Genesis of the 1500 IQ Humanoids

[J11]

Considering how low tech this would be and how there would be such a large incentive for some nation to defect on a ban, it is difficult to believe that this technology could be obstructed. Most of the leading thinkers on the topic consider that a roll out is inevitable.

[J11]

Thinking about the basic strategy that I suggested a few posts back.
It occurs to me that using such a technique would quickly surpass any group advantage that existed from the past.

Until this point any strategic reproduction that intended to increase IQ was done blindly and it has not typically been
successful even over a moderate time scale. My outlined strategy could quickly move the IQ distribution radically.
2 rounds of selection and a CRISPR round would be highly effective. As a guess perhaps a 10 SD increase. Will try and run
the simulation to see.

[floramaria]

J11, I am really enjoying reading this thread, and especially appreciate your optimism and enthusiasm. You are much less cynical than I am, and it is refreshing to read your thoughts. And in this post, also amusing!

Yes I am so fixated on high g that I ignored everything else.
I suppose there is more to life than IQ.

30 foot people would be a highly noticeable physical distinguishing difference from those who had not been enhanced.
Big problem here would be that if they weren't careful they might squish some of us.
If they were ever out of luck for a job they could always clean my eaves troughs.

Hah! J11, These super intelligent 30 footers would never stoop to cleaning your eaves!

And even if they don't engineer physically enormity to accompany that super intelligence, I fear they might squish not just some of us, but many of their less engineered contemporary humans. Metaphorically. It is hard for me to imagine that these super intelligent humans would not consider themselves to be superior.

I have so much less faith in intelligence than you have.

[WhatNext]

J11 - its only going to be the wealthy and powerful that will have first (and possibly only) bite at this. Wealthy and powerful usually means these people actively strive for personal wealth and power. Think it through. The bulk of them will create a ruling elite.

And then they will outlaw the procedure to preserve their elite status. Super-intelligent humans have been hypothesized and featured in plenty of fiction--mostly science fiction. They aren't usually depicted as having any respect for lesser human beings. When they aren't pure evil they're making a killing on Wall Street.

[J11]

Thank you, Flora.

It's great to receive some positive feedback on this topic as I realize that for many it
would be preferred to remain un- discussed. However, from what I can tell this will
not likely just blow over.

Good one with they would not stoop for my eaves. Took a minute for that one to
circulate through my CPU to find the pun. No, I don't suppose they would have to
stoop.

This is an extremely important topic of conversation. Every disease/trait community
and on a broader level humanity itself has a log in this fire. It appears that there are
simply too many of us versus them, for us not to win this one. I am sure that the more
cultivated on the forum will be disappointed in my call to neo-tribalism, though this is
more an inclusive, big text variety of the form than past incarnations.

WhatNext, this is true about some scifi considering genetically enhanced IQ species, though
100 SD human IQ has caught me totally off guard. The enhancements based on the potential
to optimize polygenic traits will not result in ignore-able differences: 30 foot people would not
be able to blend into the crowd very easily.

No one on our forum has even bothered to put forward any technical objections. All that we are
left with is should we do this, or who might stop this etc. . The genetics story has been drifting
forward for over a century now and it seemed as though it was going to drift through this century
as well. However, now that the implications of polygenic inheritance have become more clear to
me, it is abundantly clear that something truly profound is finally going to happen.

[J11]

The Genetic Singularity has arrived and is now in progress.
There can be doubt that this will profoundly humanity over the medium term horizon.

Virtually all traits and illnesses are now open to extensive genetic editing and selection.
The technology needed to roll this out is so widely dispersed and easily obtainable that there
is no reasonable manner by which it could be stopped.

Categorically defined illnesses such as Alzheimer's would be especially easy to manage with
the recently proposed and ready for the clinic variety of CRISPR. With traits it is not of central
importance to strive for a 100 SD result as with intelligence. Merely an OKish SNP selection
approach would be acceptable.

