Hacker News new | past | comments | ask | show | jobs | submit login
Super-Intelligent Humans Are Coming (nautil.us)
56 points by dnetesn on Mar 3, 2016 | hide | past | web | favorite | 80 comments

I don't know if there is any statistical correlation or not, but (as the article alludes to) it does seem that super intelligent people are more prone to some kind of other deficiencies, and particularly social issues.

Can anyone shed more light on whether this is true or not?

Anecdotally, one of my best friends (extremely intelligent, Cum Laude engineering student) went "off the rails" a few years after graduating. I always had the feeling that he was seeing things 5 steps ahead of everyone else, and had such an extremely deep insight into every day things. I think it just became too much for him to handle.

Intelligence, in real life, is quite a nebulous thing. We can perhaps measure an IQ number, observe that someone is "bright", or remark upon their achievements, but while all these things are "correlated", it is a loose correlation with many exceptions.

I am not convinced that producing people with high IQ's and declaring them "intelligent super-humans" will result in anything dramatic, unless that is followed up with a lifetime of enrichment: superb education, good parenting, and formative life-experiences.

I do volunteer work in an adult literacy program and, learning disabilities aside, some of these folks are truly very intelligent and fully capable but have been set-back in childhood and young adulthood because of circumstances beyond their control and utterly unrelated to genetics.

If we really want to create super-intelligent beings, there's a "low tech" way to do that will be far more effective than genetic engineering: treat education as a mission-critical priority and make sure kids are taken care of.

> he was seeing things 5 steps ahead of everyone else, and had such an extremely deep insight into every day things

I had a period when I experienced this for some time. I can confirm it's deeply frustrating. You see all the madness very clearly. You see theoretical solutions to everything, know it could have been different. But you also understand you're quite alone and not influential enough. Generally, you're better off being average.

I would also add that you don't necessarily 'know' solutions, but rather make sense of the limited input you have and try to extrapolate in that. Obviously this fails hard at human interactions, which makes this the frustrating part. How to think of an alternative when the 'right' solution seems so beautifully simple?

The only health problems associated with higher intelligence (as opposed to either low intelligence or completely independent of intelligence) are [1] high functioning autism, bipolar disorder and schizophrenia. Two can be treated with pills when they are disruptive with relatively high success, and the other (Autism) is to me not a particularly net negative condition when high functioning (several immediately family members have it, so this is anecdotal). So the only byproduct of selecting for these genes, is it would compound the likelihood of one of these conditions. It sounds like your friend was in a period of disintegration [2]. Happens to most gifted people at some point early in life.

[1] http://www.telegraph.co.uk/news/science/science-news/1212254... [2]: https://en.wikipedia.org/wiki/Positive_Disintegration

"high functioning" kind of autism is a misnomer - even if you aren't the kind that bangs their head on the floor / avoided being non-verbal, life can be quite challenging.

I agree that it isn't necessarily a net negative, but the success of an individual on the spectrum seems highly conditional on life circumstances.

The thing is, genetics for autism also seem to produce people with adhd (reciently rolled into the spectrum) and generally coincide with decreased ability in executive function - leading to the absent minded professor syndrome. It also has a high prevalence of sensory processing disorder which can lead to a lot of issues with dealing with day-to-day life.

Even in cases where people with ASD happen to be of above average intelligence, it doesn't necessarily lead to success. I score 99% percentile in almost all academic areas, but I can't even feed myself right now. I'm functionally retarded when it comes to interacting with people half the time and the high-stress that comes from normal every day things means I might as well be agoraphobic.

Under the right conditions I have no doubt that I can flourish. I'm actively engaged in getting my life back on track (grew up in poverty, diagnosed late in life) but there's a reason that suicide rates and general unemployment levels for people on the spectrum are astronomical compared to the general population.

I agree with this. I would be ADHD if my parents sent me to school and/or believed in those diagnoses. I have yet to meet a gifted child male that does not qualify as ADHD. It is interesting to hear what you have mastered vs. struggled with. I can't say that my anecdotal evidence suggests that people always get a net benefit, just that in the right environment, people with autism can have complete, fulfilling and engaging lives. Unfortunately it seems to take a lot of will power and brute force intelligence to make it so, and not everyone is successful.

Note that a combined diagnosis of schizophrenia and biopolar, has very high rates of suicide (over 10% if left untreated).

