To Save Time Barry Argues Both Sides

_{Barry Arrington

August 4, 2011

Intelligent Design}

Share: Facebook; Twitter; LinkedIn; Flipboard; Print; Email

In comment [25] to my last post , The ID hypothesis, Elizabeth Liddle asks about information. I think I’ve been at this long enough to predict how an exchange between me and Elizabeth would go.

Barry’s Point 1:
Let’s take the information in your comment [25]. I am sure you will agree your comment contains specified complex information. Indeed, your one little comment contains more specified complex information than we could reasonably attribute to chance and necessity working from the beginning of the universe to this moment.

Barry’s Point 2:
I am sure you will agree that the cells in your body contain more complex specified information than your comment by several orders of magnitude.

Barry’s Question to Elizabeth:
If your comment contains more specified complex information than we could reasonably attribute to chance and necessity working from the beginning of the universe to this moment, and the cells in your body contain several orders of magnitude more complex specified information than your comment, why should we attribute the complex specified information in your body to chance and necessity? Isn’t it more reasonable to attribute the CSI in the cells in your body to intelligent agency, just as we attribute the CSI in comment [25] to intelligent agency?

Elizabeth’s Probable Response 1:
The CSI in the cells in my body can reasonably be attributed to the accretion of random errors.

Barry’s Response to Her Probable Response 1:
Surely you don’t believe your comment [25] could reasonably be attributed to random key strokes by our proverbial monkey. That has a sort of first blush plausibility, but as we all know the math does not work. Then why do you attribute the CSI in the cells in your body to the accretion of random errors?

Elizabeth’s Probable Response 2:
Well, it’s not just random chance. The good random errors are selected for by natural selection and bad random errors are eliminated by natural selection. So it is not pure chance. It is chance (random genetic errors) and necessity (natural selection) combined that results in the CSI.

Barry’s response to her Probable Response 2:
What an extraordinary claim! This remarkable interaction of chance and necessity to which you allude has never been observed even over trillions of reproductive events by bacteria under intense selective pressure. What non-question begging evidence do you have for your remarkable assertion? Remember, Dawkins says that extraordinary claims require extraordinary evidence. And surely you will agree that saying the staggering amounts of CSI in living things is the result of the accretion of random errors sorted by natural selection is the most extraordinary claim ever made.

Elizabeth’s Probable Response 3:
Let’s change the subject.

Comments

So my last post bring up an interesting question. What would happen in this clock-making GA if he did not use crossover?Mung_{August 8, 2011
August
08
Aug
8
08
2011
12:22 PM
12
12
22
PM
PDT}

Elizabeth Liddle:
OK, Mung, in future I will try not to use the word “self-replicate” for sexual reproduction, even though it is widely so used. Call them breeders if you want. It makes no difference to the case or the math.
Hello passive-aggressive dishonest person. What have you done to our oh so polite, kind, and always honest to a fault Dr. Liddle? In what sense of the term "breed" do the organisms in this clock-making program breed? In what sense of the term "to reproduce via sexual reproduction" do the organisms in this clock-making program reproduce via sexual reproduction? Even the creator of this presentation doesn't say his organisms self-replicate or sexually reproduce. Since we're talking about GAs, not natural breeding and sexually reproducing populations of biological organisms, why not employ the terms appropriate to the domain? I'll tell you why. Because it helps facilitate the charade that what takes place within natural populations is analogous to what happens in a GA. iow, it's for rhetorical effect. But that's the point in dispute, so why should we allow you to beg the question?Mung_{August 8, 2011
August
08
Aug
8
08
2011
12:16 PM
12
12
16
PM
PDT}

