bookmark_borderSecond-order selection against the immortal

In his recent review of Lifespan, Scott Alexander writes:

Algernon’s Law says there shouldn’t be easy gains in biology. Your body is the product of millions of years of evolution – it would be weird if some drug could make you stronger, faster, and smarter. Why didn’t the body just evolve to secrete that drug itself?

He is talking about anti-aging research, and wondering why, if there is an easy way to stop aging, humans haven’t already evolved immortality spontaneously. There are many relevant things to say about this, but I think the evolutionary perspective is particularly interesting. Under some circumstances, it might be that immortality is inherently unstable.

The Imperium and the Horde

Suppose that it’s the future, and the FDA just approved a pill that makes you immortal. Of course people disagree about whether one should take the pill or not. As a result, humanity is now divided in two populations: the Immortal Imperium, who took the immortality pill, and the Horde of Death, who still experience the painful decay and death we all know and love.

Artist depiction of the Horde of Death.

So, people from the Horde spend their time having plenty of children to populate the next generation, while people in the Immortal Imperium try to escape their existential ennui by reading speculative blog posts on the Internet. Who will prevail?

Two orders of fitness

There are two competing phenomena at play here. One is first-order selection, which is how many of your genes are passed on to the next generation, the more the better. For the Horde of Death, there is nothing mysterious: they reproduce, then they die, and an uncertain fraction of their genes gets passed on.

What about immortal people? They don’t really pass anything to the next generation, because they don’t do the whole generation thing. On the other hand, all of their genes will still be around centuries after centuries, so for the genes involved, this is a 100% success rate. In this sense, people in the Immortal Imperium have a very high first-order fitness.

The second process is second-order selection. This is selection on evolvability. This is about how easy it is for your lineage to improve its own first-order fitness in the future. If a lineage finds a way to evolve quicker, then it may eventually take over the whole population because it will be more likely to discover new beneficial variants, and the original mechanism that granted better evolvability will hitchhike with these new variants.

If you want to see it happen with your own eyes, look at Richard Lenski’s long term evolution experiment, where people have been growing the same E. coli lineages continuously since 1988. Among the mutants that took over the population after a few thousands of generations, some were present since almost the very beginning. They are called EW, for Eventual Winners. Other mutants from the same period eventually disappeared, so they are called Eventual Losers (EL). Surprisingly, in the early days, the EL were able to grow faster than the EW. But in the long term, the EW did better. That is because the EW had mutations that made them more evolvable: they became more likely to acquire further beneficial mutations that ultimately made them grow faster than the EL. People in Lenski’s lab replayed the competition over and over, and most of the time the more evolvable strain ended up taking over.

An external file that holds a picture, illustration, etc.
Object name is nihms-322983-f0003.jpg
From Woods et al., 2011. EL: eventual losers. EW: eventual winners.

Second-order selection matters most for organisms that are not well-adapted to their environment. After all, if you are already at the top of the fitness landscape, there is no point improving your gradient-climbing abilities. Intuitively, it may look like humans are well-adapted to their environment, because we deliberately modified our environment to match our needs. But in a biological sense, current mortal humans are absolutely not well-adapted to their environment. In the First World, fertility is at an all-time low, yet we have all the resources we would need to have tons of offspring. In terms of sheer gene-copying, there is clearly a lot of room for improvement. In fact, there is a lot of genomic evidence that humans are currently under high selective pressure.

(Here is a fun way to think about it: consider that contraceptives are basically antibiotics for humans, in that they are chemicals that prevent us from reproducing. What do bacteria do when exposed to antibiotics for a long time? They evolve antibiotic resistance. So if someone gets a mutation that makes them resistant to contraceptives, they will have a fitness advantage. Realistically, we would quickly notice and switch to other contraceptives, so it’s not likely to be a large issue. But what if people get mutations that increase their parental instinct instead?)

Will the Horde win in the long run?

While the Imperium has better first-order fitness, they are pretty bad at evolving. It is likely that they’ll stop having children to avoid over-crowding. In that case, they just stop evolving completely.

Zardoz Is The 70s Sci-Fi Movie Sean Connery Wants You To ...
If I remember well, that’s what the immortals in Zardoz do. That’s enough evidence for me.