This is not at all my observation. I've noticed that the uncommonly (technically-scientifically) intelligent people I've been exposed to tended to lead very socially balanced and normal lives. They have also been some of the funniest people I've met too.

In my experience too, sense of humor correlates with broad intelligence much more than social awkwardness does. On the other hand, I know extremely gifted people in some narrow area plus few related (i.e. STEM subjects) that cannot even lead normal life without somebody else taking care of them and are below average on different subjects (i.e. languages). I think the problem is, that we don't even know, what intelligence is and how many different forms it has.

That touches on the issue I had with the article as well.

> "But psychometric studies, which seek to measure the nature of intelligence, paint a different picture. Millions of observations have shown that essentially all “primitive” cognitive abilities—short and long term memory, the use of language, the use of quantities and numbers, the visualization of geometric relationships, pattern recognition, and so on—are positively correlated."

Whilst it might have been "proven" in the lab, that's not what is apparent with day to day experiences. If we split 'intelligence' into 'academic intelligence' and 'emotional intelligence', it's rare to find someone who excels with both. It's not just a matter of capacity, our abilities are a reflection of what we devote our energy to.

To explain why it's important to make the distinction between different types of intelligence... to put it simply, academic intelligence makes a difference in abstract thinking, developing new ideas and refining existing ones. However, emotional intelligence guides us on how best to spend our time. Emotional intelligence is a close relative to wisdom. We need it to guide us. If not, we'll end up doing things because we can, rather than whether we should, and with the power of super-intelligence that's something we'll need more than ever.

The article actually addresses the issue, saying that " observations have shown that essentially all “primitive” cognitive abilities—short and long term memory, the use of language, the use of quantities and numbers, the visualization of geometric relationships, pattern recognition, and so on—are positively correlated."

So, in a nutshell : on average, being super smart at math means you are probably verbally smart too.

I believe a lot of these social issues are due to lack of understanding between society vs. a "genius". But if 20%+ of the population are on the same wave length, they may not feel so alienated. An individual's "social issues" may not be a problem for 20%+ of the population, so it may not longer be an issue.

This is a good point, although if being intelligent is correlated with being on the autistic spectrum in any way, then it doesn't really matter how many people have the same level of intelligence/condition/personality. As someone who's on the spectrum myself, I can say that you could put ten or so people with high functionality autism or what not in a room together and they'd still find it difficult to socialise.

I think that one thing that can eventually hurt very smart people is that the world is much more chaotic and complex than even they can handle, which can lead to being overwhelmed/depressed/exasperated.

Our ancestors were not so benighted as we have been told: ignorance is, indeed, bliss.

Here are a few patterns I've seen in studies or surfaced:

1. Intelligence is an ability to observe, gather feedback based on past observations, and see patterns at a speed much faster than others at a lower general level. The growth in insight can be exponential, but it also means you start thinking at levels of abstraction and patterns that are layers above people at other levels. You struggle to communicate.

2. High Intelligence is not correlated with high moral or other views - at its base, it means speed in some directions an individual is prone to have. If they find a career path that allows them to speed in that direction, they thrive, if not, you see a lot of psychological challenges over time.

3. Socially, the higher your IQ, the more likely you are to skip steps in explaining your thoughts and to be misunderstood because your thinking might be several levels of abstraction above the people your're talking to. People at the 150IQ general level are rarely happy working with or friending people below 120. Given a propensity to specialize it a field rarely touched by others - one can get quite lonely in their thoughts and lose the social skills needed to develop normal friendships. Unless you are lucky enough to be surrounded by peers, you may be very unhappy. That unhappiness can shift to negative, downright trolling behaviors for some people.

4. Given the above, a lot of potential is lost. For every Einstein, there are likely several hiding in corners wondering what is wrong with them.

I've been thinking about and working on prototypes for a platform that would be helpful to unite people with niche and advanced interests, so they can collaborate freely. Would love to connect with others interested in the topic. (Edit: typos)

Sources: 1. On Intelligence: http://www.amazon.com/On-Intelligence-Jeff-Hawkins/dp/080507... 2. https://www.quora.com/Whats-it-like-to-have-a-150-IQ-Is-life... 3. Some article shared in the MENSA group years ago I'd have to dig in to find.

Sounds like a load of bullshit with no sources that plays on peoples' stereotypical view of a genius. Your 'platform' is just a forum, irc, facebook page, or literally any social network.

I don't agree with the premise of the article. The idea that there are "better" versions of each intelligence-influencing allele - and that they simply linearly combine - seems ludicrous to me.

