly think they know what is best for the rest of us.
And nothing will stop them from pursuing political and business careers at equal terms with the rest of us. Of course, their superior genes will let them get ahead, but regardless, that's what already is happening - more genetically smart and charismatic people can get into positions of power easier than others, and we seem to be quite OK with it.
1 standard deviation difference in group distributions of intelligence already seems to result in vast disparities in power, control of resources, etc.
I shudder to think how a greatly magnified version of that will turn out.
Animals and plants physiology is still capable of doing things we're not nearly capable of reproducing in laboratory. Plants and insects production are still used in pharmacology so it seems to me there would be a huge commercial potential for plant and animal organisms made from scratch, just as there is currently for organisms just barely modified.
The human case will always be made difficult by ethicists and other party poopers, anyway.
I beheld the wretch — the miserable monster whom I had created.
He held up the curtain of the bed; and his eyes, if eyes they may be
called, were fixed on me. His jaws opened, and he muttered some
inarticulate sounds, while a grin wrinkled his cheeks. He might have
spoken, but I did not hear; one hand was stretched out, seemingly to
detain me, but I escaped and rushed downstairs. I took refuge in the
courtyard belonging to the house which I inhabited, where I remained
during the rest of the night, walking up and down in the greatest
agitation, listening attentively, catching and fearing each sound as
if it were to announce the approach of the demoniacal corpse to which
I had so miserably given life.
Of course, Frankenstein's creation wasn't the monster in that story, at least not at first. The monster was the human too wound up in his own bigotry and religiosity to accept the consequences of science on its own terms, Frankenstein himself.
> “Our ability to understand what to build is so far behind what we can build,’’ said Dr. Minshull, who was invited to the meeting at Harvard but did not attend. “I just don’t think that being able to make more and more and more and cheaper and cheaper and cheaper is going to get us the understanding we need.’’
As point of comparison, consider how well any of us would understand software engineering if it cost thousands of dollars to compile a program. That is the potential impact of this technology- talk about using this to create super-humans or whatever is pure hype and/or scaremongering.
Humans/mammals have 50 times more coding DNA than simple fungi they've syntesized so far. Plus natural mammal DNA is usually subdivided into multiple pieces on a chromosome called introns. It is unclear how important the non-coding inbetween DNA is.
Homever nearly the entire lab mouse genome has been purified into single pieces called coding DNA and can be purchased from labs. Maybe they'll try synthezing a mouse first. The human case shouldnt be more difficult.
49 comments
[ 3.0 ms ] story [ 84.5 ms ] threadThat would leave lots of time for non-biological AI to catch up.
* One thing that can stop them is the law; it could be made illegal, which does stop a lot of behavior. International treaties often work too.
* Also, if it's considered out-of-bounds by the scientific community then it could ruin the scientists' career; that's probably not a risk that a talented scientist with a promising career would take. They couldn't even publish a paper, publish in arXiv, get feedback on their work, or participate in any online discussion. They couldn't go to a conference (there wouldn't be one in their field) or even tell anyone what they are working on. What a life to choose.
* Significant progress likely requires more resources than could be mustered covertly, including financial, technological (where do you buy equipment and supplies?), and especially labor: Not only do they need talented PhD students and postdocs, but progress in most fields requires a whole community of scientists who build on each other's work, not one who discovers it all themself.
<rant>
There is potential for achieving superhuman intelligence by genetic selection. By scanning millions of people's DNA and their personal traits (predicting phenotype characteristics from genotype), they want to build a classifier that would rate a DNA before it got developed. Then they would be able to select embryos before IVF. This is done today too, but only to filter out some simple genetic defects.
A selection of 1:2 gives a 5 IQ point boost. A selection of 1:1000 embryos gives a boost of 40 IQ points. If we repeat this over a few generations, humanity will evolve much faster.
This iterative improvement by AI rating of genomes could be sped up by creating sperm and ova directly from embryo stem cells. This way we don't need to wait for a generation to pass to have the kids become parents. In a few years many DNA selection -> embryo iterations could be tested and selected and evolution would be sped up even more.
In the short run, 40 to 80 years, we face the appearance of AGI and human-computer interfaces. We will evolve together with it, and this kind of DNA synthesis would be an essential tool for the future. Even if a few people have been enhanced, when they have kids with non-enhanced people the new genetic advancements could be passed to the whole population in a few generations. The final aim is to extend life, intelligence and perhaps gain immortality and become much more than we were before.
Even if we don't do AI based DNA enhancement, we still need to prepare for the possible catastrophe at the point of singularity. We need to create databases of human embryos so we could create humanity back in case it is destroyed. I'd also put in these databases a copy of all the human produced data and culture, just in case they need to start from scratch. These databases would have to be hidden and remain incommunicado for long enough to make sure we didn't get wiped out. They could even be cared for by robots for millions of years before opening up, as an insurance ticket that we survive.
</rant>
My point is that we don't really control the evolution of these technologies. They will be developed because they can be developed. In the past they weren't developed because it was impossible. We, humans, created even the a-bomb immediately after we understood nuclear physics. We spied on people as soon as the internet (and even telephone) was available. Why would we be able to stop from making synthetic DNA? The temptation to prolong life or to enhance future children exists, the means also exist.
I know it's a popular theory but I don't know that it stands up to scrutiny. Nuclear weapons technology hasn't advanced in many years, for example; the few who could do the R&D don't seem interested. Technologies involving cannibis and LSD have not advanced in large part, AFAICT, because they are illegal. How do you overcome those and other obstacles I raised above? Another: Powerful interests freeze development in many industries.
Here's another way of thinking of it, more theoretically: There are an effectively infinite number of technologies that could be developed; only a few are chosen. Almost all tech is ignored.