Mung:
[I suspect there are no non-solutions, but I'll have to look at the code]
Of course there are non-solutions. 98% of the randomly generated genomes are non-solutions (do nothing to tell the time). So we know that only 2% of the search space contains solutions. And of that 2% a tiny fraction contain solutions that involve several hands and gears.
The potential solutions are programmed in, by design.
No they aren't They are possible (they exist in the search space, obviously, because they are found), but nothing except the Darwinian algorithm "programs them in".
So is the path to each potential solution. Programmed in, by design.
No. The only program that "designs" the search path is the Darwinian one.
We know something will be found by the search, we just don’t know what, specifically, in advance.
Yes, we do know that, and the reason we know that is that we know that Darwinian search works.
How many functional “time-pieces” are there within the search space?
Dunno, but we can put an upper bound of considerably less than 2%.
What is the actual size of the search space?
You'll have to look at the program. It'll be the total number of possible genomes. I'll try and find out, but if you think about it (unless you think the man is lying) you can probably ballpark it from the info he gives.Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
11:53 AM
11
11
53
AM
PDT}

Mung:
What you fail to see is that in order for a proof of principle to be a proof of principle it actually needs to be a proof of principle.
Which is what it is. It has all the ingredients: breeding with heritable variance in probability of breeding. The clock-part assembly part of the genome is randomly drawn at the start, with the result that only 2% of critters have any timekeeping facility at all, and that minimal. From there, the search space (every possible genome) is reliably traversed so that the final population is a population of accurate clocks (a solution from the small but still substantial solution subspace of the much larger total search space). And it obviously finds these solutions without having to randomly pick straws out of the whole haystack, therefore does it much faster than a random search would do.Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
11:45 AM
11
11
45
AM
PDT}

Elizabeth Liddle:
And the end point is NOT determined – the problem is determined, but the solution is not, and there are a very large number of potential solutions. The space of non-solutions is orders of magnitude larger. Nonetheless the algorithm finds one of the subset of solutions each time, despite these being unknown in advance to the designer of the algorithm.
[I suspect there are no non-solutions, but I'll have to look at the code] The potential solutions are programmed in, by design. So is the path to each potential solution. Programmed in, by design. We know something will be found by the search, we just don't know what, specifically, in advance. How many functional "time-pieces" are there within the search space? What is the actual size of the search space?Mung_{August 8, 2011
August
08
Aug
8
08
2011
11:43 AM
11
11
43
AM
PDT}

OK, Mung, in future I will try not to use the word "self-replicate" for sexual reproduction, even though it is widely so used. Call them breeders if you want. It makes no difference to the case or the math.Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
11:39 AM
11
11
39
AM
PDT}

Elizabeth Liddle:
Please lets not get into a lawyer quibble about whether sexually reproducing populations “self-replicate”. They do, it’s just that some genome shuffling goes on as well, which is cool.
So you want me to pretend like the "organisms" in this clock-making program are true self-replicators and not get into a liar quibble about it? We both know they don't self-replicate, so why should we say that they do?Mung_{August 8, 2011
August
08
Aug
8
08
2011
11:33 AM
11
11
33
AM
PDT}

Elizabeth Liddle:
What I find frustrating here is that when anyone produces a proof-of-principle – whether it be WEASEL or the Clock simulation – of the powers of rm+ns, the result is roundly dismissed as being too simplistic to apply to living things, which is fine.
Where have I, up to this point, argued that this clock-making demonstration is too simple? What you fail to see is that in order for a proof of principle to be a proof of principle it actually needs to be a proof of principle.Mung_{August 8, 2011
August
08
Aug
8
08
2011
11:25 AM
11
11
25
AM
PDT}

Ilioin:
Wow! Who knew that “natural selection” was so potent a force in the cosmos that artificial “natural selection” is indistinguishable from selective breeding?
Having trouble pinpointing the target of your sarcasm here, Ilion. Indeed, you could regard the clock program as an example of artificial selection aka selective breeding. That's a nice way of looking at it. Are you suggesting that the role of a ID in biology is at the level of selection?Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
01:44 AM
1
01
44
AM
PDT}

MI:
If the powers of creation are unlocked in 5000 lines of code, why can they only make little virtual clocks, for which the target functionality is largely assumed right off the bat, the starting and ending points already determined?
Because, to demonstrate the principle, it is useful to reduce the dimensions of the fitness landscape. In this case to a single dimension. And the end point is NOT determined - the problem is determined, but the solution is not, and there are a very large number of potential solutions. The space of non-solutions is orders of magnitude larger. Nonetheless the algorithm finds one of the subset of solutions each time, despite these being unknown in advance to the designer of the algorithm. Note the surprising design of the antennae in the NASA paper.
The sheer scale of creative wonders that must be present in chance and necessity accounts of evolution, if they are truly modeled in this type of simulation, should accomplish much more than being cute little toys.
And they can, and probably do, but people primarily make GAs to solve their own problems and so they design fitness functions (i.e. environments) in which the critters, by evolving to survive in that environment, will come up with a solution to the designer's problem. Although in fact we do utilise solutions that the environment has set to evolving populations, even though we didn't actually design the problem. Robotics is full of designs borrowed from evolution. But then I guess you wouldn't see it that way :)Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
01:38 AM
1
01
38
AM
PDT}

What I find frustrating here is that when anyone produces a proof-of-principle - whether it be WEASEL or the Clock simulation - of the powers of rm+ns, the result is roundly dismissed as being too simplistic to apply to living things, which is fine. However, that is a massive moving of goal posts: the claim that both WEASEL and the clock simulation were devised to refute is the claim that rm+ns cannot (in the case of weasel) produce sense out of nonsense, or (in the case of the clocks) find a clock-solution efficiently out of the Vast space of non-clock solutions. In both cases the claim is duly refuted. The clock video, in particular, falsifies the notion that rm+ns is a monkeys-on-typewriters algorithm or a needle-in-a-haystack algorithm. It demonstrates that rm+ns is an efficient search algorithm. So does the NASA paper about antenna design. All I am asking for here is for the concession that rm+ns can find design solutions in a vast search space, including solutions that an engineer would be slower (in the case of NASA) to dream up, if at all. In other words it can be creative. If you accept that notion, then indeed, we can go on to discuss whether it could account for life. Although already micro-evolution is widely accepted, so I'm not sure what the objection to the clock simulation is, frankly. So let me ask: 1) do you accept, having watched that video, that one a large number of possible clock designs is reliable "found" by the algorithm from an orders of magnitude number of clock-part combos that do not function as clocks? 2) do you accept, having watched the video, that while the fitness function and the physics-and-chemistry of the system were designed by the programmer, the clock itself was designed by the program? If not, why not?Elizabeth Liddle_{August 8, 2011
August
08
Aug
8
08
2011
01:24 AM
1
01
24
AM
PDT}

Well, I made it more than half way through that video before beginning to stare off into the horizon, like I often do when listening to long stories from some of my relatives. I can't say why a toy with 5000 lines of code (just guessing) is taken seriously as supporting some fundamental feature of evolution. And that's what these things are -- toys -- delightful little toys. If the powers of creation are unlocked in 5000 lines of code, why can they only make little virtual clocks, for which the target functionality is largely assumed right off the bat, the starting and ending points already determined? The sheer scale of creative wonders that must be present in chance and necessity accounts of evolution, if they are truly modeled in this type of simulation, should accomplish much more than being cute little toys. I confess to taking only a superficial interest to the simulation in the video, but that's all I can seem to gin up for this sort of virtual bluster. So I'll make allowance for the fact that there's more going on than I could deduce in a painful six-and-a-half minutes. I guess I would be more impressed if it evolved pocket watches, grandfather clocks, digital wrist watches, clock radios, and eventually televisions. Perhaps that's too high a bar to set, in order for something like this to surpass entertainment; but that's what we're supposed to expect from evolution, whether taking into account the intractable problem of abiogenesis, or just the trivial issues of evolving a plethora of disparate body plans from the first self-replicating, living organism. If you detected a little sarcasm toward the end of that last paragraph, then your sarcasm sensor is appropriately calibrated. You're welcome.material.infantacy_{August 7, 2011
August
08
Aug
7
07
2011
08:41 PM
8
08
41
PM
PDT}

Mung @ 128:
(3:19) Remove 3 at random and arrange them in order of their ability to accurately tell time. (3:26) The better two clocks kill the worst clock. (3:31) Mate the surviving two and produce an offspring.
Wow! Who knew that “natural selection” was so potent a force in the cosmos that artificial “natural selection” is indistinguishable from selective breeding?Ilion_{August 7, 2011
August
08
Aug
7
07
2011
08:40 PM
8
08
40
PM
PDT}

lest the point be lost Elizabeth Liddle:
What is good about it, I think, for present purposes, is that it [Webster's definition 2b] does not require a sender and/or receiver.
How does something which is communicated not have a sender and receiver? ...the attribute inherent in and communicated by one of two or more alternative sequences or arrangements of something (as nucleotides in DNA or binary digits in a computer program) that produce specific effectsMung_{August 7, 2011
August
08
Aug
7
07
2011
08:26 PM
8
08
26
PM
PDT}

In her post @107 Elizabeth directed me to an internet video which purports to demonstrate the designing power of Darwinian mechanisms. In my post @111 I offer an initial critique in which I point out at least three ways in which the GA is not at all analogous to populations of living organisms with all three objections focusing on just one single aspect of the GA, the initial population. I wrote:
1. The initial pool consists of completely random genomes (3:16).
Now Elizabeth attempted to argue that in natural populations the genomes are random. And if that is not what she is arguing I have no idea what her objection is to my first point. Elizabeth Liddle:
But in any case, you are being over-literal – the starting genomes are random with respect to clock-making, but they are nonetheless minimally functional in that they are all capable of sexual reproduction. It is only with respect to making clocks that they are random, , just as a population faced with a new environmental hazard might find itself armed with nothing better to handle it than a random bit of genome. But, finally, your objection is bizarre – the fact that the critters start with random genomes makes the point more strongly, not less – the proto-clocks start off with nothing other than the ability to self-replicate – no “pre-loaded” information in the genome. No “smuggled in” part-solution (a criticism more commonly leveled at GAs).
Mung:
Do you mean that the genomes of the initial population are random with respect to the current environment?
Elizabeth:
They are random with respect to clockmaking in the current environment.
My response:
False, false, and, yes, wait for it, false.
Elizabeth Liddle:
Well, it’s my turn to not understand this, Mung. The starting genomes are randomly generated, not generated with a bias towards assembling clocks. I don’t know what your problem is with this.
My problem with this is that it is FALSE. NOT TRUE. NOT TO BE BELIEVED. You admit so in your own words for gooodness sake and contradict what you wrote ealier!
...that little program is a perfect example of how, starting with a randomly generated set of genomes that, at best, and in a minority of the population, specifies no more than a simple pendulum,
So it is biased towards making clocks. And you know it! And yet you can say that it is not. And you can also say that the genomes are random with respect to clock-making. Also false. (1:33) Details of the simulation: The Componenents (1:33)
Gears, Ratchet, Hands, and Spring are replaced by the following symbols...
(2:21) Each clock organism consists of:
30 gears, 1 ratchet, 7 hands, 1 spring, and 1 housing
(2:39):
The clock genome is a matrix containing the information of who binds to who and what their properties are.
And the icing on the cake:
(3:15) Start with a pool of clocks with COMPLETELY RANDOM GENOMES (3:19) Remove 3 at random and arrange them in order of their ability to accurately tell time. (3:26) The better two clocks kill the worst clock. (3:31) Mate the surviving two and produce an offspring.
And yet you assert that the initial population is random with respect to clock-making. It is as plain as day that there is no "random with regard to clock-making" in this GA. the GA is designed to make timepieces. Elizabeth:
Mung – you can’t just dismiss simulations because they are too simple – the whole point of simulations is to demonstrate a principle using simple examples.
I am not dismissing it because it is too simple. I am dismissing it because it is not analogous to the way evolution purportedly works. I am showing that your repeated assertions that GAs are evidence for what evolution can do have serious and fundamental flaws. In this instance, with this particular GA, you have not offered a valid rebuttal to my first point. Nor have you ever, when I raised this exact same issue in the past. Yet you continue repeat the same claims and assert that no one has offered a convincing counter-argument. Here it is, one more time: 1. The initial pool consists of completely random genomes (3:16). How is that like a biological population? It isn’t. This dis-analogy between GAs and natural populations has been pointed out to you before. Natural populations occupy points within the search space that are close to each other. The initial populations in a GA are designed to do the exact opposite, as this GA shows.Mung_{August 7, 2011
August
08
Aug
7
07
2011
07:48 PM
7
07
48
PM
PDT}

I have to make a conscious effort to sound like a redneck, and then I don't really get it right (especially in writing). What's odd about this is that my father's people were/are; even though he had only 19 months total of formal education, and never totally eliminated the accent, he was determined to "talk proper," and set the example in our home.Ilion_{August 7, 2011
August
08
Aug
7
07
2011
07:07 PM
7
07
07
PM
PDT}

Elizabeth Liddle:
But you could watch this if you have a spare few minutes:
Also from the vid:
The theory of evolution is NOT a theory of life's origin. It is a theory of how one form of life changes over time through mutation and natural selection, into another form. (0:45)
Oh my. That's just so wrong.Mung_{August 7, 2011
August
08
Aug
7
07
2011
07:04 PM
7
07
04
PM
PDT}

Elizabeth Liddle:
But you could watch this if you have a spare few minutes:
From the video:
My favorite straw-man argument for the proof of Intelligent Design goes as follows... Take a watch... smash it with a hammer... put the pieces in a box... shake it around... open the box... and see what you get. ID advocates believe that since you do not reform a functional watch, not only must intelligent design be correct, but evolution must be wrong. Funny how such a simple experiment, supposedly able to disprove one of the most strongly supported theories in the history of science, has NEVER appeared in any scientific peer reviewed journal.
Well Elizabeth, if it's places like that where you get your ideas about ID it explains a great deal. You didn't see that and immediately think, wow, this guys is lying through his teeth? Have you come across anyone here making such an argument for ID or against evolution? Seriously.Mung_{August 7, 2011
August
08
Aug
7
07
2011
06:58 PM
6
06
58
PM
PDT}

If I did that I'd sound like a total redneck.Mung_{August 7, 2011
August
08
Aug
7
07
2011
06:47 PM
6
06
47
PM
PDT}

Ilion, Hehe. Point taken. But in casual writing I usually write the way I talk :Dmike1962_{August 7, 2011
August
08
Aug
7
07
2011
06:00 PM
6
06
00
PM
PDT}

"… And the search space is damn small." May I inject a point of grammatical pedantry here (in contrast to Miss Grundyish prissiness)? It seems to me that the grammatically proper word would have been ‘damned’, not ‘damn.’ It’s kind of like writing/saying “you were suppose to do that” when one means “you were supposed to do that”Ilion_{August 7, 2011
August
08
Aug
7
07
2011
05:44 PM
5
05
44
PM
PDT}

EL @ 119: There are only two (*) logically available ways to categorize your post #119, wherein you pointlessly try to dispute the very specific thing I said. Either: 1) you are so stupid that you simply cannot understand the noises that come out of your mouth; 2) you are so intellectually dishonest that you say, in effect, “Well, yes, you’re right about that … which just shows how wrong you are about that!” Now, I simply do not and cannot believe that you’re stupid. That leaves only one option (*). (*) For the third potential categorization (and there are only three possible in total) – that you are missing or not understanding some critical prior information – clearly does not apply here. It is clear from you pointless disputing that you understood exactly what I said.Ilion_{August 7, 2011
August
08
Aug
7
07
2011
05:33 PM
5
05
33
PM
PDT}

The video that Elizabeth cites was designed by an intelligent agent with carefully crafted fitness functions toward productions within certain quantifiable bounds. And the search space is damn small. It's interesting that all genetic algorithms that are known to produce novel F/CSI have been designed by intelligent beings, namely, humans. As for the F/CSI in nature, it's an open question as to how much intelligent intervention was required.mike1962_{August 7, 2011
August
08
Aug
7
07
2011
05:33 PM
5
05
33
PM
PDT}

Moreover, any particualar execution of it, given an identical set of inputs, will always crank out an identical set of not-ice cream and not-atomic bombs.
Only if a pseudorandom number is used and the same seed for every run. Otherwise, it won't, and doesn't, as is made clear in the video. It reliably evolves a clock, but the clocks are not identical.Elizabeth Liddle_{August 7, 2011
August
08
Aug
7
07
2011
05:18 PM
5
05
18
PM
PDT}

Mung:
Elizabeth Liddle: …the starting genomes are random with respect to clock-making… They are probably random with respect to nuclear bomb-making and ice-cream making as well.
Are you trying not to understand this, Mung?
Yet for some inexplicable reason the program doesn’t crank out ice cream, nor nuclear bombs.
Not inexplicable at all. Biological organisms don't crank out nuclear bombs either.
So much for the power of Darwinian processes to design things. So since you have so far refused to face this issue, let me put it another way. The starting genomes are random with respect to other genomes in the initial population.
Well, it's my turn to not understand this, Mung. The starting genomes are randomly generated, not generated with a bias towards assembling clocks. I don't know what your problem is with this.
How is this like an actual biological population? Answer, it isn’t.
Only in the sense that in an existing biological population you aren't starting from scratch. But that makes the program more impressive, not less.
So here we are, still at square one.
They are random with respect to clockmaking in the current environment.
False, false, and, yes, wait for it, false.
OK, Mung, time for me to give up again. You don't seem to know what I am talking about, and I don't know what you are talking about. As I see it, that little program is a perfect example of how, starting with a randomly generated set of genomes that, at best, and in a minority of the population, specifies no more than a simple pendulum, evolves, by rm+ns, a population of highly accurate clocks. That demonstrates that rm+ns can design things. And, if you don't like the clocks, take a look at those NASA radio antennae. I'm off to bed :) Sleep well.Elizabeth Liddle_{August 7, 2011
August
08
Aug
7
07
2011
05:11 PM
5
05
11
PM
PDT}

"Yet for some inexplicable reason the program doesn’t crank out ice cream, nor nuclear bombs." Moreover, any particualar execution of it, given an identical set of inputs, will always crank out an identical set of not-ice cream and not-atomic bombs.Ilion_{August 7, 2011
August
08
Aug
7
07
2011
05:07 PM
5
05
07
PM
PDT}

Elizabeth Liddle:
...the starting genomes are random with respect to clock-making...
They are probably random with respect to nuclear bomb-making and ice-cream making as well. Yet for some inexplicable reason the program doesn't crank out ice cream, nor nuclear bombs. So much for the power of Darwinian processes to design things. So since you have so far refused to face this issue, let me put it another way. The starting genomes are random with respect to other genomes in the initial population. How is this like an actual biological population? Answer, it isn't. So here we are, still at square one.
They are random with respect to clockmaking in the current environment.
False, false, and, yes, wait for it, false.Mung_{August 7, 2011
August
08
Aug
7
07
2011
04:57 PM
4
04
57
PM
PDT}

Mung
haha, how funny is it that we are talking about a GA that produces timepieces in a thread called “to save time.” Now Elizabeth, would you please be so kind as to give me the time signature for the point in the presentation at which he shows his timepieces self-replicating?
2'50". The frame that says: "Matrixes are mated just as in my first two videos." It shows Mother Father and Offspring. It also says "Every offspring receives one 1 mutation".
And would you also explain what you think he means when he says that 98% of the initial population of 10,000 genomes “do absolutely nothing” (3:38)?"
He means that in no way can they be used to tell the time. The remaining 2% are simple pendulums, which can be used to tell the time by counting swings.
Do you seriously think he means by “98% do absolutely nothing” that they do in fact “do something,” as in they each and every one of them self-replicates?
Three members of the population are drawn at random and the "best two" are mated with each other, the "best" being the ones best at telling the time. I can't remember if he says in the case of a draw, but in most algorithms, in that case, the loser is selected at random. So all are capable of replication - sexual replication in fact, and therefore of passing on part (or indeed all, as the genome is so small) to their offspring. I'm sorry if I didn't make that clear, but should be clear from the video. Please lets not get into a lawyer quibble about whether sexually reproducing populations "self-replicate". They do, it's just that some genome shuffling goes on as well, which is cool.
Elizabeth Liddle:
the starting genomes are random with respect to clock-making
I have no idea what that could possible mean. Random with respect to clock-making? How would we test that claim?
As it's a simulation you can simply ask the simulator: is set of genotypes biased in favour of clock-making? And the answer is no - he says they are random.
Do you mean that the genomes of the initial population are random with respect to the current environment?
They are random with respect to clockmaking in the current environment. The guy doesn't bother with a genome that specifies the mating procedure - he just does that, so in effect he assumes wrt to replication processes they are identical. If you are having difficulty getting your head round this, imagine that every organism has a genotype that tells it how to mate and produce offspring, but as it is identical in all cases, he doesn't show it. But that each genotype has a 9 piece slot that in 98% of cases is just junky stuff that causes it to stick bits of clock part together randomly, regardless of whether the result is any kind of clock. But in 2% of cases results in a simple pendulum. Mung - you can't just dismiss simulations because they are too simple - the whole point of simulations is to demonstrate a principle using simple examples. The point here is that the fitness function, the physics and chemistry are all designed, as is the mating system. But the "clocks" themselves are not - they evolve, given the physics and chemistry, the self-replication, and the fitness function. In other words, the Darwinian bit works. Given: Necessity, Chance and replication with variance, and a fitness function that determines what phenotypic features are most likely to breed, you get a designed clock. The Darwinian algorithm "finds" a clock in the Search Space, if you like. It doesn't have to search every single combination of parts in the hope that a clocky one turns up. It moves steadily towards clockiness, although, interestingly, not producing identical clocks each time.Elizabeth Liddle_{August 7, 2011
August
08
Aug
7
07
2011
04:23 PM
4
04
23
PM
PDT}

Elizabeth Liddle:
the starting genomes are random with respect to clock-making
I have no idea what that could possible mean. Random with respect to clock-making? How would we test that claim? Do you mean that the genomes of the initial population are random with respect to the current environment?Mung_{August 7, 2011
August
08
Aug
7
07
2011
03:56 PM
3
03
56
PM
PDT}

haha, how funny is it that we are talking about a GA that produces timepieces in a thread called "to save time." Now Elizabeth, would you please be so kind as to give me the time signature for the point in the presentation at which he shows his timepieces self-replicating? And would you also explain what you think he means when he says that 98% of the initial population of 10,000 genomes "do absolutely nothing" (3:38)? Do you seriously think he means by "98% do absolutely nothing" that they do in fact "do something," as in they each and every one of them self-replicates? pleaseMung_{August 7, 2011
August
08
Aug
7
07
2011
03:49 PM
3
03
49
PM
PDT}

Prev 1 2 3 4 … 6 Next

You must be logged in to post a comment.

Leave a Reply