It is equally plausible (and to my mind much more likely) that the many parameters influencing intelligence are linked in complex ways with no simple global optimum in sight.

I've read that many gifted children are utterly unfit for normal school. Some are even prone to killing themselves if not taken care of properly.

Assuming we can indeed select for gifted embryos, and do so, the consequences are going to be… interesting.

I'm a bit wary of "giftedness" in intelligence discussions...

I think the author's argument is very weak, analogous to:

"Scientists have identified thousands of parameters influencing the efficiency of petrol-powered cars. Each parameter seems to influence the outcome by a small amount. If we simply chose the best value for each parameter, I predict that we could see cars with efficiency in the range of 500 miles per gallon".

A related argument (I think the one you are making) is that generally optimising one thing is going to trade off other things. So going all-in on fuel efficiency might impact crash resistance, for example.

The author's argument for general intelligence might be true if the basic architecture of our brains has a lot of room for improvement. This might be true, but it seems that evolution should have already exploited any low-hanging fruit.

This whole thing reminds me of the "humans only use 10% of their brains" flavour of reasoning.

But what if normal school becomes what used to be a gifted school. if "gifted" is normalized, will these children still feel alienated and be "unfit" for school?

Screw genetic modification, I'd be interested to see what wacky deviations we see in a generation or so without it. Right now we've got assortative mating going on at a scale relatively unprecedented in history.

If I understand the wiki on assortative mating, it supports your statement.

However, in the last couple centuries there have also been greatly increased migrations globally, leading to increased mating across cultures and races. These events will be carried in the gene pool, maybe forever, and repeated. Intuitively, even if they're outliers, these matings will trend the gene pool towards more homogeneity.

Is this intuition correct? How do biologists reconcile the views?

Right now we've got assortative mating going on at a scale relatively unprecedented in history.

What exactly does this mean? Is there a place where I can learn more about it?

I recommend reading Coming Apart by Charles Murray, where assortative mating is discussed extensively.

There are also some interesting articles you might want to read: http://www.nytimes.com/2015/12/27/upshot/marriages-of-power-...



It means dating sites and college preferences are putting similar people together at an unprecedented rate:


Some of the assumptions in this article are very ambiguous and makes the reading quite disturbing. I don't grasp the author's notions of IQ. Maybe it's me misunderstanding the writing, but IQ is more of a normative scale than an infinite scale. You don't push your IQ to 1000. It's not an anime energy level. This makes no sense.

Then there are things like "Right now we've got assortative mating going on at a scale relatively unprecedented in history" (assumption or data-based fact?)...

With only 7 billion people on the planet the theoretical IQ maximum is around 190. You can calculate this value from the CDF (cumulative distribution function) of a normal probability distribution having a mean of 100 and std of 15 (the definition of the IQ graph). Unfortunately the integral of normal pdf is the error function which is difficult to solve with high precision.

Allowing for all the assumptions in the article I'm not sure I'd want to go ahead with it.

Living in a world where you are that much smarter than anyone else could be intensely isolating, imagine been so smart that you would find every conversation with anyone who isn't similarly boosted boring.

> The alternative would be inequality of a kind never before experienced in human history.

That is a pretty chilling sentence and it's hard to argue, if you could engineer children for intelligence and it was fantastically expensive then that has some really nasty potential sociological outcomes (something that has been a Trope of sci-fi for at least a hundred years).

> This means that there must be at least thousands of IQ alleles to account for the actual variation seen in the general population.

Or that IQ is only governed by genes to a very small extent in which case the rest of the argument goes out of the window.

You're right that the author (unless I missed it) didn't say what proportion of variance in IQ is explained by additive genetic effects (narrow sense heritability) . I think a quick search would likely answer that.

The issue I see with changing so many alleles is that we may not know all of the phenotypic consequences of each allele. Moving 10,000 alleles to the best intelligence settings might make people smarter, but genes are highly pleiotropic. Mutating 10,000 genes could have 10,000 phenotypic side effects.

How do you optimize: maximize intelligence subject to not messing up the organism too much. How would you do this for multiple phenotypes? Maximize height and intelligence subject to not messing up the organism too much.

Superhuman intelligence needs to have the corresponding nurturing factor as well. It won't do much good if you are born super-smart but are trapped in poverty or circumstances which limit your potential.

Plus there is also the EQ (emotional) quotient. Not many people are going to want your company if you are a super-intelligent asshole either.