I think that calling it inevitable is a way to duck difficult questions and obligations, similar to the way that some blame the free market and self-interest for their choices (as if those are gods who forced the mortal's hands).
1. You can sense atomic weapons materials at a distance.
2. It takes significant capital equipment to make atomic weapons.
3. It takes precision equipment to make atomic weapons.
4. While theoretical knowledge of atomic weapons is readily disseminated, it is easy to find and target experienced experts in the making of atomic weapons.
I can't find corresponding "choke points" that could control genomics, even on the scale of a synthetic human genome.
You don't need to create sperm or ova to create a genome. That's adding layers of indirection. Bostrom/Sandberg should study biology instead of handwaving like that. It's much faster to forego some of those steps completely. They are so busy being excited about "you don't need to wait an entire generation" that they don't realize "you don't need a sperm nor ova to rate a genome".
You'd have to find a country that cares about blue-sky health research and not bioethics, which sounds tricky. Despite the hype, this sort of technology wouldn't have any useful immediate applications- even if we could make a human with a complete custom genome, we don't have a good idea a what genes to give him/her in order to produce useful qualities like intelligence, strength, etc.
The participants must have some sense of where the lines are and also respect that the public has a right to make decisions for their own society.
[1] http://www.imdb.com/title/tt0090305/
Truly a terrifying possibility.
Just like GMO doesn't mean bad (I eat organic food by the way), a lot of this goes to intent. If science wants to push us forward and make us better. I'm all for it. If they want to engineer us to be resistant to industrial chemicals or pollution so that companies can pollute more, or make super soldiers that is when you have to be terrified . Not just because we have something artificial.
I'm scared of being replaced. I'm also worried such humans would be so different than us, they wouldn't identify with us. Maybe even come to despise us. They would form their own ingroup and view us as enemies even.
I've met someone that had high IQ, and identified with high IQ like some people identify with their country. He legitimately hated low IQ people, was ok with exterminating them, and said he would prefer to die or commit suicide than lose his intelligence. It was such an alien way of looking at the world to me.
You're on Hacker News and you're afraid of nerds? Come on. We're right here. Ask us what we think of you.
>I've met someone that had high IQ, and identified with high IQ like some people identify with their country. He legitimately hated low IQ people, was ok with exterminating them, and said he would prefer to die or commit suicide than lose his intelligence. It was such an alien way of looking at the world to me.
To be honest, he probably went through a lot of pain and maltreatment to acquire that kind of point-of-view. He's not an alien! He's a person. Just because his cognition has less noise in it and he needs fewer samples to reach the same inferences doesn't make him some fundamentally different kind of creature: all the algorithms are the same, only their free parameters are different.
Smart people have empathy and feelings too, and in fact, most of the time the most genuinely hurtful thing we experience (speaking as someone who is too often told that he's the "smart guy" in the group, even among nerdy, intelligent people) is precisely the Straw Vulcan stereotype about intelligence. We're not cold-hearted, we're not aliens, we're just people. Putting us in gifted programs as kids is a great idea. Forcing us to compose prodigal concertos at age 12 is probably a bad idea. Treating us with respect and fellowship works best.
What happened to this guy I was talking about, was high IQ defined his identity, what separated him from others. Fortunately most high IQ people do integrate fine into society, we all still have a lot on common.
But it's likely the superintelligent babies would not integrate well. They would only be able to communicate and find things in common with other supers, and not idenitfy with us normies. We'd be subhumans to them.
The day we're able to make batman a reality someone will also make The Joker a reality.
While some might think, "oh well they would just guide us on how to treat Earth better and we all gonna live happily ever after", I would rather bet there would be no good reason for such investment on their part; you don't invest time and energy into training monkeys not to leave the Zoo perimeter; you just put high fences and plug them to electricity.
Honestly, and selfishly, the better race of human is just that - a better version of humans. Sadly I personally would accept their superiority.
It's not as if you'd have a choice in the matter, any more than neanderthals had a choice when we came along.
By that same logic, there are 0.0001 or 10000. Its completely baseless.
So a 1,000 IQ would mean that the probability someone had been naturally born before of equal or greater intelligence would be more than 1 in a trillion trillion trillion...
This is possible because genes are mixed up randomly every generation. The probability of getting the exact optimal combination of genes, out of a thousand genes which affect intelligence, is like flipping a coin and getting heads a thousand times in a row. It will never happen unless you intelligently select every gene artificially, intentionally place every coin down on heads.
But that's irrelevant to this. Because we are dealing with the actual genes that determine intelligence, we can estimate the exact distribution of intelligence, and predict where a set of genes would fall on that distribution. See my comment below.
We are like kids playing with loaded machine guns, currently we are working on removing that fiddly thing called a safety.
And we're like how hard can i be, our parent mother nature does it all the time.
I personally believe that smart people don't feel any need to get privileges of any kind, since they are by definition more capable to attain their goals anyway. You may not give a dog lots of rights, but you don't see it as a threat or a competitor either.
And nothing will stop them from pursuing political and business careers at equal terms with the rest of us. Of course, their superior genes will let them get ahead, but regardless, that's what already is happening - more genetically smart and charismatic people can get into positions of power easier than others, and we seem to be quite OK with it.
I shudder to think how a greatly magnified version of that will turn out.
Animals and plants physiology is still capable of doing things we're not nearly capable of reproducing in laboratory. Plants and insects production are still used in pharmacology so it seems to me there would be a huge commercial potential for plant and animal organisms made from scratch, just as there is currently for organisms just barely modified.
The human case will always be made difficult by ethicists and other party poopers, anyway.
As point of comparison, consider how well any of us would understand software engineering if it cost thousands of dollars to compile a program. That is the potential impact of this technology- talk about using this to create super-humans or whatever is pure hype and/or scaremongering.