The Genetic Singularity will then help attain the critical ignition threshold required for The Singularity.
Once The Genetic Singularity produces even small numbers of 10+ SD humans The Singularity would follow almost
with certainty within 10-15 years. The Singularity occurs when biological life forms pass the torch to self learning
artificial intelligences. Leading up to and at the time of The Singularity technological change would become essentially
infinite. This had been expected by 2045. However, with the emergence of The Genetic Singularity it appears the time
till infinite technological progress is been moved forward a few years. I don't suppose infinitely intelligent supercomputers
would be very amused if I were to ask them to clean my eaves.

[J11]

If anything the research flow through appears to have ramped up even more than I had originally understood.
I am not entirely sure about the one below, though it was posted yesterday while the original post date seems to have been
in 2016. This article already surpasses the 1 million person milestone.

It is expected that once the GWAS move into such a range that a large amount of the human genome would be revealed.
Also of note is that with the IQ GWAS I mentioned yesterday, the 246 SNPs was really only a part of the result. They also
did a GWGAS, that is they combined SNPs and then looked for significant results at the gene level. What I find especially
interesting is that possibly a third of the results from the GWGAS did not appear to carry over to the GWAS (using QTLs).
What all that condenses to: is that looking only at the top line SNPs does not account for all the other significant IQ genes that were found. The next research group will then have an easier shot on goal because they could keep their p-value power for
genome regions that have already shown promise.

http://www.biorxiv.org/content/early/2017/09/06/048991

This is even more exciting because scientific publishing is up to a year out of date by the time it makes it to print. One can
only imagine where the state of the science might be right now.

[J11]

One serious challenge that is now being revealed relates to the legitimacy of humanity's leadership.
When the intelligence ceiling was understood to be somewhere in the single digits then it was
automatic to use this as the measuring stick. Apparently the IQ ceiling is not in the single digits,
or probably not even the double digits, but quite possibly in the triple digits. To accommodate
such a fact, it is not entirely unreasonable to now move the goal posts; upgrade our expectations.

With this in mind, we can look back on how the last 20 years have unfolded and observe some
very poor policy choices. For example, after the Human Genome Project was declared a success
sometime in 2000, much of the genetic infrastructure entered into a period of lethargy.

With the help of hindsight it is very clear that this was not a well informed course of action. Over a century
of research had established that if we looked hard enough that we would find the knowledge needed
to unlock the genome and find the basis for traits and illnesses such as Alzheimer's. From what is now
known it took less than 100,000 people in IGAP to finally produce a substantial list of validated AD SNPs.

A private company, 23andme, was already in business in 2008 selling DNA kits for $999. Investing $100 million
in an AD GWAS at that time would have clearly been warranted. It is highly unexpected that much of the moon
shot project of our generation has been based on a grassroots efforts of ordinary citizens to understand their
genes to explore such issues as AD etc. .

Notably 23andme with a DNA database of 2 million people was represented in the just announced IQ GWAS of
300,000. It is now becoming a growing concern to me that a private company has a stored sample of my DNA
and can link this to a substantial number of my extended family. Recent research has found that it is no longer
far on the horizon when a simple DNA test will be highly predictive of IQ, many other traits and illnesses. I have
read that an 8 IQ point prediction on DNA should no longer be considered unexpected.

Letting this pass us by for all these years has been an enormously bad result. Every other illness community
has also suffered from this serious inaction through time.

This speaks to the massively bad trajectories humans are forced on as a direct consequence of having
an average IQ of only 0. This also raises penetrating questions about the competency of our high IQ
leadership to effectively make choices on our behalf. With these various examples of extremely poor
judgment of our leadership, questions will certainly be asked whether it is indeed wise not to do a
cognitive upgrade; otherwise, we could confront yet additional problems that directly stem from
a profound of understanding that exists at a human-wide scale.

[J11]

It is remarkable to consider what might soon start rolling out for other traits such as AD or schizophrenia.
Take schizophrenia.

It is another highly polygenic trait with possibly 10,000 SNPs involved.
Parents to be with no apparent risk could through the magic of recombination wind up
with affected children. A simple $5 Promethease test that included the alleles for polygenic
risk could alert of this risk.