> It won't do much good if you are born super-smart but are trapped in poverty or circumstances which limit your potential.

I'd argue the opposite. If you're in a tougher situation, there's a lot more for you to gain by applying your intelligence.

Poverty causes much more than can be helped. Bad schools that are falling apart, a lack of opportunity, and the effects of long term stress on the brain are all things that will affect any person regardless of intelligence.

You would have more to gain, sure. But you would not be able to reach the level you could have reached with the right education, or with good nutrition, etc...

The article actually also mention this, in the passage about the Flynn effect.

There appears to be considerable disagreement as to whether EQ measures any real underlying trait or has any independent utility in predicting/determining outcomes, e.g., in work.

Tldr? Title seems clickbaity and I don't want to reward that.

Nautilus make consistently good articles, you can excuse the headline this time.

TL;DR: Scientists have discovered some parts of DNA that reflect intelligence and advances in DNA manipulation could lead to being able to create super-intelligent humans.

I wish there were a way to quantify that impulse, because that is often my first thought when presented a click-baity link: "Shall I award my attention to this"?

Sometimes I want to click on an enticing clickbait article from buzzfeed, but can't let myself because its buzzfeed and a listicle.

Several comments have mentioned that high intelligence can result in a very lonely, isolated, and depressing existence, as you find that you can communicate with very few people, and how sometimes very bright people wind up either anti-social or suicidal.

I wonder about the implications of this for super-intelligent AI...


Haven't we heard this before?

And it hasn't lost its allure one bit.

Can confirm. Daughter turns 13 in 2 weeks. Pretty soon she's going to know everything.

I would gladly fork out the cost of a university tuition for CRISPR therapy to give my children the edge. But many would be adverse to this, probably to the extent of being a same sex marriage/abortion level of debate.

If I understand correctly, techniques like CRISPR might allow for for gene editing in adult humans.

If so, how much would you be willing to pay to have your IQ increased by say 10 points? I think I would be willing to pay $10,000 if the risks of something bad happening because of the treatment would be low.

You would also need some way of re-inducing plasticity in the brain's connectome. This seems less likely to be feasible after adulthood.

Also, IQ is a lousy measurement scale. Ten points means a lot more to someone with an 85 or 95 IQ than to someone with 120.

Sci-fi story plot twist: it's possible, but your personality completely changes and you lose your earlier memories. You don't get actually any smarter, but you have a brief period in your life where you learn super fast.

The concept in the movie "Code 46" where the global elite take a virus to learn a skill always stuck with me.

Skills based enhancement - becoming more empathetic or maybe more mathematical or acquiring novel light-sensitive retinal cells or biological radio seems more plausible than boosting "general intelligence" which might not be a thing at all.

Y modifying genes of a living creature is obviously going to be a low risk procedure.

Probably not, but it's not a problem, because you are only going to realize after the fact.

interesting study .. but it seems to be heading to a rather narrow direction .. humans are multifaceted beings that are far beyond what IQ can describe .. I wonder if the authors could conduct a study to measure IBM's Watson and see conclusion one could arrive at .. Have fun!

And why exactly should we assume the "additivity of small effects" assumption holds far from the center of the current g distribution?? Stack up enough of these variants and you're into unknown territory, including possibly inviable embryos.

I just hope that they watched enough Planet of the Apes movies to not test it on monkeys first.

“the case for substantial genetic influence on g is stronger than for any other human characteristic.”

How does that fit with the popular notion that intelligence is independent of race? Isn't the definition of race essentially just different genetic background?

if you haven't seen it, the film Gattaca covers this very well http://www.imdb.com/title/tt0119177/

Unintended side effects are the risk here. You can't legally 'test' this out, so it would have to be trial and error by parents. Wouldn't want to be an early adopter here.

That must be a somewhat daunting procedure for scientists in the genetic field; creating humans that are much, much smarter than themselves, likely rendering themselves obsolete very quickly.

There'd still be at least a generation before the kids are up to speed. Even if they're of the "finish HS by 12, BS by 13, PhD by 16" types, they still have to accumulate a large amount of knowledge and practice, to actively participate in the science.

And I think science is not a zero sum game.

There's only so much science to go around, so don't go hogging all the discoveries.

Unrelated to the article but the design of that website is so nice

From the article: ". . . if combined in an ideal fashion, could lead to individuals with intelligence that is qualitatively higher than has ever existed on Earth: Crudely speaking, IQs of order 1,000, if the scale were to continue to have meaning."

