## Saturday, April 09, 2016

### P.Z. Myers believes that genetic engineering is impossible

P.Z. Myers is an aggressive, far left-wing blogger and a would-be scholar in life sciences who proudly describes himself as "a randomly ejaculating liberal scumbag of godless šit." I've interacted with this individual about 10 times in the past. He even tried to influence the discussions about the sum of integers and the zeta-function regularization, topics that this classic genius of mediocrity has no chance to master.

But he has something to do with biology and he should know something about it, right? He must know about it if he has attacked the conservatives and Christians and others about 1 million times. Well, no. This guy is a complete crackpot in biology, too. A clear new proof came from his arguments about the genetic editing of human intelligence with physicist Steve Hsu:
Hsu about IQ in Nautilus (Superintelligent humans are coming, 2014)

P.Z. Myers' reply at his Pharyngula blog 4 days ago (Superbrains will not come out of a test tube)

Hsu's blog: a fresh reaction (This is for PZ Myers)

Myers' super-new counter-rant (Nothing but us big fat liberal chickens around here)
The genetic modifications of humans are morally controversial. I don't know whether I should celebrate them and endorse them or be worried. But it seems clear to me that those efforts can't be completely stopped, all the know-how is basically ready, and – just like with the nuclear bomb – it's better if relatively good guys are ahead of the bad guys.

But the argument between Hsu and Myers wasn't really about "whether it's desirable to do such things". It was about "whether it's possible". Yes, Myers has basically revolted against some of the very basic facts that biology has learned about cognitive skills; about evolution; about genetics; and he has denied the possibility of the technological progress, too.

Hsu wrote that e.g. John von Neumann was unusually smart. Von Neumann could remember every word from a book he just read. I usually can't do that – it's a success for me to remember that the fashionable word for the photographic memory is "eidetic memory". There are clearly genes that influence the human intelligence (and its slightly inequivalent but correlated aspects such as the memory, abstract thinking potential etc.). They change somewhat randomly in Nature. But if geneticists manually and carefully edit them in a desired direction, they may produce much smarter humans.

These general claims seem absolutely obvious to me. Most of the cognitive skills boils down to genes – even if it were just a "substantial part", not "most", it would be enough. And some people who were smart clearly had genetic reasons for that fact. You could use and combine some of their genes to get smarter clones or improved variations.

Steve helpfully gave us this incredible picture of the increasing size of chickens. Currently chickens are some 30 (old) standard deviations above what they were half a century ago. It seems obvious to me that other variables such as the IQ may be affected by genetic engineering analogously to the size of the chicken. After all, the size is not the only thing that matters but it matters, too. The very size of the human brain affects the IQ and it may be influenced by genes just like the size of the chickens.

How smart the people could be made?

Hsu sketches some models but let me give you a simplified model of mine. The human DNA may contain some 10,000 genes that influence the IQ significantly. Let's assume that each of them may be realized either in "plus" or "minus" version, thus increasing or reducing the intelligence. You get about $$5,000\pm 100$$ pluses; the statistical deviation goes like $$\sqrt{N}$$, OK?

So the standard deviation of about "100 genes" is what produces the standard deviation in the IQ we know and denote as "15 IQ points" (away from the mean value of 100 IQ points). In principle, you could turn all the remaining 5,000 genes to "plus". That's 50 times the "root mean square" improvement of 100 genes, so you expect to increase the IQ by 50 times 15 if you create this "cognitively optimized human". The IQ could go to $$100+50\times 15=850$$, not far from one thousand.

(I suspect it's a slight overestimate because in the populace, many of the genes are correlated with each other, so the actual genetic variability is higher than estimated above – many genes tend to flip simultaneously – and therefore the potential for "additional" improvements is a bit smaller. But let's ignore technicalities.)

I am not saying that we will be there next year etc. but I think it is a fair approximate calculation of the potential hiding in this kind of genetic modification and I am confident that Hsu will agree that my sketched estimate is at least as good as his.

The IQ around 1,000 is a problematic thing because the right "possibly nonlinear change of variables" you may apply to parameterize such high IQs is not specified. In the calculation above, I assumed that the IQ is basically a "linear function of the good IQ genes". However, the dependence could be more nonlinear.

At the end, the IQ should be defined in such a way that "all the humans" have the IQ normally distributed around the mean value of 100 and with the standard deviation of 15. In this sense, many people having the IQ about 1,000 is an "oxymoron" because the probability that you find an IQ 1000 person now is the 50+ sigma effect we mentioned, basically zero.

You may artificially produce the first humans with the IQ around 1,000 – relatively to the current population that is meant to be normally distributed around 100 – but you won't really be able to figure out whether his IQ is 1,000 or 3,000. The person will be much smarter than all genetically unmodified people but his "advantage" can't be really quantified, if you get my point. And if you want to get from "current moderate IQs" to the "high IQs of the future" (around 1,000, but they would be renormalized back to 100), the precise numbers you assign depend on all the details how the world population will evolve.

But let's ignore the technical details concerning the quantification of the IQ. Regardless of these ambiguities, it's clear that there would be a huge and immediately visible qualitative difference between the normal people and the monster minds with the IQ around 1,000.

These things are amazing to speculate about and they will be fascinating – and maybe terrifying – if and when they are turned into reality. But another shocking thing is the complete denial of all the basic facts about genetics by P.Z. Myers, a guy who has marketed himself as a staunch defender of evolution (the word "evolution" is the first word in the subtitle of his blog). Myers wrote:
Stephen Hsu thinks super intelligent humans are coming.

...

Good god, what a load of crap. Lots of people seem to think it’s brilliant, though. It isn’t.
Notice that when his foe actually understands genetics better than P.Z. Myers, godless P.Z. Myers suddenly starts to whine and beg God for assistance. Myers continues by saying that the intelligence is pretty useless because the voters of Trump are rather intelligent but he doesn't like them, anyway, but then he returns to the key point:
Then the question becomes one of whether such an increase is possible, and whether genetic engineering is a practical way to achieve it. My answer to the first is that it’s unlikely, and the second is a flat no.
As announced in the title, he just doesn't think that genetic engineering may work for humans.
Our evolutionary history suggests that there was a period when Homo‘s brain was undergoing a long period of gradual enhancement. It wasn’t Homo sapiens though; it was Homo erectus. Fossils of that species over its 2 million year history.

...

But taking a million years to figure out how to put a sharper edge on a stone hand axe isn’t exactly a rapid development cycle.
This sort of complaint is pretty much identical to what you often hear from fans of creationism. Evolution doesn't really work because we didn't see it taking place quickly. "Have you ever seen a bird turning into a squirrel?" some creationists at my Alma Mater would sometimes ask before they concluded that evolution was ludicrous. And now you hear it from a scumbag who has been trying to pretend that he was smarter than the creationists.

Sorry but it is absolutely OK for the evolution to proceed slowly. The Earth has reserved billions of years for that process. Naturally, it has indeed taken something like millions of years to create notable changes. But that's because the changes of the genes are mostly chaotic and most of them cancel out.
A key novelty of genetic engineering is that the modifications aren't really chaotic and they're done in such a way that most of them don't cancel out.
It is obvious from Myers' rant that he doesn't get this elementary point. He doesn't really understand that genetic modifications may be done and why their effects are fast.

Let me continue to discuss the model with "10,000 binary genes underlying the human IQ". About 5,000 of them were said to be "positive" for a human person. Imagine that after each 25 years, a generation, the average number of positive genes in the population changes by one – but in a random direction. And apply almost no selection, at least in the short run. How quickly will the IQ evolve?

Well, it's a random walk. After $$N$$ generations, the typical change of the number of "positive genes" will be about $$\sqrt{N}$$. How much time does it take for this random walk to increase the IQ by one standard deviation which we said was equivalent to 100 excess positive IQ genes? Well, 100 is the square root so you need 10,000 generations or 250,000 years in this model.

You should in no way assume that this number is precise. It may differ by an order of magnitude – but not much more than that. The rough conclusion is correct: it takes intervals of time that are comparable to one million of years for the natural "random walk" to change the average by the amount that is comparable to the standard deviation at a given moment.

If you need e.g. 250,000 years to increase the IQ by 15 – assuming that the natural selection was very weak but it preferred the positive trend for the IQ – you may immediately see that some 500,000 years ago, the human IQ could have been 70. At those times, about one-half of humans were morons. Some 10% were imbeciles and the percentage of idiots (who are stupider) was much higher than today, too. P.Z. Myers would feel much more welcome over there.

When Myers asks
One has to wonder, if IQ is such a great boon to humanity, why hasn’t the biological basis for it shown much improvement in the last 100,000 years?
the answer is, once again, obvious to everyone who has understood Darwin's theory at least at some basic level. Because in Nature, most of the effects of mutations cancel out, like in a random walk.

But with the genetic modification, it's an entirely different story. First, the step doesn't need one generation or 25 years. The step may be 1 day in a laboratory or less. Equally importantly, the steps don't add up as a random walk. They may be added in the positive direction systematically, so the number of required steps isn't the square of the required number of improvements, but the number of required improvements itself.

You can easily imagine that all those 5,000 "negative" IQ genes of a human DNA are fixed in a year or something of the sort. When Myers writes
Evolution is far better at tinkering than humans are, and has been tweaking our species for a long, long time, but super-brains haven’t emerged yet,
it's just insane. In Nature, evolution may be said to be "safer" or "more natural" if you embrace some old-fashioned environmentalist (or Luddite) thinking. But it is definitely not true that Nature is "better [faster] at tinkering" things than the geneticists. Just look at the damn chickens at the picture above. Once again, Nature would need millions of years for that, not just 50 years. Genetic engineers may be said to be "less wise" or "less safe" or "less careful" than Nature but they have surely become much more efficient in tinkering the genes than Nature.

In the following paragraphs, Myers talks about the origin of people in Africa – probably right – and claims that the improvements didn't result from the increases of the IQ but from the creation of the language and social structures.

Well, that's wonderful except that the chimps are still unable to invent and use their own Czech language – and, as far as I know, no community of chimps has built a stock market that would also allow the trading of derivatives yet.

It's obviously right to say that the language and social structures have done a lot for the mankind to become civilized. But it does not follow that the creation of the language or social structures were the primary causes or that they had no prerequisites. There obviously have been lots of prerequisites. The members of the homo species first had to become sufficiently intelligent – they had to become sapiens – and only when it was good enough, the language and the social structures could develop. And such a development was arguably straightforward and unavoidable because the human communities with these achievements had big advantages over others.

But all these achievements depend on some degree of intelligence so the increase of the intelligence when "humans were becoming humans" was clearly an essential process that was necessary for the existence of the civilization today. It's remarkable but P.Z. Myers basically denies this self-evident fact, too.

Because he sort of realized that he hasn't found any argument against Hsu's picture that would be defensible, he tried to look for problems in chaotic ways. What about this one?
Note his estimate of the number of genes that contribute to IQ: 10,000. That’s half the human genome! Hmmm. I wonder if any of those genes play a role in other processes in human physiology that might be affected by his plan?
Right. Humans have 20,000-25,000 genes that affect the production of proteins. I can assure you that almost all of these genes have numerous consequences. When the size of the chicken was being inflated, something else about the chicken could have changed, too. But the chickens were still capable of living happily.

All the "positive" versions of the 10,000 genes that Hsu wants to include to his "monster mind" already exist out there. He just wants to combine them – all of them or a big fraction of them. It's hard to believe that the combination of genes that are fine for survival individually would lead to death. It's just possible to change most genes safely.

Some of these high-IQ genes may be related with conditions that could be called "diseases". But if the main goal is to create a monster mind, some diseases or things like skin rashes are secondary. At the end, the real reason of Myers' hysteria is that he hates people who are smarter than him. That's why he hates about 80% of the mankind.

But there's nothing unavoidably wrong about the hypothetical humans with the IQ around 1,000. The improved versions of the genes may have other consequences, but as long as they may live and reproduce, they just help to shape the identity of those folks. To describe these conditions negatively means to be prejudiced. Because of the higher IQ, these alleles are at least "mixed packages". I don't know how happy the supersmart folks would feel – but I am sure they would be very useful and consequential.

In the following, even more laughable, part of Myers' rant, he claims that it is impossible to construct fast cars. To see how ludicrous Hsu is, Myers says, imagine that engineers want to make a fast car so they dismantle a car to pieces and try to make every piece faster. Myers says that it's absurd!

He must have never heard of Formula One. Or Koenigsegg Agera One:1 (picture above) whose top speed is said to be 270 mph or 435 kilometers per hour. What the hell is absurd or impossible about the construction of fast cars and the optimization of each component of the car for the speed? How totally f*cked-up idiot P.Z. Myers must be to make similar statements? Like the statement that it's impossible to make cars faster by tweaking their components? It's the kind of tasks that many car designers do all the time – and the genetic engineers and other occupations do tasks that are indeed vaguely analogous, too.

P.Z. Myers hasn't done anything useful in his life – let alone a contribution to the scientific or technological progress – so he prefers to believe that it's impossible to do useful things.

Myers also tells us that it's impossible to increase the speed of the cars by playing with each component because... because he hasn't been given a flying car yet. So the techno-optimists are idiots. Wow. Except that even his claims about the flying cars are not just inequivalent to the claims about the fast cars but wrong as well. For example, a Slovak engineer named Mr Štefan Klein has built a flying car, Aeromobil 2.5, for himself, play the video over there. The version 3.0 is actually out. (I am about 50 times more thrilled by such a flying car than by another generic electric car from Musk.) All the things Myers says to be impossible are possible and many of them are already out there.

Another "argument" of Myers against Hsu's picture is some story about an egg and who is going to take the responsibility for that. Well, some people who organized, paid for, or did the work are surely responsible for these acts and they may face moral and legal hurdles. Many other people in the world do face such things, too. How can it change the fact that it's possible and it's probably going to happen at some point?

The stupidity of some leftists who otherwise want to call themselves "pro-science" is incredible.

By the way, in his second reply yesterday, Myers started by an attack on "twin studies". Such an attack is insane, too. Twins are priceless for research and these studies have irreversibly shown that many things are really up to genes because the agreement between many quantities describing the twins was just stunning. But Myers just doesn't like the conclusions so the twin studies must be bad and he links them to racists, Libertarians, and techno-fetishists. Cool.

Later, Myers basically says that we only inherit our membership in homo sapiens but nothing more detailed than that. Wow. And he says that the IQ can't be influenced because it's not "quantitative" but "due to many factors". Wow. If someone is a complete idiot who has a job based on the illusion that he's smart, you may see that quite a lot of efforts goes into "proofs" that the difference between idiots like him and the smart people where he should belong doesn't exist. But it does exist.

You may see that he has some equally retarded fans. One of them denounced powerful brains because "brains consume too much energy". Well, no doubt, many readers of Myers' blog prefer to send much more energy to their aß than to their peabrain.