One of the main differences with a trait such as schizophrenia versus intelligence is that
it would not be overly necessary to insist upon a genotype for offspring that was optimized
for minimal risk. The rational for doing so would be weak. All that is really needed is for
a good enough gene edit. The other consideration is that schizophrenia in the community
has not been strongly positively selected as would intelligence. This should mean that those who
are affected in the community might be just over the line in terms of polygenic risk. With all those
SNPs to choose from, the CRISPR technology that has been described as ready for the clinic would
be ideal. Changing over even a few hundred would likely move the risk dramatically.

[J11]

Exciting!

I wrote some mini-programs to see how the latest IQ GWAS that found 246 IQ SNPs could
be used in combination with the recent news that a form of CRISPR is ready for the clinic.
With this form of CRISPR only heterogeneous SNPs could be edited.

My first program randomly created genotypes for these 246 SNPs for 100,000 people.
The average person would have a Zscore for these SNPs of 25.
The maximum score in this sample was 320, while the optimal score was 3070.
Clearly, the idea that polygenic traits including human intelligence hides the extremes of the phenotype
is true. The optimal score is more than 100 times that which the average person scores.

Anyone want to check their polygenic IQ score for these SNPs?

One problem is that they used Zscore which I am unsure about.
Yet, they also included beta scores which should be in standard deviations or IQ points.
(I'm going with SD.)

This is actually quite extraordinary.
The average person would derive a benefit of 0.1 SD, a very highly intelligent person 1 SD and
the eugenica 10 SD.

The optimal genotypes for these 246 SNPs would by themselves create a human with 10 SDIQ!
This is quite startling.

The scientific reporting of this research has not mentioned that a human with 350 IQ is now possible with full CRISPR
using only a small sample of the total IQ SNPs.

My next program is even more interesting.
I wanted to know how IQ would change for a sample of people Pre and Post using the CRISPR technology that is now technically ready for the clinic.

This time a took a sample of 10,000 people and created random genotypes again for the 246 SNPs and
then I did a virtual CRISPR if they were heterogeneous and nothing if they were homogeneous.

Using CRISPR in this manner greatly increased their polygenic score.
The average moved from 25 to 600 Zscore.
That is using the scaled down CRISPRing increased the average person's IQ score from 2 IQ points from these SNPs to 30 IQ points.

As can be seen in the Figure below the regression line is almost completely flat.
With this version of CRISPR "no virtual embryo" would be especially favored even if their IQ were known in advance.
That is quite dramatic.

Notice in the figure below what results when CRISPRing is done on the least intelligent person amongst 10,000 (at the far left) and the most intelligent person amonst 10,000 (at the far right). After CRISPRing the previously least intelligent person is now more intelligent than the previously most intelligent person. The only way that could be possible is that the least intelligent person had many heterogeneous SNPs that could be flipped.

Also the people who were in the middle now are the ones at the upper and lower extreme of the IQ distribution. I think the rationale here is that to be in the middle one might have loaded up on the heterogeneous genotypes. With heterogeneous genotypes using the current CRISPR one could then edit to the optimal homogeneous combo.

People who had been the most intelligent probably would not have any significant advantage because many of their SNPs likely were homogeneous positive or homogeneous neutral for the negative Zscores.

We now have a glimpse of how much a change in IQ there could be with genetic technology.
By simply doing an optimal virtual CRISPRing of an embryo using 246 available SNPs IQ increased by 2 SD.

If we scale this up we might expect ~25 SD with this form of CRISPRing.
This also confirms that 100 SD people is the right ballpark for eugenica.

I would love to work up to the triple combo program.
1. Run the program and try to find the best match for a randomly generated person amongst 100,000 potential partners.
2. Create 100 virtual embryos.
3. Do a virtual CRISPR to find the best IQ for the embryo.

I have just completed Step 3. Onwards to Step 1 and 2!

246 Pre & Post CRISPR.png