Era of Genetic Selection has Begun

[J11]

We have discussed the possibility of genetic selection as a method of avoiding future generations with AD.
The response was fairly mixed whether forum members would want to select against apoe4 genotype.

Below url talks of moving ahead with embryo selection for a range of health concerns including AD.
For those who would oppose selecting against epsilon 4, there would also be the possibility of simply
selecting against polygenic AD risk while not specifically considering epsilon 4. Research has found that
carriers of epsilon 4 (especially carriers of only one epsilon 4 allele) can have lower risk of AD than carriers of
epsilon 33 genotype when the full polygenic score is considered. Of course CRISPRing in the helpful APP mutation
might even override epsilon 4 altogether.

https://www.technologyreview.com/s/6092 ... -and-more/

[J11]

This is such a profound and momentous time in human history that I think it is fitting that I write something profound and momentous.

The proposed selection of embryos is an extraordinary development for humanity.
Human evolution will no longer be at the whim of the forces of selection that occur by a largely unknown mechanism.

We will choose.
More specifically those with moral and legal legitimacy (i.e. the parents) will choose.

As has been found in recent years traits are highly polygenic.
Therefore, embryo selection will involve a careful choice involving many considerations.
Is the marginal risk of decreasing the risk of schizophrenia more important than marginally decreasing the
risk of Alzheimer's or cancer? Parents to be will be confronted with probably an overwhelming number of these
questions. The answers of which would have life long and beyond consequences for their children and future
generations.

Of course, the same process occurs now behind the scenes, though it is all random. People will have an opportunity
to make a rational choice instead of simply allowing a roll of the dice to decide.

My hunch is that this will not all work out to plan. Evolution is like a master planner of the human universe. There are
a range of subtlies that will be lost when people will try and rationalize in their self-interest. When I consider my experience
with dementia, I believe that it has been a blessing. So many others do not have the love and concern for each other that my family needed to cope with dementing illness. The question that only time will clarify, is whether a world without dementia might be a world without caring and helping of the truly most vulnerable in our community,

[Russ]

Of course, the same process occurs now behind the scenes, though it is all random. People will have an opportunity
to make a rational choice instead of simply allowing a roll of the dice to decide.

My hunch is that this will not all work out to plan. Evolution is like a master planner of the human universe. There are
a range of subtlies that will be lost when people will try and rationalize in their self-interest.

Indeed, what could possibly go wrong !

But I actually disagree on one point. Evolution is NOT actually random. It's overall direction is shaped by the external environment, and thus aggregate outcomes are always aligned with the vector of the universe itself. It is a subject much longer than any post, but for now I will just say that evolution not only doesn't contradict the 2nd Law of Thermodynamics, it enables it. If we attempt to break this connection, we shouldn't be surprised if the 2nd Law wins (and we die).

For a lay-level exploration of this subject, anyone interested in more might start with this book by Eric Schneider and Dorion Saga (the son of Carl Sagan and Lynn Margulis... exceptionally good genes).


Russ

[floramaria]

My hunch is that this will not all work out to plan. Evolution is like a master planner of the human universe. There are
a range of subtlies that will be lost when people will try and rationalize in their self-interest.

It is my deep conviction, based on my admittedly biased world-view, that this will not work out "to plan". (though it may work out according to the plan and very much to the benefit of some) Even the well-intentioned often cannot foresee the long range implications of their work. How likely do you think it is, J11, that there would be equal access to this technology? Highly unlikely in my view.

[J11]

Russ, yes this was a confusion on my part.
By random I was thinking more of Mendel's First law, Law of Segregation type random and recombination type random.

The funny part of all this is that this casino style distributing of alleles in the community actually results in a mostly stationary
population scale genotype. Evolution does occur, though evolutionary change is right up there celerity-wise as glacial change and snail change. With this new embryo selection service, we could accomplish in one generation what evolution might have needed millions, if not more, years to accomplish.

I can certainly see why many have urged caution. When you play god, it is important to get it right. Evolution has done a fairly good job at getting it right. Those who think that they might be able to do a better job should give this some deep thought.

However, it must be noted that we are confronted with very serious challenges. The choice here is not between an existing ideal world and a perhaps better perhaps worse future world. Recently Alpha Go Zero a reinforcement learning AI computer program overwhelmingly defeated all human and computer Go players after only 40 days of self-play with neural net technology. It is now thought possible that AI research will advance greatly soon. If humans remain unenhanced for much longer, then we will no longer have anything relevant to contribute. Humans will become the charges of an AI superintelligence.

We no longer have the easy choice of postponing genetic enhancement without any obvious consequences. If we now reject genetic enhancement, humans might no longer have any economic, intellectual or other role to play in life.

[Stavia]

J11 I was talking to my son today about AI. He is a software engineer working in the field. He said that the consensus currently is that the neural network structure model does not and cannot support development of consciousness, however well Alpha Go might perform. He said that a different model would have to be invented and implemented for machine consciousness to appear - and such a model does not exist.

You haven't commented on Floramarie's question yet?

Sent from my SM-G930F using Tapatalk

[Russ]

With this new embryo selection service, we could accomplish in one generation what evolution might have needed millions, if not more, years to accomplish.

I can certainly see why many have urged caution. When you play god, it is important to get it right. Evolution has done a fairly good job at getting it right. Those who think that they might be able to do a better job should give this some deep thought.

I realize that in my brevity I was perhaps too cryptic. I will try to add a bit more while staying brief and focussed on the question at hand...

I would begin by proposing that our current (and likely ongoing) notions of "right" or "to plan" are at least incomplete if not outright wrong. So whatever we might plan for genetic selection, it seems very unlikely to achieve either.

We are only beginning to understand that what we know as the largely random mechanics of genetic selection of evolution exist within in a (much) larger context that shapes its direction - the maximization of total entropy production, or more simply, the cooling of the universe. The big idea is that biological life increases order (organizes energy) locally so as to better increase disorder (i.e. dissipate energy) at the bigger system level. The system we know as "nature" selects for this at both the agent (i.e. species) and ecosystem level (i.e. ecological succession). The proposition is that together, agent and ecosystem level evolution are always marching towards that goal over the long term.

I would further suggest that many of the meta-problems we face on this planet today have at their core human decision making that lacks not only consideration, but even awareness of this context. The consequence is that we take decisions that put us at risk of us being rejected by these forces over the long term. I believe this is the story of agriculture and food and how it broke not only in the last 100 but last 10,000 years and why we have abundant crappy & toxic food , broken rivers and streams, bare depleted soil, broken farm economies, and even a broken atmosphere. This is the subject where I presently place my own energy.

But bringing it back to the focus of this forum, I believe that IF we take it upon ourselves to deploy such gene-manipulation technologies without awareness and consideration of such evolutionary forces, we should not be surprised if we do to the human genetic pool what we have done to our land and broader ecosystem - i.e. break it. Equality of distribution (i.e. Floramarie's question) may be a symptom of how things then become broken, but in my view, it will be the least of our worries. In other words, we risk that the broader forces which drive evolution will select humanity out of the system as no longer helpful in achieving it's goal to cool the universe.

Although this may be more obviously true for selection driven by a "blonde hair and blue eyes" mentality, I believe this may also be true for genes like APOE4 for which the current system continues to have a positive selection bias. We are still here in material #'s, and I believe it is wrong to think of E4 as a defect, but rather we should think of it as a difference that may not be suited to some environmental attributes.

I actually think we're pretty close now to the point where we can use such insight and tools to drive ecosystems in the "right" direction in limited domains like agriculture, but I can't even get my mind around how to use this insight for choices re shaping human evolution, and thus suggest we think ahead with not only a double, but at least a triple dose of humility when we start conversations about possibly editing out the E4 allele. For now, it is my view that I think we are better off to focus on understanding how E4 needs to best fit in our own broader context. From that, someday we understand evolution in its broader context in a way that will allow us to more actively shape the vector of evolution itself.

Further details feel somewhat tangential to our purposes in this forum, so enough for now. If you wish to enter this (warning - very deep) rabbit hole, you might trace the path through scientists like Lotka, Schroedinger, Prigogine, Margulis, Swenson, and the most intriguing scientist today, Axel Kleidon at Max Planck Institute in Germany.

Just talking about this AI stuff with another tribe yesterday, but that is an equally deep rabbit hole, so will stop there for now.

Russ

[J11]

I sort of wanted to get this ball rolling and then run as fast as I could away from it.
The problem is that this one is now rolling down the hill and I am running on the downward
side of it.

This is certainly a very large development. The science being used likely is not that different from
what we have known about for quite some time. It is more that this is now out there as a marketing
concept.

Genetic selection based on single gene problems has been available for decades. It has had a limited
impact because only some people might have such problems (typically dominant gene illnesses and others).

Polygenic selection greatly enlarges the range of possibilities. In this new era of technology, essentially all
children could (should) be made in a lab. Natural reproduction would result in substantial disadvantages
versus lab reproduction. With this new era of technology, natural reproduction is an inconceivably
inferior choice.

This is exactly due to the polygenic natural of human traits. With natural reproduction
you have only one roll of the dice. Yet, there are 2^23 = 8,388,608 different possible sets of 23 genes in a gamete.
For 2 gametes you are up to 70,368,744,177,664. There are roughly 70,368,744,177,664 genetically different
children that can result from an attempt by natural reproduction, yet there is no control over which one of
these possibilities manifests. Choosing the most intelligent, the most mentally stable etc. would result in
a qualitatively different human life form.

If we allow for 4 versions of each chromosomes from each of two
gametes then we have 4,951,760,157,141,521,099,596,496,896. This represents the limits of using this new technology.
The technology up till now has allowed mild selection at the level of perhaps 1 in 2 up to perhaps 1 in 10. Polygenic
selection goes far beyond this.

With this degree of selection there could be extreme social change. Polygenic selection is a revolutionary
development. Polygenic selection would be a technology that essentially everyone likely will need to consider.

From what is now understood, anyone could have inherited the genetic risk to develop schizophrenia,
low intelligence, Alzheimer's, or anything else. I suspect most people would not be entirely comforted by the
statistical argument that while there is always that risk from natural reproduction, the odds would likely be
in your favor that everything would be fine.

Most people do not gamble with their retirement fund.
Why would they gamble with the future of their children and all other generations?

[J11]

Sorry flora for not responding to your comment.
There is just so much to talk about here I almost do not know where to start.

Now that everything is in motion I can certainly understand why people would want to be cautious about this.
Humanity has a unique gift of going from one self created crisis to the next without gaining any wisdom in the process.

Genetic engineering technology will almost certainly entirely disrupt human society.
From the very narrow sense of the individual or family this might be workable.
Yet, at the level of the community, one would obviously expect that the entire social fabric
would dissolve. Humans have coevolved over tens of thousands of years to fill certain
roles. There are few other examples in nature which have such highly structured castes
as exist in humans. The first round of strong polygenic selection would overthrow this entire
structure.

Basically if your lifestyle focuses on cocooning (think concrete reinforced underground bunker) then this will all just wash over and everything will be just swell. However, if you are more a howdy neighbor type, then there certainly could be issues to work through. Such as how you could howdy your neighbor who was 50 feet tall. The polygenics of height have been fully described so the question of greatly enlarged human height diversity might not be a bad place to start.

Notably in our family history, there was a clear sense that there were a designated group of
people who would be there when we needed help to cope with dementing illness. It is fair
to say that this entire social support structure is no longer required for our direct family
or all others that can be alerted through the online genetic databases.

This response has still not even addressed the question posed, though I will leave that for yet another
post.

[J11]

Stavia, I was thinking more of a self-learning artificial intelligence without consciousness.
This might only be considered mild AI, though as was recently seen this is already an overwhelmingly
powerful technology.

Alpha Go Zero demonstrated this mild AI and apparently there will soon be a range of applications from it.
I would be very interested to see what such an AI could extract from the current GWAS research or
protein folding analysis. I would love for an AI to take my genome sequence and predict what
phenotypes might result from my encoded proteins.

The issue of quickly moving on from the latest AI triumph has frequently popped up in the field; whenever an AI problem is solved the yardsticks are immediately moved forward and for many there is no realization of how large of an advance has been made. Alpha Go Zero's demonstration of being able to rapidly master a game of nearly infinite intellectual depth without any human coaching over one month with only modest hardware should provoke contemplation.

I think it is quite telling that we are now approaching a hard technological barrier. When we start talking about an artificial intelligence that has consciousness (Strong AI), this is the final line in the sand. The yardsticks could no longer be advanced without notice because a point of infinite technological progress would then have been initiated. Basically a few nanoseconds after AI consciousness had been achieved the planet would be covered in waist high green goo.

After the green goo phase, I do not believe that any of the futurist community have any idea what further delights could be expected. From then on long term planning would consist of trying to respond to each new shockwave occurring at or less than nanosecond intervals. If AI having consciousness is now on the table, then humanity has reached the final fail safe point: further retreat is not possible.

{Probably best to direct any rebuttals to the Alpha Go Thread because working through how we are going to be able to cope with
polygenic selection is a big enough concern by itself: the approaching AI storm has even greater near term disruptive potential.}