That's an amazingly ignorant comment. First of all, the better evidence is behind the idea that currently observed high human IQ, when the "g" factor is factored into consideration, consists mostly just of luckily lacking deficiencies in function that other human individuals have. In other words, the centuries-long factor-analytic approach mostly suggests that high-IQ individuals luckily lack "general stupidity," but still are subject to hard limits on human cognition (Mackintosh, IQ and Human Intelligence second edition 2011, pages 155-156).

And the hard limits on human cognition are intractable, and actually illustrated by the article kindly submitted here. Psychologist Keith R. Stanovich points out in his book What Intelligence Tests Miss that even very high-IQ individuals (studied in studies of members of high-IQ societies) are still subject to most human cognitive biases, and they can be MORE subject to delusional beliefs than members of the general public because they can make up elaborate rationalizations for anything they choose to believe, including a belief in superintelligence.

I doubt the statements in this article also because there is a long line of research that goes back to before when the article author was born on selective breeding of mice and rats for high intelligence in their behavioral context. (One kind of test of intelligence given to rats and mice is running through mazes to find food.) Since before the author was born, there have been purebred strains of "maze-bright" and "maze-dull" rats that have been tested by psychologists. Now that there are gene technology methods in animal research, there are transgenic mice with knocked-out genes and added genes, and the author significantly fails to mention any experimentation in animal models that shows that his speculation is likely to become true.

The Wikipedia article on IQ classification adequately cites references to demonstrate that the IQ scale is strictly ordinal, so indeed a notion of "IQ 1000" is meaningless.

If you'd like to read a more thoughful article on genetic influences on human IQ, by a researcher who has been closer to the research for longer and who has published in better journals than the author whose essay was kindly submitted here, see Wendy Johnson's article "Understanding the Genetics of Intelligence: Can Height Help? Can Corn Oil?"[1] which has an excellent review of decades of research on genetics and on human intelligence.

[1] https://scholar.google.com/scholar?q=Understanding+the+Genet...

I dont think we need smarter humans, i think we need more normal-intelligence humans who like working hard.

I respectfully disagree. We need a society of people who hate working hard but, very importantly, are deeply empathic so they won't just leave the hard work for other people. Those are the people who automate the hard things away and share their work so other people can do the same, to leave time for science, abstract creative pursuits, recreation, etc.

>share their work

If you mean share for free, then I don't think that's ever likely on any large scale. You're talking about altruism, which is very exploitable by the scrupulous.

You're talking about altruism

Not at all. We share things so others can build on our work and share their changes back again. The work gets improved (directly by contributions, or by testing, or just through feedback and criticism), and we benefit from those improvements. It isn't altruism because the reason we share the code is to get those improvements. Without community feedback very few people would bother opening up their source and putting it out in the world.

Yes, we definitely need more mindless wage slaves and worker drones who like being used and exploited.

"You should just feel lucky to have a job at all, who cares that they aren't paying you in level with the amount of service/product you produce for them..."

I hate this embrace your servitude mentality.

Time for a "Brave New World" indeed.

I think we need people who have more empathy for others, have more patience, have more respect for fellow humans, animals and the environment in general. I feel working hard, intelligence etc are secondary to empathy.

And what will they do, when robots take all their jobs in some time?

this is highly subjective topic, nobody really knows what mankind truly needs long term in current state. we can state our own personal preferences that would improve things we care for, but that's another topic with possibly different outcomes.

For example, I can easily say we need more people with high moral standards who don't lie, cheat partners, friends,family and do overall positive added value to mankind. This can be anything from freeing a seat in full bus to granny/pregnant woman to curing some specific type of cancer, and anything in between. They don't even need to work hard on this, even small positive action here and there, multiplied by 7 billions, would change our lives gradually for the better. But that's just my subjective 2 cents :)

>normal-intelligence humans who like working hard

Because working hard is a virtue, right?

You may have said that sarcastically, but there is hard work other than manual labor. I think 99% of what we call "intelligence" can be improved with hard work directed toward learning and practice.

Who said anything about manual labour?

Have you watched CNN lately? I don't think so.

As if intelligence per se were some kind of virtue. All it does is give you an edge over the less intelligent, like a boatload of money gives you an edge over poorer people. The utility lies in the difference. If you lift everyone's level simultaneously, nothing changes. When all is said and done, human stupidity will still be infinite.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact