Is Natural Selection smarter than we are?

[Mr Skeptic]

So, lucaspa and I seem to disagree with this one. Rather than bring this up on every thread where the topic appears, I'd rather discuss it in a new thread.

 

I didn't know whether this thread belongs in the evolution forum, psychology forum, or computer science forum, since aspects of each are relevant.

 

I think Eugenics is just stupid. It is based on faulty science -- in this case faulty evolutionary biology and natural selection. The basis of eugenics is that some traits are always "good" and should be encouraged and some traits are always "bad" and should be eliminated. In evolution, traits are "good" or "bad" depending on the environment. Each trait comes with a cost as well as a benefit. Natural selection is constantly evaluating the cost vs benefit of multiple traits and how they interact.

 

So eugenics assumes that we are smarter than natural selection. However, the data emphatically says we are are not.

 

Here you are comparing our intelligence with that of natural selection. Separating them as two different things, and judging one smarter.

 

We turn to natural selection (in the form of genetic algorithms) when the design problem is too tough for us and we have no idea how to make something. We let natural selection do the designing for us. And we often end up with designs that we can't even figure out how they work, much less would have known how to make them.

 

"I'm really exploring what evolution can do that humans can't," he [Thompson] explains. "There are properties that humans have great trouble designing into a system, like being very efficient, using small amounts of power, or being fault tolerant. Evolution can cope with them all." Evolving A Conscious Machine BY Gary Taubes Discover 19: 72-79, July 1998

 

Now, if natural selection is better than us at relatively simple electrical engineering problems, why would anyone think that we would be better at natural selection in the much more complicated area of the entire human genome? Yet that is exactly what eugenicists do think.

 

And here, you are lumping a human-made algorithm running on a human-made computer with natural selection. Nevermind that the algorithm is human-made, it is a different algorithm from that of natural selection and evolution.

 

The power of massive amounts of trial and error, huh?

First, Darwinian selection is more than "trial and error". That is a demeaning soundbite.

 

Second, that "trial and error" is how humans design! Whether you realize it or not, the program of human designed genetic changes still relies on Darwinian selection -- "trial and error".

 

Darwinian selection is the ONLY method for gettting design. The issue is whether it happens within a brain or outside of one. Human design -- including that of the DNA sequences to insert -- is Darwinian selection.

 

Basically, within your brain you generate variations with your imagination and then run them thru selection against what you want. You pick the ones that seem to fit the criteria (will "work") and then make variations on those in your imagination and test them. Finally, at some point you decide to manufacture what you consider the best design. And then that gets tested in the real world instead of your imagination. In the case of DNA sequences, that real world would be the first individuals with the new sequences (genes). And then we find out if there are unanticipated effects on that trait or interactions with other traits (polygenic and pleiotrophic).

 

Then you begin the imaginative process again by imagining new variations within your brain. So yeah, the power of Darwinian selection is amazing.

 

And now you lump all intelligence together, contradicting your previous statement that natural selection is smarter than we are, since you are now saying our intelligence is also natural selection.

 

Remember what Thompson said: human imagination is limited. There are some things that human engineers have a difficulty including. Natural selection doesn't have that limitation.

 

And now, natural selection has no limits?

 

----

 

Now, I've got plenty of respect for natural selection since it has billions of years of experience. But I disagree that it is in any way smarter than we are. And they are definitely separate.

 

For one thing, evolution/natural selection is a very inefficient algorithm. Favorable mutations are selected slowly, while unfavorable ones are eliminated slowly. Sometimes, beneficial mutations are lost, or unfavorable ones selected. And the mutation rate is extremely low.

 

Compared to that, evolutionary algorithms are very efficient. It can sustain an incredibly high mutation rate, since it will remember the best permutations even as it makes new permutations with them. Its population size and selection pressure can be adjusted as necessary. More importantly, it is guided by humans and is also able to be used with many kinds of problems, whereas evolution is limited to only living things.

 

As to the human mind, we are closer to neural nets than to evolutionary algorithms IMO. In any case, the folks here in Psychiatry and Psychology should know more about this.

[foodchain]

So, lucaspa and I seem to disagree with this one. Rather than bring this up on every thread where the topic appears, I'd rather discuss it in a new thread.

 

I didn't know whether this thread belongs in the evolution forum, psychology forum, or computer science forum, since aspects of each are relevant.

 

 

 

Here you are comparing our intelligence with that of natural selection. Separating them as two different things, and judging one smarter.

 

 

 

And here, you are lumping a human-made algorithm running on a human-made computer with natural selection. Nevermind that the algorithm is human-made, it is a different algorithm from that of natural selection and evolution.

 

I think a better question is weather or not we can grasp our behavior as being the prime source or NS as something that we could even control in time. I think using the reality of time alone is enough, it may not sound to qualify but to honestly view the aspects of this with honesty that it would be wise to point out that being NS means you have to understand what humans will do constantly then for say forever. I mean when we have turned even the autotroph level of microbes in some trophic system into being our version of NS don’t you think it would require constant care forever?

 

Anyways human life of that magnitude already is a global variable in regards to NS with issues like pollution alone. This has not worked out very well, and giving how treatment of this issue I could hardly imagine how NS would be treated in the form of GM by us giving such an environment.

 

I truly think the test to pass my standard for allowing GM would entail a society(prolly unified global) that does not need it in any way already. Such as already possessing enough understanding to avoid things like massive extinction via greenhouse gases and what not.

[CharonY]

I have not read the other thread, but let me barge in uninvited.

Natural selection has a an advantage compared to top-down approaches: what works, works. One basic disadvantage of eugenics is that usually a reduction of variance. If the environmental conditions were constant forever and we knew the complete results of any given eugenic approach, then maybe more alleles which increase fitness may accumulate faster.

But this is only very theoretical. We do not know hardly enough to actually even try to do that on a population level.

[lucaspa]

Here you are comparing our intelligence with that of natural selection. Separating them as two different things, and judging one smarter.

 

I'm comparing the ability to design, not overall "intelligence". In this area, NS is a "smarter" designer than we are. I think you have taken the term "smarter" and applied it to "intelligence" in general, instead of the narrow area I intended. So I hope I have cleared that up: when I say "NS is smarter than we are" I am referring to the specific area of making designs.

 

And here, you are lumping a human-made algorithm running on a human-made computer with natural selection. Nevermind that the algorithm is human-made, it is a different algorithm from that of natural selection and evolution.

 

No, it's not a different algorithm. The whole point of the algorithm was to make it identical with NS. The only difference is that humans set the environment, instead of nature. But the variation, selection, and inheritance are what is found in nature. BTW, you do realize that NS is an algorithm, right?

 

And now you lump all intelligence together, contradicting your previous statement that natural selection is smarter than we are, since you are now saying our intelligence is also natural selection.

 

What I am saying is that humans design by Darwinian selection. As I said "Darwinian selection is the ONLY method for gettting design. The issue is whether it happens within a brain or outside of one. Human design -- including that of the DNA sequences to insert -- is Darwinian selection."

 

Yes, we design by Darwinian selection. But we don't do it as well as NS (Darwinian selection) working outside our brain does. I think the reason is the generation of variation. Our variations come from our imagination, and our imagination is limited.

 

And now, natural selection has no limits?

 

No. I had said what I just ended the last paragraph with: our imagination to generate variation is limited. NS doesn't have such a limit on the generation of variation.

 

For one thing, evolution/natural selection is a very inefficient algorithm. Favorable mutations are selected slowly, while unfavorable ones are eliminated slowly. Sometimes, beneficial mutations are lost, or unfavorable ones selected. And the mutation rate is extremely low.

 

You seem to have some misconceptions about NS.

 

1. Efficiency is different than competence.

2. And selection can work very fast. Recent studies in the lab and the wild show that directional NS can work at rates >10,000x faster than we generally see in the fossil record. This raises the question: why is NS so slow in the fossil record? I think the answer lies in purifying selection keeping fit populations stable.

3. No, unfavorable variations are not picked by natural selection! Slightly deleterious mutations can be fixed by genetic drift, but not by NS.

4. Mutation rates vary. They are lowest in bacteria and highest in humans. In humans the mutation rate is about 20 per individual.

 

Compared to that, evolutionary algorithms are very efficient. It can sustain an incredibly high mutation rate, since it will remember the best permutations even as it makes new permutations with them.

 

NS also remembers the best permutations -- that is what inheritance is. And some biological systems have incredibly high mutation rates. IN fact, in some cases an increased mutation rate is itself selected for.

 

Its population size and selection pressure can be adjusted as necessary. More importantly, it is guided by humans and is also able to be used with many kinds of problems, whereas evolution is limited to only living things.

 

1. These are differences in details, not differences in kind. Population sizes and selection pressures are part of natural selection.

2. It most definitely is not "guided" by humans. Humans set up the system and then let it work. No tinkering. If the system was guided by humans, then humans would know how the end product worked!

3. That Darwinian selection can be used in many venues is the hallmark of an algorithm. That's why I try to use Darwinian selection as the general term and natural selection as referring to what happens in nature in living organisms. However, I have sometimes confused the terms and used natural selection when I could have said Darwinian selection.

4. Remember that this discussion started because you advocated genetic engineering on H. sapiens! You wanted to do evolution in living things. So the situation you set up put human designs in direct competition with natural selection.

 

As to the human mind, we are closer to neural nets than to evolutionary algorithms IMO. In any case, the folks here in Psychiatry and Psychology should know more about this.

 

Again, I'm afraid you made a generalization (and therefore accidentally a strawman) I did not make nor intend to make. You took the word "smarter" out of the context I was using it -- design and particularly genetically designing humans (your original context) -- and made it about "intelligence".

 

Yes, our minds probably work by neural nets. But we design by Darwinian selection. We use the neural nets as the computer to generate variations (in designs) and test them. I notice in the literature that humans are using Darwinian selection to design computer analogs to the human brain so that they can get machine intelligence.

 

Mr. Skeptic, you have never addressed the core of my argument: IF humans were better designers than NS (Darwinian selection), then why do we use Darwinian selection when the design problem is too tough for us? Why, after we use Darwinian selection, do we end up with designs that we don't know how they work? If we were smarter designers, we wouldn't have to use Darwinian selection, would we? And, having used it, our superior design skills (and intelligence) would be able to tell us how the design worked.

 

Thompson still doesn't know how the chip works. It appears to use principles of physics we humans haven't discovered yet.

 

Foodchain, I'm really puzzled by your post. Could you try again?

 

Mr. Skeptic started the thread because he is a champion of genetic modification of humans. He very much wants to use genetic engineering to change H. sapiens to make us, what was it?, "faster, stronger, smarter". So he wants to re-design H. sapiens. Why we need to be redesigned at the genetic level is not clear.

 

One of my objections to this is based on the data that natural selection is a better designer than we are. Pre-empting natural selection to letting humans design H. sapiens strikes me as ultimately a very stupid idea.

 

I have not read the other thread, but let me barge in uninvited.

Natural selection has a an advantage compared to top-down approaches: what works, works. One basic disadvantage of eugenics is that usually a reduction of variance. If the environmental conditions were constant forever and we knew the complete results of any given eugenic approach, then maybe more alleles which increase fitness may accumulate faster.

But this is only very theoretical. We do not know hardly enough to actually even try to do that on a population level.

 

I agree. And this is one of my objections to Mr. Skeptic's advocacy of wholesale genetic engineering to change H. sapiens. He doesn't advocate eugenics so much (which is artificial selection) but genetic engineering. He wants us to consciously change the DNA to make H. sapiens "better". This approach also reduces genetic diversity (variation) by having everyone have the same alleles so that they are all the same intelligence or physical attributes.

 

As you noted, we don't know enough about the current environmnent, much less future ones, to try this approach. I can easily see us gengineering H. sapiens to a situation where a sudden change in the environment renders H. sapiens extinct. We wouldn't have the variation among individuals so that some individuals would survive because they were lucky enough to have the appropriate variations for the new environment.

Edited July 3, 2008 by lucaspa
multiple post merged

[Mr Skeptic]

I'm comparing the ability to design, not overall "intelligence". In this area, NS is a "smarter" designer than we are. I think you have taken the term "smarter" and applied it to "intelligence" in general, instead of the narrow area I intended. So I hope I have cleared that up: when I say "NS is smarter than we are" I am referring to the specific area of making designs.

 

Either way, I disagree. At best, NS has had a few billion years more time to work than we have. If you gave humanity 2 billion years, we could outdo anything that natural selection has done. And, of course, NS only works on things that are self-replicating.

 

No, it's not a different algorithm. The whole point of the algorithm was to make it identical with NS. The only difference is that humans set the environment, instead of nature. But the variation, selection, and inheritance are what is found in nature. BTW, you do realize that NS is an algorithm, right?

 

Of course I know its an algorithm. That's why I called it an inefficient algorithm in the post you replied to. The point of the genetic algorithm was not to make it identical to NS. It was to allow computers to design things with almost no human inputs. The most notable difference is that genetic algorithms have an arbitrary, human-designed fitness algorithm, whereas NS has only reproductive capability of a self-replicating thing in the real world as its fitness algorithm. Also differenare how variability is introduced (also arbitrary), how reproduction occurs (also arbitrary), and how selection occurs (also arbitrary).

 

What I am saying is that natural selection is a specialized type of genetic algorithm. Not the other way around. Just like in biology, a monkey is a type of primate but a primate is not a type of monkey.

 

What I am saying is that humans design by Darwinian selection. As I said "Darwinian selection is the ONLY method for gettting design. The issue is whether it happens within a brain or outside of one. Human design -- including that of the DNA sequences to insert -- is Darwinian selection."

 

Do you have any evidence for that? I don't think that I use Darwinian selection for 1+1=2, which is something used for many designs. I don't use Darwinian selection for using the equations of physics.

 

Yes, we design by Darwinian selection. But we don't do it as well as NS (Darwinian selection) working outside our brain does. I think the reason is the generation of variation. Our variations come from our imagination, and our imagination is limited.

 

I won't agree that we design by natural selection until you provide some proof. Our imagination is not completely random, which is why it is better than natural selection. That's what gives us the ability to plan. We can recognize when we have part of a problem solved and then work on the missing parts. And we can design computers and computer programs to do modeling, trial and error, etc. for us.

 

No. I had said what I just ended the last paragraph with: our imagination to generate variation is limited. NS doesn't have such a limit on the generation of variation.

 

That's because we don't use natural selection to think. It is too inefficient to get things done with our limited resources. By inefficient, I mean that NS wastes many computations. Not that the results are inferior, but that there are ways to get similar results with less computations.

 

You seem to have some misconceptions about NS.

 

1. Efficiency is different than competence.

 

Yes, I never said that NS was incompetent, only inefficient.

 

2. And selection can work very fast. Recent studies in the lab and the wild show that directional NS can work at rates >10,000x faster than we generally see in the fossil record. This raises the question: why is NS so slow in the fossil record? I think the answer lies in purifying selection keeping fit populations stable.

 

Whatever the reason, NS can only work on the variability that is created by mutations. And the generation of variability is very slow, compared to that created by either genetic algorithms. Genetic engineering is rather slow to create variability, but it is almost always useful variability as opposed to mostly neutral or bat variability generated by mutation.

 

3. No, unfavorable variations are not picked by natural selection! Slightly deleterious mutations can be fixed by genetic drift, but not by NS.

 

Which is part of natural selection. As I said, very inefficient. If you don't want deleterious mutations to be fixed by your selection algorithm, then you need a different selection algorithm.

 

4. Mutation rates vary. They are lowest in bacteria and highest in humans. In humans the mutation rate is about 20 per individual.

 

Yes, and they can vary by portion of the genome too. I never said they didn't vary, I said that the mutation rate in evolution is far, far lower than that in genetic algorithms in general.

 

NS also remembers the best permutations -- that is what inheritance is. And some biological systems have incredibly high mutation rates. IN fact, in some cases an increased mutation rate is itself selected for.

 

But not always. Sometimes an accident will eliminate an individual or even an entire population with a beneficial mutation. In a genetic algorithm you can ensure that the best permutations are always preserved.

 

1. These are differences in details, not differences in kind. Population sizes and selection pressures are part of natural selection.

 

Yes, whereas in genetic algorithms, they can be adjusted for optimum results.

 

2. It most definitely is not "guided" by humans. Humans set up the system and then let it work. No tinkering. If the system was guided by humans, then humans would know how the end product worked!

 

Humans set up the algorithms for fitness, creation of new permutations, mutation, genetic exchange, selection, population size, etc. Humans can also insert some non-random permutations. After that, it's all on its own.

 

3. That Darwinian selection can be used in many venues is the hallmark of an algorithm. That's why I try to use Darwinian selection as the general term and natural selection as referring to what happens in nature in living organisms. However, I have sometimes confused the terms and used natural selection when I could have said Darwinian selection.

 

Ah, I was using genetic algorithm as the general term, and natural selection as a specific implementation of it. This seems correct, because a genetic algorithm has arbitrary algorithms for fitness, creation of new permutations, mutation, genetic exchange, selection, population size. Evolution and natural selection have these algorithms, but they are specific rather than arbitrary. Were you using Darwinian selection for this?

 

4. Remember that this discussion started because you advocated genetic engineering on H. sapiens! You wanted to do evolution in living things. So the situation you set up put human designs in direct competition with natural selection.

 

Correct. And despite putting forth only a miniscule fraction of human ingenuity on genetic engineering, and for that matter, despite it being a very new science, we are already making huge improvements in various crops. More importantly, natural selection is often at odds with our own objectives.

 

Again, I'm afraid you made a generalization (and therefore accidentally a strawman) I did not make nor intend to make. You took the word "smarter" out of the context I was using it -- design and particularly genetically designing humans (your original context) -- and made it about "intelligence".

 

And I meant it in both ways. And when we genetically engineer humans to be smarter, we'll be even smarter than natural selection.

 

Yes, our minds probably work by neural nets. But we design by Darwinian selection. We use the neural nets as the computer to generate variations (in designs) and test them. I notice in the literature that humans are using Darwinian selection to design computer analogs to the human brain so that they can get machine intelligence.

 

If you mean that we select the best designs we can come up with, than yes. If you mean that we create random designs and evaluate them, then no.

 

Mr. Skeptic, you have never addressed the core of my argument: IF humans were better designers than NS (Darwinian selection), then why do we use Darwinian selection when the design problem is too tough for us? Why, after we use Darwinian selection, do we end up with designs that we don't know how they work? If we were smarter designers, we wouldn't have to use Darwinian selection, would we? And, having used it, our superior design skills (and intelligence) would be able to tell us how the design worked.

 

I've addressed this before. We fall back to trial and error when we don't know a better way to proceed. Your argument contradicts itself -- we use trial and error as a last resort, because it works on any kind of problem but is extremely inefficient. But we prefer more efficient ways to solve problems if they work.

 

As for why we use genetic algorithms rather than a human-like intelligence, it is because they are much simpler and we don't know how to program a human-like intelligence anyhow.

 

Thompson still doesn't know how the chip works. It appears to use principles of physics we humans haven't discovered yet.

 

Lies. If we didn't know how it worked, it could not have been created by a genetic algorithm designed by us. The fitness function would have said it wouldn't work, and it would have been discarded.

 

Now, maybe he doesn't know why it works, but we do know the principles of physics (unless it was constructed by real-world trial and error).

 

Mr. Skeptic started the thread because he is a champion of genetic modification of humans. He very much wants to use genetic engineering to change H. sapiens to make us, what was it?, "faster, stronger, smarter". So he wants to re-design H. sapiens. Why we need to be redesigned at the genetic level is not clear.

 

To be faster, stronger, and smarter, of course. And maybe glow in the dark. And maybe have gills. Maybe redesign ourselves to live in zero gravity. Whatever we want, really. I don't want to wait around for evolution/natural selection to do its thing.

 

One of my objections to this is based on the data that natural selection is a better designer than we are. Pre-empting natural selection to letting humans design H. sapiens strikes me as ultimately a very stupid idea.

 

I've yet to see any of that data. I think it is just your opinion.

 

Anyhow, we can't really pre-empt natural selection. We will just provide very good variability as compared to background mutation.

 

I agree. And this is one of my objections to Mr. Skeptic's advocacy of wholesale genetic engineering to change H. sapiens. He doesn't advocate eugenics so much (which is artificial selection) but genetic engineering. He wants us to consciously change the DNA to make H. sapiens "better". This approach also reduces genetic diversity (variation) by having everyone have the same alleles so that they are all the same intelligence or physical attributes.

 

Do mutations also reduce genetic variability? Eugenics, artificial selection, and natural selection decrease variability. Mutation and genetic engineering increase it.

 

I do not think everyone should have the same alleles. I think we should create new alleles ourselves rather than wait for background mutation to do it.

 

Please quit saying I approve of removing our genetic variability.

 

As you noted, we don't know enough about the current environmnent, much less future ones, to try this approach. I can easily see us gengineering H. sapiens to a situation where a sudden change in the environment renders H. sapiens extinct. We wouldn't have the variation among individuals so that some individuals would survive because they were lucky enough to have the appropriate variations for the new environment.

 

And I can see us not engineering ourselves and still going extinct, like most species have. Or, not engineering ourselves and then being outcompeted/exterminated by humans who have engineered themselves to be stronger, faster, or smarter and decided that they wanted more land.

[immortal]

Either way, I disagree. At best, NS has had a few billion years more time to work than we have. If you gave humanity 2 billion years, we could outdo anything that natural selection has done. And, of course, NS only works on things that are self-replicating.

 

 

Do you have any evidence for that? I don't think that I use Darwinian selection for 1+1=2, which is something used for many designs. I don't use Darwinian selection for using the equations of physics.

 

 

I won't agree that we design by natural selection until you provide some proof. Our imagination is not completely random, which is why it is better than natural selection. That's what gives us the ability to plan. We can recognize when we have part of a problem solved and then work on the missing parts. And we can design computers and computer programs to do modeling, trial and error, etc. for us.

 

 

Whatever the reason, NS can only work on the variability that is created by mutations. And the generation of variability is very slow, compared to that created by either genetic algorithms. Genetic engineering is rather slow to create variability, but it is almost always useful variability as opposed to mostly neutral or bat variability generated by mutation.

 

Yes, whereas in genetic algorithms, they can be adjusted for optimum results.

 

Humans set up the algorithms for fitness, creation of new permutations, mutation, genetic exchange, selection, population size, etc. Humans can also insert some non-random permutations. After that, it's all on its own.

 

If you mean that we select the best designs we can come up with, than yes. If you mean that we create random designs and evaluate them, then no.

 

 

 

I've addressed this before. We fall back to trial and error when we don't know a better way to proceed. Your argument contradicts itself -- we use trial and error as a last resort, because it works on any kind of problem but is extremely inefficient. But we prefer more efficient ways to solve problems if they work.

 

As for why we use genetic algorithms rather than a human-like intelligence, it is because they are much simpler and we don't know how to program a human-like intelligence anyhow.

 

 

Now, maybe he doesn't know why it works, but we do know the principles of physics (unless it was constructed by real-world trial and error).

 

 

 

To be faster, stronger, and smarter, of course. And maybe glow in the dark. And maybe have gills. Maybe redesign ourselves to live in zero gravity. Whatever we want, really. I don't want to wait around for evolution/natural selection to do its thing.

 

 

 

And I can see us not engineering ourselves and still going extinct, like most species have. Or, not engineering ourselves and then being outcompeted/exterminated by humans who have engineered themselves to be stronger, faster, or smarter and decided that they wanted more land.

 

Mr Skeptic you have to understand that Natural selection is cumulative i.e. accumulation of good designs. You can't expect any algorithm to design gills for you in one step. Humans can only introduce genes which are created by nature. You can't produce a new gene or a protein in one step because it requires some knowledge of the solution some thing which we don't have. You have to take a random walk. I am not saying that there are no efficient algorithms other than Natural selection. Just as there are higher adaptive peaks for a population there can be other algorithms.

 

Our brains do work like neural networks and thats how a micro processor works it has a transistor which works like an axon and a variable resistor which works like a synapse. Infact learning is nothing but just further neuronal differentiation each neuron very well knows to which other neurons it has to connect this is due to lecithin receptors present between the synaptic junctions.

 

 

No matter how much you are hard wired you still have to make some trial and error and optimise it so that you can give an accurate solution. Actually Imagination is done in the pre-motor area of the brain and it is in the Wernicke's area where your ideas or memes originate and you express it through your speech and our imagination has some limits because we don't pass on our memes efficiently we don't ensure that our memes replicate and mutate and this truly corresponds to the fact that the number of people in the scientific community are very less compared to the population of the world.

 

We do create designs randomly (that does'nt mean that we create designs that does'nt make any sense) and learn from our previous mistakes and come up with better designs which works better than the old one.

 

You can't expect algorithms to be simple. Life is information and information leads to complexity.

 

Actually the speed of biological evolution not only depends on rate of mutations but also on how much of those mutations get fixed.

 

Even Natural selection can insert some non random changes to its genome and we have good proof that transposons do produce programmed genome modifictions.

 

Infact I am in favour of genetic engineering. What's the point in studying life if we don't convert our science into technology. I don't really study these things just for the joy of discovering something I study them because I think its worth studying them. Every technology as a risk factor involved in it its all about reducing the risks and move on. Our ability to survive on this planet depends on our ability to manipulate it something which we are doing since the new stone age. The reason why the species are going extinct at a faster rate because our culture evolution has become extremely fast and if we have to save these species we have to accelerate biological evolution which is slow but very good at finding an optimal solution.

Edited July 4, 2008 by immortal

[pioneer]

Is natural selection smarter than we are? If we argue that we are smarter than natural selection, then since humans are making species extinct, this is smarter than natural selection which might decide to keep them around. If it is a stupid idea to kill species, then we are less intelligent than natural selection. In this case, we have to work hard just to keep stride.

 

What it comes down to, is not better or worse than natural selection, but a blend of the two. Humans have their share of bloopers. But they also have their share of things that rise above. If we average it all out, the fact we are even concerned about violating nature seems to indicate somewhat below natural selection. When we were not concerned for nature because there seemed to be a sense of balance, we found a zone where we were maybe slightly above.

 

Here is a paradox. Most, if not all scientists, like evolution. Humans are creating environmental stresses yet nobody trusts evolution. New nature always find a way to blend in even if it is bugs, birds and weeds. We have this need to micro manage because selective advantage, natural selection, and genetic drift is not trusted in practice. Instead we try to do the morally right thing; ironical isn't it. The concern stems around the environment being stronger than genetics, therefore shifting the life distribution. The historical cataloging doesn't have to walk the walk, just talk the talk.

[foodchain]

Here is a paradox. Most, if not all scientists, like evolution. Humans are creating environmental stresses yet nobody trusts evolution. New nature always find a way to blend in even if it is bugs, birds and weeds. We have this need to micro manage because selective advantage, natural selection, and genetic drift is not trusted in practice. Instead we try to do the morally right thing; ironical isn't it. The concern stems around the environment being stronger than genetics, therefore shifting the life distribution. The historical cataloging doesn't have to walk the walk, just talk the talk.

 

I get lost on this second paragraph. Do you mean by history really phylogeny? It would seem if biology could be reduced at all to one single field with the best possible ability to retain biology as a field of science it would be biological evolution. This is what I don’t understand, plus really looking at evolution or the idea that life really goes extinct more then it thrives, or at least in time a majority of life that has existed on earth in some particular form such as a specie has gone extinct. So really how would you even logical equate that past objective understanding really?

 

So I do not see even how to start thinking about if we are or not are smarter then natural selection. As far as I know the only way you could tell would be to attempt such as see if it could remain fit in time regardless of change in any form. I don’t think we can guarantee the future globally down to each individual specie, so really all I could desire from GM is work towards ending human disease really, I don’t really even think I would enjoy its scope of influence including the endangered species list or conservation.

 

Overall if you cant remove natural selection massive scale GM would entail at least to me that humanity always would have to know what its doing and where its going with that. I just don’t see us fit to do such really, currently at least.

[PhDP]

The problem is;

 

- Natural selection (and this is a strong argument against creationism) cannot plan in advance. It can push species in evolutionary dead ends, lead to the fixation of allele which would be deleterious in the long run, et cetera...

 

- Natural selection is not always there, it can be overwhelmed by drift, draft, mutational pressure...

 

So I do think we should look at genome and modify it and perhaps we could start producing our own vitamin C again.

 

"Darwinian selection is the ONLY method for gettting design. The issue is whether it happens within a brain or outside of one. Human design -- including that of the DNA sequences to insert -- is Darwinian selection.

 

Technically speaking, this is not true, but it's fair to say that Darwinian selection is probably involved in pretty much all cases of adaptation.

[lucaspa]

Either way, I disagree. At best, NS has had a few billion years more time to work than we have. If you gave humanity 2 billion years, we could outdo anything that natural selection has done. And, of course, NS only works on things that are self-replicating.

 

It's not the time. It's that NS keeps track of hundreds/thousands of variables and balances cost/benefit analyses of each. Humans don't do that. When you talk of gengineering humans, you pick only 3-4 traits you want to change.

 

The point of the genetic algorithm was not to make it identical to NS. It was to allow computers to design things with almost no human inputs. The most notable difference is that genetic algorithms have an arbitrary, human-designed fitness algorithm, whereas NS has only reproductive capability of a self-replicating thing in the real world as its fitness algorithm.

 

No,the point was to make the genetic algorithm like NS. The human input is the environment in which the selection takes place. In nature, NS also has a "fitness algorithm". It's sometimes described as the "fitness peak" by evolutionary biologists. Remember, the individuals who survive and replicate have the best designs in that generation. Designs to fit the design problem posed by the environment.

 

Also differenare how variability is introduced (also arbitrary), how reproduction occurs (also arbitrary), and how selection occurs (also arbitrary).

 

Selection is not arbitrary. That's the whole point of natural selection. The individuals who do best in the competition for scarce resources are the ones selected.

 

What I am saying is that natural selection is a specialized type of genetic algorithm.

 

And here I thought you were arguing that genetic algorithms are not natural selection! Now you say they are. I wish you'd make up your mind. Rather than this, I would say that both genetic algorithms and natural selection are Darwinian selection in their respective substrates.

 

Do you have any evidence for that? I don't think that I use Darwinian selection for 1+1=2, which is something used for many designs. I don't use Darwinian selection for using the equations of physics.

 

Think back to when you were first learning math. There were many possible answers you could provide for 1 + 1 = 2 (and you can find many of those on tests of 1st graders!) but the environment meant that you selected "2" each time because that was the "correct" answer. Now, of course, this is memorized and so you select that answer every time. When you were learning physics, you did have to select the correct physics equation in the appropriate environment -- problem you were facing. And, I bet, on some tests you chose the wrong equation for some of the problems. But then, you had already done some selecting in the environment of physics problems as you did practice problems. So now, in a particular environment, you go back to the selection that worked before.

 

I won't agree that we design by natural selection until you provide some proof. Our imagination is not completely random, which is why it is better than natural selection. That's what gives us the ability to plan. We can recognize when we have part of a problem solved and then work on the missing parts. And we can design computers and computer programs to do modeling, trial and error, etc. for us.

 

Think about a baby. It makes random noises until it hits upon variations that elicit a positive response in the environment -- parents and other people. "mama" and "dada" are usually the first words. That environment might be provided by a parent. But when the baby in the variations hits the correct sound, it gets selected. And that continues as more and more words are learned. Many are initially mispronounced. My daughter did "elphedent" for a few days. Positive environment at home but a different environment at day-care. So she selected "elephant" instead. Remember, Darwinian selection is cumulative. It builds. So by the time you hit school you already have selected a basic vocabulary. Now comes new selection of proper spelling and grammar. So as I write these sentences I am building upon all the cumulative selection that has gone before as I choose the sentences I make in my imagination to say what I want to say (the environment I am selecting in).

 

As I said, Darwinian selection is cumulative. So yes, when part of the design fits the environment we stop tinkering with it. So does natural selection.

 

By inefficient, I mean that NS wastes many computations. Not that the results are inferior, but that there are ways to get similar results with less computations.

 

And how fast can our brains compute? Also, I think you think we have to do everything from scratch. In designing these sentences, I already have the grammar and spelling down. I even have previous sentences I have written to get similar meanings down. I don't have to go back to random variations of sound, but just variations in the words and order from a limited set of them.

 

Yes, I never said that NS was incompetent, only inefficient.

 

Good to hear. And I said that NS is more competent than we are.

 

Whatever the reason, NS can only work on the variability that is created by mutations. And the generation of variability is very slow, compared to that created by either genetic algorithms. Genetic engineering is rather slow to create variability, but it is almost always useful variability as opposed to mostly neutral or bat variability generated by mutation.[/qutoe]

 

NS also works on variations generated by recombination, and that is very large. And no, in terms of generations, that generation of variability is not that slow. That's a myth. So far, genetic engineering has been a notable failure. So if you are going on past performance, the clinical trials have been stopped because the genetic engineering has not been useful.

 

But again, with genetic engineering "useful" is defined by you. And you are not the best judge.

 

Which is part of natural selection. As I said, very inefficient. If you don't want deleterious mutations to be fixed by your selection algorithm, then you need a different selection algorithm.

 

No, genetic drift is different from natural selection. Genetic drift is pure chance. When the ways that a Hardy-Weinberg equilibrium can be disturbed, genetic drift is separate from natural selection: which it is. A selection algorithm will NOT fix a deleterious mutation.

 

Yes, and they can vary by portion of the genome too. I never said they didn't vary, I said that the mutation rate in evolution is far, far lower than that in genetic algorithms in general.

 

But a lot of variation comes by recombination. However, saying that genetic algorithms can have a higher variation rate is a difference in degree, not kind. It doesn't make genetic algorithms qualitatively different from natural selection.

 

But not always. Sometimes an accident will eliminate an individual or even an entire population with a beneficial mutation. In a genetic algorithm you can ensure that the best permutations are always preserved.

 

Perhaps, but again a quantitative difference, not a qualitative one. And, as I pointed out in another thread, if the population is "large" (over 100) it is unlikely that beneficial variations will be lost. As we know, 100 is not all that large for any population.

 

Yes, whereas in genetic algorithms, they can be adjusted for optimum results.

 

So? Quantitative difference, not qualitative.

 

Humans set up the algorithms for fitness, creation of new permutations, mutation, genetic exchange, selection, population size, etc. Humans can also insert some non-random permutations. After that, it's all on its own.

 

Which means that the designing is all on its own! Which is the point. Humans aren't doing the designing, Darwinian selection is. The cases I've seen, humans have not done any interferring by non-random permutations.

 

Ah, I was using genetic algorithm as the general term, and natural selection as a specific implementation of it. This seems correct, because a genetic algorithm has arbitrary algorithms for fitness, creation of new permutations, mutation, genetic exchange, selection, population size. Evolution and natural selection have these algorithms, but they are specific rather than arbitrary. Were you using Darwinian selection for this?

 

I'm saying that both genetic algorithms and natural selection are substrates of Darwinian selection.

 

Correct. And despite putting forth only a miniscule fraction of human ingenuity on genetic engineering, and for that matter, despite it being a very new science, we are already making huge improvements in various crops. More importantly, natural selection is often at odds with our own objectives.

 

Whereupon I question our objectives. I was thinking of genetic engineering in humans. That has been a failure so far. What you call "improvements" are exactly the problem I am raising for humans: those crops do exactly what we want in a very controlled environment. Change the environment and those crops are toast. Even then, some of our genetically engineered crops are susceptible to diseases we didn't anticipate. So yes, genetically engineer corn or cotton because, if those varieties go extinct, we haven't lost anything. But do this to ourselves and the environment changes and we lose the human species. OUR species! The risk I'll accept for crop species is not a risk I think we should entertain for our own species.

 

And I meant it in both ways. And when we genetically engineer humans to be smarter, we'll be even smarter than natural selection.

 

Don't think so. Again, remember Thompson. There are some things humans are not good at. We can't keep track of the thousands of cost/benefit relationships that interact. We can't even get a computer to do it, much less our brains. We can modify some few traits in plants and animals but, again, if we lose those it's OK. Losing the human species to your hubris is just too much to risk.

 

If you mean that we select the best designs we can come up with, than yes. If you mean that we create random designs and evaluate them, then no.

 

This is the limitation of human imagination I was referring to. We may not make totally random designs in our imagination. And that may be our undoing because we miss good designs due to our limited imaginations. Our best human designers are the ones that can think "out of the box", right? That means they are ones that come closest to making "random" designs. Ones that more conventional people don't. Darwinian selection, by using "random" variations, is more likely than us to hit upon really novel designs. Which may be why we turn to Darwinian selection when we are stumped.

 

I've addressed this before. We fall back to trial and error when we don't know a better way to proceed. Your argument contradicts itself -- we use trial and error as a last resort, because it works on any kind of problem but is extremely inefficient. But we prefer more efficient ways to solve problems if they work.

 

As for why we use genetic algorithms rather than a human-like intelligence, it is because they are much simpler

 

No, it's because the other method does not work, as you noted. It has nothing to do with efficiency and everything to do with competence. When our Darwinian selection doesn't work we turn to genetic algorithms. In terms of "efficiency", you don't know how fast our brains operate and therefore how many variations we flash thru in a second in our own Darwinian selection.

 

Lies. If we didn't know how it worked, it could not have been created by a genetic algorithm designed by us. The fitness function would have said it wouldn't work, and it would have been discarded.

 

You should be careful about throwing around that word. Especially when you haven't read the article! The "fitness function" was recognition of a spoken word! Not a mathematical function, but an actual operation in the real world. Here, I'll quote from the article:

 

"Thompson realised that he could use a standard genetic algorithm to evolve a configuration program for an FPGA and then test each new circuit design immediately on the chip. He set the system a task that appeared impossible for a human designer. Using only 100 logic cells, evolution had to come up with a circuit that could discriminate between two tones, one at 1 kilohertz and the other at 10 kilohertz.

 

To kick off the experiment, Thompson created a population of 50 configuration programs on a computer, each consisting of a random string of 1s and 0s. The computer downloaded each program in turn to the FPGA to create its circuit and then played it the test tones (see Diagram, below). ... By generation 2800, the fittest circuit was discriminating accurately between the two inputs, but there were still glitches in its output. These only disappeared completely at generation 4100. After this, there were no further changes.

 

Once the FPGA could discriminate between the two tones, it was fairly easy to continue the evolutionary process until the circuit could detect the more finely modulated differences between the spoken words "go" and "stop".

 

So how did evolution do it? If a human designer, steeped in digital lore, were to tackle the same problem, one component would have been essential--a clock. The transistors inside a chip need time to flip between on and off, so the clock is set to keep everything marching in step, ensuring that no transistor produces an output between 0 and 1. A human designer would also use the clock to count the number of ticks between the peaks of the waves of the input tones. There would be 10 times as many ticks between the wave peaks of the 1 kilohertz tone as those of the 10 kilohertz tone.

 

In order to ensure that his circuit came up with a unique result, Thompson deliberately left a clock out of the primordial soup of components from which the circuit evolved. Of course, a clock could have evolved. The simplest would probably be a "ring oscillator"--a circle of cells that change their output every time a signal passes through. It generates a sequence of 1s and 0s rather like the ticks of a clock. But Thompson reckoned that a ring oscillator was unlikely to evolve because it would need far more than the 100 cells available.

 

So how did evolution do it--and without a clock? When he looked at the final circuit, Thompson found the input signal routed through a complex assortment of feedback loops. He believes that these probably create modified and time-delayed versions of the signal that interfere with the original signal in a way that enables the circuit to discriminate between the two tones. "But really, I don't have the faintest idea how it works," he says. ... That repertoire turns out to be more intriguing than Thompson could have imagined. Although the configuration program specified tasks for all 100 cells, it transpired that only 32 were essential to the circuit's operation. Thompson could bypass the other cells without affecting it. A further five cells appeared to serve no logical purpose at all--there was no route of connections by which they could influence the output. And yet if he disconnected them, the circuit stopped working.

It appears that evolution made use of some physical property of these cells--possibly a capacitive effect or electromagnetic inductance--to influence a signal passing nearby. Somehow, it seized on this subtle effect and incorporated it into the solution."

 

"It wasn't just efficient, the chip's performance was downright weird. The current through the chip was feeding back and forth through the gates, "swirling around," says Thompson, and then mov-ing on. Nothing at all like the ordered path that current might take in a human-designed chip. And of the 32 cells being used, some seemed to be out of the loop. Although they weren't directly tied to the main circuit, they were affecting the per-formance of the chip. This is what Thompson calls "the crazy thing about it"

Thompson gradually narrowed the possible explanations down to a handful of phenomena. The most likely is known as electromagnetic coupling, which means the cells on the chip are so close to each other that they could, in effect, broadcast radio signals between them-selves without sending current down the interconnecting wires. Chip designers, aware of the potential for electromag-netic coupling between adjacent compo-nents on their chips, go out of their way to design their circuits so that it won't af-fect the performance. In Thompson's case, evolution seems to have discovered the phenomenon and put it to work.

It was also possible that the cells were communicating through the power-supply wiring. Each cell was hooked in-dependently to the power supply; a rapidly changing voltage in one cell would subtly affect the power supply, which might feed back to another cell. And the cells may have been communi-cating through the silicon substrate on which the circuit is laid down. "The cir-cuit is a very thin layer on top of a thicker piece of silicon," Thompson ex-plains, "where the transistors are dif-fused into just the top surface part. It's just possible that there's an interaction through the substrate, if they're doing something very strange. But the point is, they are doing something really strange, and evolution is using all of it, all these weird effects as part of its system."

 

Notice that Thompson has only "most likely" explanation. He doesn't know.

 

Now, I'll let the ad hominem pass -- this once. Next time you use the word "liar" I'll report you.

 

To be faster, stronger, and smarter, of course. And maybe glow in the dark. And maybe have gills. Maybe redesign ourselves to live in zero gravity. Whatever we want, really. I don't want to wait around for evolution/natural selection to do its thing.

 

I noticed you used the word "want". I asked why we needed genetic changes. You responded with want. Different things. We don't need to be "faster, stronger, smarter, glow in the dark, have gills, redesign to live in zero gravity". We have technology that will do all that. IOW, you want to play god with the human species. You want to shape them in your image -- because you want to. Never mind the good of the species.

 

Anyhow, we can't really pre-empt natural selection. We will just provide very good variability as compared to background mutation.

 

How naive! What you are doing is replacing ALL the alleles with the ones you want. You are reducing variation, not increasing it. And again, "very good variability" is only within your narrow parameters of what is "good".

 

I've yet to see any of that data. I think it is just your opinion.

 

Other examples of where Darwinian selection was used when the design problem was too tough for humans to solve:

1. MJ Plunkett and JA Ellman, Combinatorial chemistry and new drugs. Scientific American, 276: 68-73, April 1997. Summary of article: "By harnessing the creative power of Darwinian selection inside a test tube, chemists can now discover compounds they would not have known how to make. The key is combinatorial chemistry, a process that allows them to produce and screen millions of candidate molecules quickly and systematically."

2. GF Joyce, Directed molecular evolution. Scientific American 267: 90-97,July 1994.

3. AI Samuel, Some studies on machine learning using the game of checkers. IBM Journal of Research Development, 3: 211-219, 1964. Reprinted in EA Feigenbaum and J Feldman, Computers and Thought, McGraw-Hill, New York, 1964 pp 71-105.

6. CW Petit, Touched by nature: putting evolution to work on the assembly line. US News and World Report, 125: 43-45, July 27, 1998. Use "genetic algorithms" (cumulative selection) to get design in industry. Boeing engineers had cumulative selection design a wing forthem for a jet to carry 600 passengers but have a wing the same size as a 747.

9. FS Santiago, HC Lowe, MM Kavurma, CN Chesterman, A Baker, DG Atkins,LM Khachigian, New DNA enzyme targeting Egr-1 mRNA inhibits vascular smooth muscle proliferation and regrowth after injury. Nature Medicine 5:1264-1269, 1999. Used Darwinian selection to design a DNA enzyme (not found in nature) that degrades mRNA for use in treating hyperplasia after balloon arthroplasty. Humans have no idea what the nucleotide sequence of the DNA enzyme because they didn't make it --Darwinian selection did.

10. Breaker RR, Joyce GF.A DNA enzyme that cleaves RNA. Chem Biol 1994 Dec;1(4):223-9

11. Ronald R Breaker, Gerald FA Joyce DNA enzyme with Mg2+-dependent RNA phosphoesterase activity Chemistry & Biology 1995, 2:655-660.

 

You'll especially like this one:

13. http://www.discover.com/aug_03/gthere.html?article=feattech.html Use of Darwinian selection to evolve of the ability to think in computers. If humans think so well, why use Darwinian selection to get the ability to think in computers?

 

14. Jr Koza, MA Keane, MJ Streeter, Evolving inventions. Scientific American, 52-59, Feb 2003 check out http://www.genetic-programming.com

15. A. Thompson, P. Layzell and R. S. Zebulum Explorations in Design Space: Unconventional electronics design through artificial evolution. IEEE Trans. Evol. Comp., Vol 3, No 3, (1999) http://www.informatics.sussex.ac.uk/users/adrianth/TEC99/paper.html

 

Do mutations also reduce genetic variability? Eugenics, artificial selection, and natural selection decrease variability. Mutation and genetic engineering increase it.

 

Natural selection promotes variation. You need to read Chapters 14, 15, and 22 in Futuyma's Evolutionary Biology for all the details. Or start a new thread. Genetic engineering decreases variations within populations. After all, every individual has the same genetically engineered alleles.

 

I do not think everyone should have the same alleles. I think we should create new alleles ourselves rather than wait for background mutation to do it.

 

Please quit saying I approve of removing our genetic variability.

 

Then you are not promoting genetic engineering. Tell me, do genetically engineering crops have more variation than wild type? Of course not. Because the alleles in every individual is replaced.

 

If you are going to introduce alleles for "faster, smarter, stronger", then isn't everyone going to have them? Would you restrict them to a few? How would you do that? If not restricted, then everyone would have to have them in order to be equal in the society, wouldn't they? Wouldn't there be job discrimination against those individuals without the "smarter" alleles?

 

So what do you plan, several subpopulations of humans? One with gills, one with alleles for low gravity, one that can glow in the dark?

 

And I can see us not engineering ourselves and still going extinct, like most species have. Or, not engineering ourselves and then being outcompeted/exterminated by humans who have engineered themselves to be stronger, faster, or smarter and decided that they wanted more land.

 

Then there goes your "increased variation" argument, doesn't it? Everyone would have the alleles. Less variation. Yes, it's possible we might go extinct. But the odds are much, much less if we let natural selection keep the variation -- and thus our ability to use lots of environments -- rather than engineer us into one niche. So we are all adapted to zero g and we lose space travel. How do those individuals live in a gravity well?

 

I've said this before and you didn't respond. To get smarter means a bigger brain or more brain cells packed into a smaller volume. Either way, the brain's requirement for energy would go up. That's OK now in our environment of plenty of food. But suppose global warming does play havoc with crops and suddenly we have a lot less food. Now we have everyone needing more calories per day than they did and the calories are not available. Congrats. You have put H. sapiens into a situation where it starves! Whereas if we had NOT gene engineered, there would have been a few people who may be "dumb" but could survive on fewer calories and thus, H. sapiens survives.

 

The problem is;

 

- Natural selection (and this is a strong argument against creationism) cannot plan in advance. It can push species in evolutionary dead ends, lead to the fixation of allele which would be deleterious in the long run, et cetera..

 

1. Natural selection does not plan long term, but neither is the genetic engineering being advocated by Mr. Skeptic.

2. NS does have a short term plan -- adaptation to the environment. Since it can't predict future environments, there is no long term plan.

3. The argument is that NS is less likely to push H. sapiens into an evolutionary dead end than genetic engineering is. Where we have used genetic engineering on crops, we are pushing them to an evolutionary dead end. We keep them going by our technology.

 

Natural selection is not always there, it can be overwhelmed by drift, draft, mutational pressure..

 

Natural selection is always there. Like you say drift is always there. I know of no case where natural selectionis "overwhelmed" by drift. Even in founder effects, where N=2, natural selection still works to adapt the population.

 

Technically speaking, this is not true, but it's fair to say that Darwinian selection is probably involved in pretty much all cases of adaptation.

 

Darwinian selection is being used for the process whereever it occurs. As Dennet shows in Darwin's Dangerous Idea, Darwinian selection is an algorithm to get design. Natural selection is one application of Darwinian selection and, yes, it is responsible for all adaptations (which are designs). I am claiming that human design is Darwinian selection operating within a brain.

[Mr Skeptic]

It's not the time. It's that NS keeps track of hundreds/thousands of variables and balances cost/benefit analyses of each. Humans don't do that. When you talk of gengineering humans, you pick only 3-4 traits you want to change.

 

That's not true. Humans can keep track of many different things at once, as evidenced by the complicated things we build all the time. Again, I am not advocating we replace natural selection, only that we create new alleles that random mutation would not be able to do. Natural selection can still have its say.

 

Just to keep you happy -- we are perfectly capable of using genetic algorithms to help us design humans.

 

 

No,the point was to make the genetic algorithm like NS. The human input is the environment in which the selection takes place. In nature, NS also has a "fitness algorithm". It's sometimes described as the "fitness peak" by evolutionary biologists. Remember, the individuals who survive and replicate have the best designs in that generation. Designs to fit the design problem posed by the environment.

 

Yes, which makes the genetic algorithm a more general algorithm than NS.

 

Selection is not arbitrary. That's the whole point of natural selection. The individuals who do best in the competition for scarce resources are the ones selected.

 

The fitness function in a genetic algorithm is arbitrary, whereas NS has only one fitness function. Which makes the genetic algorithm a more general algorithm than NS.

 

And here I thought you were arguing that genetic algorithms are not natural selection! Now you say they are. I wish you'd make up your mind. Rather than this, I would say that both genetic algorithms and natural selection are Darwinian selection in their respective substrates.

 

Natural selection is a type of genetic algorithm. Not the other way around. Like a tiger is a feline, not the other way around. One is a more general category.

 

Think back to when you were first learning math. There were many possible answers you could provide for 1 + 1 = 2 (and you can find many of those on tests of 1st graders!) but the environment meant that you selected "2" each time because that was the "correct" answer. Now, of course, this is memorized and so you select that answer every time. When you were learning physics, you did have to select the correct physics equation in the appropriate environment -- problem you were facing. And, I bet, on some tests you chose the wrong equation for some of the problems. But then, you had already done some selecting in the environment of physics problems as you did practice problems. So now, in a particular environment, you go back to the selection that worked before.

 

No, when I was learning to add, I added on my fingers. Then I counted how many fingers there were. Addition in base one is extremely simple.

 

Think about a baby. It makes random noises until it hits upon variations that elicit a positive response in the environment -- parents and other people. "mama" and "dada" are usually the first words. That environment might be provided by a parent. But when the baby in the variations hits the correct sound, it gets selected. And that continues as more and more words are learned. Many are initially mispronounced. My daughter did "elphedent" for a few days. Positive environment at home but a different environment at day-care. So she selected "elephant" instead. Remember, Darwinian selection is cumulative. It builds. So by the time you hit school you already have selected a basic vocabulary. Now comes new selection of proper spelling and grammar. So as I write these sentences I am building upon all the cumulative selection that has gone before as I choose the sentences I make in my imagination to say what I want to say (the environment I am selecting in).

 

Partly true. However, babies primarily learn by copying.

 

And how fast can our brains compute? Also, I think you think we have to do everything from scratch. In designing these sentences, I already have the grammar and spelling down. I even have previous sentences I have written to get similar meanings down. I don't have to go back to random variations of sound, but just variations in the words and order from a limited set of them.

 

Babies do not go through a phase where they string random words together. Their grammar is very crude, but they make intelligible requests, not random words.

 

Good to hear. And I said that NS is more competent than we are.

 

Again I disagree. NS is slow but competent. We are fast, and more competent. For example, we have the ability to plan ahead. We can see patterns. We can look at our billions of base pairs and figure out which changes lead to which medical conditions. We have the ability to copy good ideas (eg to take useful genes from other species -- again to make you happy, these were created by natural selection).

 

NS also works on variations generated by recombination, and that is very large. And no, in terms of generations, that generation of variability is not that slow. That's a myth. So far, genetic engineering has been a notable failure. So if you are going on past performance, the clinical trials have been stopped because the genetic engineering has not been useful.

 

Sexual type recombination (including plasmids) does not generate any new alleles, and that is the only type that is rapid. Consider that much of our genetic engineering is recombination of genes from unrelated species -- which, to make you happy, were created by natural selection.

 

But again, with genetic engineering "useful" is defined by you. And you are not the best judge.

 

We better let the process that let 99% of species go extinct be the judge then. How can you put your faith in that? Natural selection can neither prepare for the future, nor learn from the past. We can.

 

No, genetic drift is different from natural selection. Genetic drift is pure chance. When the ways that a Hardy-Weinberg equilibrium can be disturbed, genetic drift is separate from natural selection: which it is. A selection algorithm will NOT fix a deleterious mutation.

 

You can say it is different if you like. But you can't have natural selection without genetic drift.

 

But a lot of variation comes by recombination. However, saying that genetic algorithms can have a higher variation rate is a difference in degree, not kind. It doesn't make genetic algorithms qualitatively different from natural selection.

 

As I said, the genetic algorithms we use to design are more efficient, and abstract, compared to natural selection. They're designed to be efficient.

 

Also, a difference is degree is significant if you are arguing about which is smarter. Quantitative differences are important too!

 

Perhaps, but again a quantitative difference, not a qualitative one. And, as I pointed out in another thread, if the population is "large" (over 100) it is unlikely that beneficial variations will be lost. As we know, 100 is not all that large for any population.

 

No, this one is a qualitative difference. Genetic algorithms need not have accidents, but natural selection will have accidents. Even if a beneficial allele is established in a large population, accidents can still remove it.

 

Which means that the designing is all on its own! Which is the point. Humans aren't doing the designing, Darwinian selection is. The cases I've seen, humans have not done any interferring by non-random permutations.

 

You forget that humans designed the designer.

 

Non-random permutations can be used to optimize an already existing design.

 

I'm saying that both genetic algorithms and natural selection are substrates of Darwinian selection.

 

Fair enough. Then what I am calling genetic algorithms you are calling Darwinian selection.

 

Whereupon I question our objectives. I was thinking of genetic engineering in humans. That has been a failure so far. What you call "improvements" are exactly the problem I am raising for humans: those crops do exactly what we want in a very controlled environment. Change the environment and those crops are toast. Even then, some of our genetically engineered crops are susceptible to diseases we didn't anticipate. So yes, genetically engineer corn or cotton because, if those varieties go extinct, we haven't lost anything. But do this to ourselves and the environment changes and we lose the human species. OUR species! The risk I'll accept for crop species is not a risk I think we should entertain for our own species.

 

But the situation would be completely different. Crops are so vulnerable mostly because they are a monoculture, not because of any engineering. Monocultures have been vulnerable way before genetic engineering existed.

 

We would never allow humans to become a genetic monoculture. Also, the original alleles should and will be preserved, be it in poor people who cannot afford genetic enhancement, people who are opposed to genetic engineering for religious, cultural, or scientific reasons, sperm banks, etc.

 

Eugenics is a completely separate issue from genetic engineering.

 

Don't think so. Again, remember Thompson. There are some things humans are not good at. We can't keep track of the thousands of cost/benefit relationships that interact. We can't even get a computer to do it, much less our brains. We can modify some few traits in plants and animals but, again, if we lose those it's OK. Losing the human species to your hubris is just too much to risk.

 

There are things natural selection isn't good at either. Think computers (programmable with negligible calculation error), nuclear power, space travel/habitation, supersonic travel, etc. Think "irreducible" complexity, or if you prefer, "finding non-local maxima". Natural selection has a strong pressure to remain at a local maxima, which prevents it from seeking out a non-local maxima which may be far better.

 

In fact if it wasn't for the genetic drift which you seem to despise, natural selection would be stuck at a local maxima until environmental conditions moved the local maxima.

 

This is the limitation of human imagination I was referring to. We may not make totally random designs in our imagination. And that may be our undoing because we miss good designs due to our limited imaginations. Our best human designers are the ones that can think "out of the box", right? That means they are ones that come closest to making "random" designs. Ones that more conventional people don't. Darwinian selection, by using "random" variations, is more likely than us to hit upon really novel designs. Which may be why we turn to Darwinian selection when we are stumped.

 

Thinking outside of the box has its limitations -- notably it is less efficient (but has potential to be more competent). We simply cannot afford to think outside the box all the time, as we would get nothing done. However, thinking outside the box once in a while can make the box bigger for everyone.

 

Us humans are such ingenious creatures that when we run out of calculating capacity or creativity, we can turn around and design a computer and computer program to compensate for it.

 

No, it's because the other method does not work, as you noted. It has nothing to do with efficiency and everything to do with competence. When our Darwinian selection doesn't work we turn to genetic algorithms. In terms of "efficiency", you don't know how fast our brains operate and therefore how many variations we flash thru in a second in our own Darwinian selection.

 

Trial and error is a last resort because it works for all cases but is far less efficient. It is "last" because it is less efficient, and a "resort" because it works.

 

You should be careful about throwing around that word. Especially when you haven't read the article! The "fitness function" was recognition of a spoken word! Not a mathematical function, but an actual operation in the real world. Here, I'll quote from the article:

 

[...]

 

Now, I'll let the ad hominem pass -- this once. Next time you use the word "liar" I'll report you.

 

A very interesting article, thank you. I'd like to point out that there are no "physics we humans haven't discovered yet" mentioned anywhere. The new principles are principles of design, not of physics, as far as I can tell.

 

Another flaw was that the result was based on the specific FPGA he used, rather than the circuit in general. Had he used several FPGA's he might have gotten a circuit design rather than a design for that specific FPGA.

 

As for the comment, "Lies." That was directed at whatever lazy or sensationalist reporter replaced "principles of design" with "principles of physics", not at you or Thompson.

 

I noticed you used the word "want". I asked why we needed genetic changes. You responded with want. Different things. We don't need to be "faster, stronger, smarter, glow in the dark, have gills, redesign to live in zero gravity". We have technology that will do all that. IOW, you want to play god with the human species. You want to shape them in your image -- because you want to. Never mind the good of the species.

 

Sometimes there is a very fine line between "want" and "need". I used the term "want" in the same sense that people "want" electricity, clean running water, and nutritious food. I do not want to play god with the human species. I don't want anyone else playing god with the human species. I just want people to have the freedom to change their genetic makeup. The degree to which it will be necessary or desirable in the future is unknown. The Amish choose to go without much of modern technology, but that puts them at a severe disadvantage for most jobs (any that require transportation, new technology, power tools).

 

How naive! What you are doing is replacing ALL the alleles with the ones you want. You are reducing variation, not increasing it. And again, "very good variability" is only within your narrow parameters of what is "good".

 

How is creating a new allele the same as destroying the others?

 

Natural selection promotes variation. You need to read Chapters 14, 15, and 22 in Futuyma's Evolutionary Biology for all the details. Or start a new thread. Genetic engineering decreases variations within populations. After all, every individual has the same genetically engineered alleles.

 

No, natural selection decreases variability, and genetic engineering increases it. That is a tautology, you can't argue against it. Hint: the permutations that natural selection selects against and eliminates are variability. Less permutations == less variability. Genetic engineering creates new permutations. More permutations == more variability.

 

The fact that people artificially select in favor of genetically engineered genes is also a selection process that reduces variability, but it is distinct from the actual creation of new, genetically engineered genes.

 

Then you are not promoting genetic engineering. Tell me, do genetically engineering crops have more variation than wild type? Of course not. Because the alleles in every individual is replaced.

 

Yes, they do. However, the genetically engineered genes are artificially selected for, to the point of eliminating the other genes.

 

If you are going to introduce alleles for "faster, smarter, stronger", then isn't everyone going to have them? Would you restrict them to a few? How would you do that? If not restricted, then everyone would have to have them in order to be equal in the society, wouldn't they? Wouldn't there be job discrimination against those individuals without the "smarter" alleles?

 

No. The new genes will come at a cost, one of which will be increased risk. Many people will not have them. Discrimination will be a problem, but isn't that also a problem with people who are "naturally" smarter?

 

So what do you plan, several subpopulations of humans? One with gills, one with alleles for low gravity, one that can glow in the dark?

 

Possibly. We could end up with a society of merfolk, one of spacefarers. and one of glowy cave people. Or the glowing could become a form of communication replacing speech via sound with speech via light (though I included glowing as an example of a silly modification).

 

It would be sad if we fragmented into separate societies, however, as it could end in a giant war.

 

Then there goes your "increased variation" argument, doesn't it? Everyone would have the alleles. Less variation. Yes, it's possible we might go extinct. But the odds are much, much less if we let natural selection keep the variation -- and thus our ability to use lots of environments -- rather than engineer us into one niche. So we are all adapted to zero g and we lose space travel. How do those individuals live in a gravity well?

 

The society that allows genetically engineered humans would have both their original alleles, and the new ones -- more variety. The only way it could end in less variety would be if one genetically engineered allele replaced several natural ones, which without artificial selection would only happen if they were superior.

 

I've said this before and you didn't respond. To get smarter means a bigger brain or more brain cells packed into a smaller volume. Either way, the brain's requirement for energy would go up. That's OK now in our environment of plenty of food. But suppose global warming does play havoc with crops and suddenly we have a lot less food. Now we have everyone needing more calories per day than they did and the calories are not available. Congrats. You have put H. sapiens into a situation where it starves! Whereas if we had NOT gene engineered, there would have been a few people who may be "dumb" but could survive on fewer calories and thus, H. sapiens survives.

 

Or we could improve the ion pumps in the body. We use an obscene amount of energy to pump ions, especially in neurons. An improved ion pump could result in both more intelligence and lower food requirements. And again, not everyone would necessarily have the new genes.

 

----

 

I am in favor of creating new alleles via genetic engineering, not of replacing natural alleles in the entire population with a single genetically engineered allele.

 

Lucaspa, how can you keep insisting that creation of new alleles implies the elimination of all the other ones?

[foodchain]

 

I am in favor of creating new alleles via genetic engineering, not of replacing natural alleles in the entire population with a single genetically engineered allele.

 

Lucaspa, how can you keep insisting that creation of new alleles implies the elimination of all the other ones?

 

 

Because I think you are confused a little bit. The traits or the phenotype or whatever of those new alleles will be treated the same as normie ones. If you produce a new allele do you think adding it into a population would somehow produce a result where it was not processed by NS at all?

 

If you were to do it in a fashion in which the new allele would have noting to do with meiotic cells or meiosis in general. I could not imagine currently how you would pull that off. Basically I think you would have to master morphogenesis and development so you could insert such on the fly, or the allele would have to be tagged in some way as to avoid ending up in some genome as hereditary material, if you could subtract that you could attack one way NS impacts, but you would still be left with the impact it would hold via concentration in individuals of a population right? Simply put even though its not occurring naturally such genes themselves would be occurring regular in generations of populations.

[lucaspa]

I am in favor of creating new alleles via genetic engineering, not of replacing natural alleles in the entire population with a single genetically engineered allele.

 

Lucaspa, how can you keep insisting that creation of new alleles implies the elimination of all the other ones?

 

You answered that one yourself later in your post: "However, the genetically engineered genes are artificially selected for, to the point of eliminating the other genes." You think the new alleles are going to be better: you call them "genetic enhancement" The inevitable result of genetic engineering humans is that, in order not to be inferior to other humans, each parent is going to engineer their children with the new alleles. So within a few generations everyone is going to have the new alleles -- and a loss of genetic variability.

 

The way to avoid that is to have genetic engineering be so expensive that only a fraction of the population can afford it. You have that solution below. But then you have other problems: you are creating a genetic aristocracy. These people would think they are "better" than the proles and, eventually, change society to reflect their superiority. Of course, you might have a revolt from the proles. In which case, the result would be no genetic engineering.

 

Also, the original alleles should and will be preserved, be it in poor people who cannot afford genetic enhancement, people who are opposed to genetic engineering for religious, cultural, or scientific reasons, sperm banks, etc.

 

So this is your proposed mechanism to ensure genetic variability. The "religious, cultural, or scientific reasons" aren't going to hold under your scenario. In the short term, you are talking "enhancement". I might have objections to my kids being "enhanced", but they are going to see that they don't perform as well as the "enhanced". So they are going to do one of 2 things:

1. Make sure their kids are enhanced.

2. If they can't do that then exert political pressure to have enhancement stopped so that their kids have a level playing field to do well in society.

 

Your reasons (other than poverty) are only going to last one generation.

 

That's not true. Humans can keep track of many different things at once, as evidenced by the complicated things we build all the time.

 

Because the things are complicated doesn't mean that we keep track of all the variables. Look again at what Thompson said. We tend to specialize our artifacts in a few specific traits and then ignore the other interactions.

 

Again, I am not advocating we replace natural selection, only that we create new alleles that random mutation would not be able to do. Natural selection can still have its say.

 

It can't work this way, unless you artificially restrict the number of people who get the new alleles. If you don't think the new alleles are "better", how are you going to get parents to accept something that might be harmful to their lineage in the long run? After all, when the individual doesn't reproduce, not only do the artificial alleles get lost, but all the alleles. So you are asking parents to risk all their genetic heritage on whether the introduced alleles are, in fact, "better".

 

Sorry, but you can't have it both ways. You either have to say that the alleles are "better", in which case everyone wants them, or you say that the new alleles are a crapshoot in which case no rational person is going to want to gamble that way.

 

Just to keep you happy -- we are perfectly capable of using genetic algorithms to help us design humans.

 

If we don't know the complete variables of the environment, how do we set it for genetic algorithms? And we don't. Also, if you do, then you back to decreasing genetic variability. After all, won't you be saying that the humans designed by the genetic algorithm is "better" than the current humans? Then every parent is going to want their children designed to the new standard.

 

The fitness function in a genetic algorithm is arbitrary, whereas NS has only one fitness function. Which makes the genetic algorithm a more general algorithm than NS.

 

The environment of a genetic algorithm is not "arbitrary". It is determined by the function the individual needs to do. NS does not have "one fitness function". Rather, it has many fitness functions, many of which are contradictory. Remember, every trait comes with a cost as well as a benefit. The eyesight of a hawk comes at the cost of finer control of development of the eye. The speed of a cheetah comes at the cost of more energy used. What's the correct balance between speed to catch a prey (to get food for energy) and energy expenditure. If the cheetah expends more energy to catch the prey than it gets from eating the prey, then the cheetah loses overall.

 

No, NS is a more general and complicated algorithm than genetic algorithms. And, in turn, genetic algorithms are more complicated and general than human designing.

 

Natural selection is a type of genetic algorithm.

 

That's quite a reversal for you! You started out saying that genetic algorithms weren't anything like NS. Genetic algorithms are a simulation of NS.

 

Babies do not go through a phase where they string random words together. Their grammar is very crude, but they make intelligible requests, not random words.

 

Babies go thru a phase where they make random sounds. When those random sounds match the environment so that they get positive feedback -- as when they say "da-da", then those sounds are selected for.

"Four months

At this stage, your child will start to babble, combining consonants and vowels (such as "baba" or "yaya"). The first "mama" or "dada" may slip out now and then, and though it's sure to melt your heart, your baby doesn't quite yet equate those words with you. That comes later, when she's almost a year old. Her attempts at talking will sound like stream of consciousness monologues in another language, endless words strung together. http://www.babycentre.co.uk/toddler/development/speechandlanguage/milestonetalking/

 

And no, when children first learn to talk, many of their requests are not intelligible. Ask any parent who has raised kids what it was like trying to understand children between the ages of 1 and 2.

 

NS is slow but competent. We are fast, and more competent.

 

Again, go back and read Thompson. We aren't that competent.

 

We can look at our billions of base pairs and figure out which changes lead to which medical conditions. We have the ability to copy good ideas (eg to take useful genes from other species -- again to make you happy, these were created by natural selection).

 

All this is irrelevant. What we can't do is model the cost vs benefit analysis of a trait with every other trait in the individual or against the environment.

 

Sexual type recombination (including plasmids) does not generate any new alleles, and that is the only type that is rapid.

 

Remember, most traits are polygenic. You are thinking of one allele = one trait. Also, most genes are pleiotrophic: they contribute to more than one trait. So, recombination will produce new traitswithout ever having to generate a new allele. Now, if we do like you say -- genetic engineer an allele from an unrelated species -- how is that allele going to affect not only the trait we want but also all the other traits that it contributes to? That's part of what natural selection can do that humans can't.

 

And, if you say that you want to let natural selection work that out after you introduce the allele, remember that natural selection is pretty hard on the individual. But the individuals here are people. If the allele is harmful, it is analogous to poisoning that individual (which is unethical).

 

We better let the process that let 99% of species go extinct be the judge then. How can you put your faith in that? Natural selection can neither prepare for the future, nor learn from the past. We can.

 

Considering that we are the result of 3.8 billion years of winners, yes. We can't prepare for the future, either, because we have a very imperfect idea what future environments will be like. For instance, 40 years ago would we have predicted HIV?

 

You can say it is different if you like. But you can't have natural selection without genetic drift.

 

Of course you do! Mr. Skeptic, please read an evolutionary biology textbook. Genetic drift and natural selection are different. I think you are confusing that genetic drift starts with a mutation and natural selection can start with a mutation with the idea that you must have genetic drift to have natural selection. Yes, both start with a mutation. But what happens after that is completely different. Under natural selection, the change in frequency of the allele is deterministic and related to the selection coefficient of the allele while in genetic drift the change in frequency of the allele is pure chance.

 

Genetic algorithms need not have accidents, but natural selection will have accidents. Even if a beneficial allele is established in a large population, accidents can still remove it.

 

If a beneficial allele is established in a large population, accident can't remove it. Instead, accident can remove the allele while it is present at low frequency.

 

You forget that humans designed the designer.

 

Stated like a true IDer! All the humans did was simulate natural selection. In genetic algorithms the permutations are "random". That is, humans don't have control over them

 

But the situation would be completely different. Crops are so vulnerable mostly because they are a monoculture, not because of any engineering. Monocultures have been vulnerable way before genetic engineering existed.

 

How does "monoculture" differ from "low genetic diversity"?

 

We would never allow humans to become a genetic monoculture.

 

How would you stop it? As I've said before, if the allele is "better", then every parent is going to be sure that their children get the genetic engineering so that they are not "inferior". Who would you designate to keep the old, inferior alleles and how would you enforce it? As you follow the logic chain of your claims, genetic engineering becomes undesirable from an ethical and societal aspect as well as a biological aspect.

 

There are things natural selection isn't good at either. Think computers (programmable with negligible calculation error), nuclear power, space travel/habitation, supersonic travel, etc. Think "irreducible" complexity, or if you prefer, "finding non-local maxima". Natural selection has a strong pressure to remain at a local maxima, which prevents it from seeking out a non-local maxima which may be far better.

 

First, as you should be aware, there is no "irreducibly complex" structure that can't be reached by natural selection: http://www.cbs.dtu.dk/dave/JTB.html

 

Second, you are trying to compare manufactured artifacts to biological entities that arose by natural selection. There are no evolutionary processes in the environment that will produce computers or nuclear power, machines that travel at supersonic speeds, etc. But that is irrelevant because you are comparing apples and oranges. Natural selection did produce programmable computers with negligible calculation error: the human brain.

 

In fact if it wasn't for the genetic drift which you seem to despise, natural selection would be stuck at a local maxima until environmental conditions moved the local maxima.

 

So? That isn't what we are talking about. We are talking about humans trying to pick alleles that will be at a new "maxima" based on their imperfect knowledge of the environment and the interaction of genes with other genes.

 

BTW, how did you get the idea that I "despised" genetic drift? I just said it was not the same as natural selection. Also, genetic drift will not move a population from one fitness peak to another. Remember, genetic drift is only effective at very small populations -- less than 10 effective individuals. And here the most common result is fixing neutral or even slightly deleterious effects.

 

Thinking outside of the box has its limitations -- notably it is less efficient (but has potential to be more competent). We simply cannot afford to think outside the box all the time, as we would get nothing done. However, thinking outside the box once in a while can make the box bigger for everyone.

 

But you are not thinking "outside the box". You are very much within the traditional box that says "humans are great designers, we will just apply this human ingeniuosness to our own species".

 

Outside the box in this case is realizing that natural selection is a better and more ingenious than we are.

 

Trial and error is a last resort because it works for all cases but is far less efficient. It is "last" because it is less efficient, and a "resort" because it works.

 

As I showed, humans always use trial and error. The question is whether that Darwinian selection occurs in our minds and the trial and error are ideas in our heads or whether it happens out in the physical universe.

 

You are very much in the "box" that views humans as the smartest things around. Out of the box here is that natural selection is smarter than we are. That's why using natural selection to design things was so innovative.

 

I'd like to point out that there are no "physics we humans haven't discovered yet" mentioned anywhere. The new principles are principles of design, not of physics, as far as I can tell.

 

My apologies. That was in a bit I put in ellipses. The problem comes with the circuit getting a clock without the elements to make a clock. To do so, it might be using physics we haven't discovered yet.

 

Another flaw was that the result was based on the specific FPGA he used, rather than the circuit in general. Had he used several FPGA's he might have gotten a circuit design rather than a design for that specific FPGA.

 

First, an FPGA is general. The whole point of an FPGA is that it can change: "the logic gates in an FPGA and their interconnections can be changed at will. The transistors are arranged into an array of "logic cells" and simply by loading a special program into the chip's configuration memory, circuit designers can turn each cell into any one of a number of logic gates, and connect it to any other cell. So by loading first one program, then another, the chip can be changed at a stroke from, say, an amplifier to a modem ("Software, who needs it?", New Scientist, 2 November 1996, p 41).

 

Second, for what I want to demonstrate, that is irrelevant. The point was that natural selection did the designing and, as evidence of that, Thompson doesn't know how it works.

 

I just want people to have the freedom to change their genetic makeup.

 

Do you also want people to have the freedom to own other humans as slaves? "Freedom" does not mean that we should do whatever we want. There are limits on freedom.

 

How is creating a new allele the same as destroying the others?

 

I've told you several times: since you insist that genetic engineering is "enhancement", the result is going to be replacement of all other alleles with the genetically engineered one.

 

No, natural selection decreases variability, and genetic engineering increases it. That is a tautology, you can't argue against it.

 

Sure you can. With data. Remember heterozygote superiority. This has kept both the sickle cell and normal hemoglobin alleles in the population. Also remember polygenic traits and pleiotrophy. This too keeps many alleles in the population. Also remember disruptive selection. In the case of humans, this one is very important. Humans have such a wide range that they face different environments. Natural selection fosters different alleles for the different environments and gene flow keeps them all in the population.

 

Genetic engineering creates new permutations. More permutations == more variability.

 

The fact that people artificially select in favor of genetically engineered genes is also a selection process that reduces variability,

 

THANK YOU for showing how genetically engineered genes are going to reduce variability. Thank you for agreeing with what I have been saying. What you are trying to do is separate the generation and insertion into the population of genetically engineered genes from the artificial selection that you admit it is going to follow. But, when looking at overall genetic variability in the population, you can't separate the two.

 

However, the genetically engineered genes are artificially selected for, to the point of eliminating the other genes.

 

And that answers your question above "how can you keep insisting that creation of new alleles implies the elimination of all the other ones?"

 

The new genes will come at a cost, one of which will be increased risk. Many people will not have them. Discrimination will be a problem, but isn't that also a problem with people who are "naturally" smarter?

 

But now "smarter" is not determined but any family can produce a smart kid. Now you are making a trait deterministic.

 

The society that allows genetically engineered humans would have both their original alleles, and the new ones -- more variety.

 

You contradicted that above. The society that allows genetic engineering is going to have artificial selection for the "enhanced" allele: less variability.

 

The only way it could end in less variety would be if one genetically engineered allele replaced several natural ones, which without artificial selection would only happen if they were superior.

 

It would happen if they were perceived as superior. But then again, why would anyone introduce what they thought was an "inferior" allele? What you have said, and what I have said, is that replacement is inevitable within a few generations -- far too little time to see the possible detrimental effects.

[Mr Skeptic]

I just don't find that convincing. The crux of your argument seems to be that because one person thinks one allele is superior, that everyone else is going to agree and artificially select for that allele at the expense of all others. Yet your argument contradicts its own main point, repeatedly. Your secondary point, that natural selection is smarter than human intelligence, is no more than an argument from personal incredulity. For it to even matter, it depends on your primary point, that genetic engineering must result in people replacing natural selection with artificial selection.

 

For example, as you said, an allele is not likely to be all out better than any other, which counters your point -- some people will like its benefits, but others will not like its drawbacks. So no monoculture. A case in point is our own disagreement: I think an allele that is intelligently designed by us clever humans would be good, you think that an allele that has withstood the test of time would be better. This would work for any genetically engineered alleles compared to natural ones. Even among engineered alleles, different people would choose different alleles, as they might value intelligence, beauty, strength, dexterity, resistance to disease, etc. all differently.

 

How about the following argument:

Premise: You aren't going to genetically "upgrade" your own kids just because I do.

Given above premise is true, my point stands that not everyone will artificially select for new and shiny genetically engineered allele, and you concede your argument that everyone or no one would do so, with yourself and myself as the counterexamples.

If the above premise were true, you concede that you think human ingenuity beats evolution.

 

So long as, as shown above, we do not replace natural selection, we can add to the mutation step of evolution intelligent, human intervention. Since mutations are random, it stands to reason that our intelligence could be better -- it is not hard to do better than random. In addition, there will still be mutations; it's not like genetic engineering will eliminate mutations. If intelligence is better than randomness, then we have improved upon evolution by improving on the mutation step, leaving natural selection unchanged.

[Mokele]

(fair warning: this thread is really long, so I'm just going to jump in)

 

IMHO, the most important point is simply that we don't know enough about how the genome works yet. We don't fully understand all the metabolic pathways and their roles in development, and we're only just beginning to understand how gene regulation works.

 

Of course, technically nature doesn't "know" anything, but over the millennia, it's selected against non-viable phenotypes.

 

So, basically, natural selection isn't really "smarter" or "dumber", but rather it's functional, and until we actually know a lot more than we do now, it's probably best to be cautious in any tinkering we do. However, molcular biology advances very, very fast, so it may not be too long.

[Mr Skeptic]

A fair point, and I agree. I would probably balk if someone wanted to change my Cytochrome C -- a small, ancient, and vital gene, exactly the kind that evolution ought to perfect. On the other hand, if someone wanted to change my hemoglobin, superoxide dismutase, catalase, or some such, which I think have a specific function that could easily be measured, I'd be much more inclined to approve. I'd probably want animal testing done first though.

 

Another point is that with intelligence we could design some biology that would be extremely unlikely to evolve ("irreducibly complex"). An extreme example would be a living creature that can perform fusion reactions -- such a thing would be pretty much impossible for it to evolve, but we could in theory design. A more realistic example would be a tree with a plug in it, that we could put a plug in and get solar generated electricity out. Maybe we could modify some cells to serve as intermediaries between computer chips and neurons. Such things probably wouldn't evolve on their own, but we could in theory design.

[Mokele]

But remember that evolution isn't the only constraint - there's developmental constraints. You can't just *make* a tree with a plug in it, you have to *grow* it, from a seed, and there are certain limits to what the developmental system can and cannot do.

 

Of course, it's possible to alter the developmental system too, but that comes under the 'cytochrome C' example, especially for very old, important systems, like HOX genes.