Meanwhile, the Horde does a full cycle of variation/selection/reproduction every 30 years or so. This makes them pretty effective at discovering beneficial variants and become more adapted. To makes things worse, humans have a tendency to constantly change and remodel their own environment. This would explain why the rate of human evolution became higher in the last few thousands years: civilization is changing all the time. Our genomes are always adjusting to human-made changes in technology, environment, agriculture and social organization. The Horde would have no problem finding new genomes to stay up-to-date. The Imperium must do with the same old genomes they have had since the late 21st-century. For example, it’s easy to imagine that the mortals can physically adapt to global warming, while the immortals will not have this chance.

If the Immortals do continue to have babies, their second-order fitness is still pretty bad: if the centuries-old generations still reproduce or mate with the newer generations, the average generation time is still much longer than the Horde’s, so they still evolve slower, and it only gets worse as the population ages. Also, the original immortals still have to compete for resources with the younger, better-adapted immortals, so we are back to the problems we had with the Horde.

Cultural evolution

Anyways, genomic evolution is only one part of the picture. There is also cultural evolution, which is how cultures with higher fitness reproduce (or spread) faster, selecting for more adapted cultural norms. The main reason why humans are so good at colonizing everything is that cultural evolution is faster and more efficient than genetic evolution, so that’s an important thing to have.

For the Horde of Death vs Immortal Imperium conflict, I am not sure who would benefit more from cultural evolution. On one hand, the Imperium has a lot of experience. They have seen everything and had plenty of time to discuss every problem. They have all read the Sequences. They have maximum wisdom.

On the other hand, the Horde gets fresh brains. We all know that young scientists and mathematicians tend to do the most groundbreaking discoveries, and that scientific fields tend to have booms in creativity following the death of established leaders. So what happens if they never die?

Here is a hint of evidence from tennis: when composite rackets were introduced, it altered the way people play in a subtle way, so that the previous way to play was no longer optimal. According to that one study, older players had trouble adapting to the new way to play, which favored younger players. I don’t know how well it generalizes – at the very least, it implies that the Imperium would suck at tennis.

Another hint of evidence: moral values seem to be acquired at young age. When asked moral dilemmas (is it ok to eat the corpse of your pet after it was killed by a car?), people are more morally conservative when the question is asked in their native language, as opposed to another language they learned later in life. This suggests that some of our beliefs and values are shaped by the things we learn while our brains are still developing, and it is not clear whether that can ever be fully overwritten. Perhaps it will be much more difficult for the Imperium to update to new moral norms, which would hinder their cultural evolution. If the Overton window remains stuck in the same place, it would also hamper technological progress: at some point, the Immortals will see all the new gadgets the Horde constantly comes up with, and of course find them absolutely disgusting and immoral.

Elephants and mice

Altogether, it is hard to tell how humanity would continue its evolution if we discovered a way to immortality. There is a decent chance that an immortal population is inherently unstable, but there could also be cultural workarounds. One possible path that I didn’t explore is that mortals and immortals end up occupying different ecological niches. Elephants are practically immortal compared to mice, yet both of them coexist without out-competing each other. If the Horde and the Imperium ever reach such an equilibrium, what would their respective niches look like?

One last quirk: what if the Immortal Imperium, in a last-resort strategy, decides to put immortality drugs in the Horde’s drinking water? Then the Horde become immortal too, and lose their second-order advantage. Problem solved. Unless, of course, people start developing resistance to the immortality pills – such a mutation would be selected for because it helps selecting for mutations that help selecting for mutations that are beneficial. I have never heard of any third-order selection occurring in nature, but maybe humans will make it happen.

bookmark_borderIs top-down veganism unethical?

« Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away. »
– Antoine de Saint-Exupéry

Remember how the vegetables we eat everyday are very different from their ancestors from a few centuries ago? The same is true for animals. In half a century, farmers bred increasingly large races of chicken. Here is a comparison of the size of bones for modern and ancestral chickens:
Scale bar: 2 cm (source)

The leg on the left belongs to a modern broiler chicken. The one on the right belongs to a wild jungle chicken.

From the perspective of meat production, this is an improvement. From the perspective of animal suffering, things are more uncertain. Contemporary chicken are reaching pantagruelian proportions and now they have trouble walking and their legs often break under their own weight. One might even go as far as worrying this is a little bit unethical. Fortunately, there are solutions. I can think of three of them – the first two, you already know. The last one, however, I never see discussed anywhere.

1. Non-meat

The most fashionable solution now is to replace meat with plant-based construction materials that are claimed to look and taste similar to meat. My main problem is that plant-based meat is, at best, overlapping with real meat: the best-quality plant-meat is comparable to the lowest-quality meat. If you think the vegan burgers make accurate simulacra of meat, I’m afraid you are eating too much heavily processed shitty meat. We are still very far from the impossible® A5-rated wagyu, the impossible® pressed duck, the impossible® volaille de Bresse “en vessie” (which must be gently cooked in a plant-based impossible® pork bladder to be valid). As a typical Westerner, i have the opportunity to eat only about 90,000 meals in a lifetime, there is no way I’m wasting any of them on sub-delicious food. Still, this approach deserves some praise for actually existing and working, which cannot be said about the second approach –

2. Lab-grown meat

To be fair, the interest in lab-grown meat is increasing, slowly and steadily. Perhaps it will eventually catch up on sexbots. Here is a Frontiers review from last year, whose title alone drives the point home: “The Myth of Cultured Meat”. It is not that bad, really, but the current prototypes look like attempts at emulating the vegan attempts at emulating real meat. I don’t see any lab-grown marbled beef appearing in the foreseeable future.

3. Top-down vegan meat

Lab-grown meat was the bottom-up approach. Here, I will inquire into the feasibility of a top-down approach. Rather than starting from cell cultures and engineering them into a sirloin steak, I suggest starting from whole animals and using genetic engineering to remove all the things we find ethically questionable, one by one. Our end goal is, of course, to turn the live animals into warm, squishy, throbbing blocks of flesh devoid of anything that could possibly be construed as qualia. If we can give them a cubic shape for easy packaging and storage, that’s even better.

The path to success is long, but straightforward:

Perhaps the easiest, short-term solution is to make the animals insensitive to pain. We’ve known for a long time that some genetic variants in humans make pain disappear completely. The most famous one, a mutation in the gene SCN9A, was discovered on a Pakistani street performer who would literally eat burning coals and stab himself for the show (he did not live very long). Earlier this year, Moreno et al managed to make mice insensitive to pain using a CRISPR-based epigenome editing scheme (basically, they fused an inactivated Cas9 to a KRAB repressor, so it binds to the DNA just next to the SCN9A gene and inhibits transcription). As we can see from the street performer kid, disrupting the pain sensitivity pathway is totally viable, so I see no technical reason we couldn’t try that on farm animals too.

Of course, pain is not the only form of suffering. If we really want to persuade the PETA activists, we might want to make the animals permanently happy, whatever the circumstances. This is where it gets tricky. I found this genome-wide association study which identifies variants associated to subjective well-being in humans, but it’s not clear whether these variants have a direct effect on happiness, or if they just make you more likely to be rich and handsome. In the later case, it would not be particularly useful for our next-gen farm animals (it can’t hurt, though). It is pretty clear that some genetic variants have a direct effect on personality traits like depression and anxiety, so maybe there is room for action. To optimize happiness in farm animals, we would of course need a way to measure the animals’ subjective well-being, so that’s another obstacle in the way of convincing the vegans (vegans, I’ve been told, can be extremely picky). Also, there is another problem: if we find a way to make animals permanently happy, we might be tempted to apply it to ourselves instead, and then, nobody will care about factory farming anymore.

If removing pain and sadness is not enough, the next logical step is to get rid of consciousness entirely. Any chemical used to induce coma is probably not an option, since we don’t want people to fall into a coma themselves after eating lunch (I’m already close enough to a comatose state after lunch with regular food, let’s not make this worse). A more radical approach is just to remove as much of the nervous system as possible. In humans, there is a rare condition called anencephaly where a fœtus develops without most of the brain, and in particular without a neocortex. It is pretty clear that these kids have no consciousness, yet they can survive for a few hours or even a few days. There is also evidence that some mutations or recessive variants can trigger anencephaly, so the prospect of developing animal lineages without a cerebrum does not seem completely impossible. A major challenge, of course, would be to extend the life of the organism for more than a few hours. Moreover, it would require a lot of effort from the marketing department to make such a monstruosity appealing to consumers.

Sadly, this will not be enough for most vegans. Most of the vegans I personally know put the edibility frontier somewhere between the harp sponge Chondrocladia lyra and the egg-yolk jellyfish Phacellophora camtschatica, that is, anything with a nervous system is formally off-limit. This criterion does not make things easy for our master plan: we can remove as much of the nervous system as we can, I can’t think of any way to get rid of the cardiac automatism or the part of the nervous system in charge of respiratory function. Unless, of course, we dare enter into cyborg territory. Is the world ready for alimentary cyborgs? The future is full of surprises.


Let’s be honest, this post started as fun speculation and gratuitous vegan trolling, but I am actually very serious about the central point. GMOs are mainly discussed in terms of cost, environmental impact or health properties, yet very rarely as an avenue to reduce animal suffering. Many of the ideas discussed here are still beyond what is possible with our current understanding of genetics. Still, we can already identify some interesting research paths that are just waiting to be explored. So, what makes this approach so disturbing? As often, the moral questions turn out more difficult than the technical barriers. The major obstacle is not so much the actual genetic engineering, but the lack of good metrics for success – how do you even measure suffering to begin with? On the other hand, if the outcome of a problem cannot be measured or even defined in any meaningful way, maybe it does not matter that much, after all. I would be happy to hear what ethical vegans think about the general approach. What would it take for a top-down reduction of animal suffering to be acceptable to you?

bookmark_borderThe Holy Algorithm

As it will surely not have escaped your insight, this weekend is Easter. Why now? The date of Easter is determined by a complicated process called the Computus Ecclesiasticus. I will just quote the Wikipedia page:

The Easter cycle groups days into lunar months, which are either 29 or 30 days long. There is an exception. The month ending in March normally has thirty days, but if 29 February of a leap year falls within it, it contains 31. As these groups are based on the lunar cycle, over the long term the average month in the lunar calendar is a very good approximation of the synodic month, which is 29.53059 days long. There are 12 synodic months in a lunar year, totaling either 354 or 355 days. The lunar year is about 11 days shorter than the calendar year, which is either 365 or 366 days long. These days by which the solar year exceeds the lunar year are called epacts. It is necessary to add them to the day of the solar year to obtain the correct day in the lunar year. Whenever the epact reaches or exceeds 30, an extra intercalary month (or embolismic month) of 30 days must be inserted into the lunar calendar: then 30 must be subtracted from the epact.

If your thirst of knowledge is not satisfied, here is a 140-page document in Latin with more detail.

As far as I understand, during the Roman Era the Pope or one of his bureaucrats would perform the computus, then communicate the date to the rest of Christianity and everybody could eat their chocolates at the same time. Then, the Middle-Ages happened and communication became much harder, so instead they came up with a formula so people could compute the date of Easter locally. Of course, the initial formulas had problems – with the date of Easter dangerously drifting later and later in the year over centuries, and don’t even get me started on calendar changes. Eventually Carl Friedrich Gauss entered the game and saved humanity once again with a computationally-efficient algorithm (I am over-simplifying the story so you have more time to eat chocolate).

But now is 2021, and I’m wondering how they run the algorithm now, in practice. I looked up “how is the date of Easter calculated” but all the results are about the algorithms themselves, not about their practical implementation. I have a few hypotheses:

  1. There are responsible Christians everywhere who own printed tables with the dates of Easter already computed for the next few generations. If your Internet goes down, you can probably access such tables at the local church.
Here is such table from 1907 (Wikimedia commons)

Of course this does not really solve the problem: who comes up with these tables in the first place? Who will make new ones when they expire?

2. There is a ceremony in Vatican where a Latin speaker ceremoniously performs the Holy Algorithm by hand, outputs the date of Easter, prints “Amen” for good measure and then messengers spread the result to all of Christianity.

3. Responsible Christians everywhere own a Computus Clock, a physical device that tells you if it is Easter or not. When in doubt, you just pay a visit to that-guy-with-the-computus-clock. Then, it is like hypothesis 1 except it never expires.

4. There is software company (let’s call it Vatican Microsystems®) who managed to persuade the Pope to buy a license for their professional software solution, Computus Pro™ Enterprise Edition 2007 – including 24/7 hotline assistance, that only runs on Windows XP and they have a dedicated computer in Vatican that is used once in a while to run these 30000 lines of hard Haskell or something. Then, it goes just like hypothesis 2.

(Of course, all of these solutions are vulnerable to hacking. It might be as easy as sneaking into a church and replace their Easter tables with a fake. A talented hacker might even have it coincide with April fools.)

If an active member of the Christian community reads this and knows how it is done in practice, I am all ears.

Anyways, happy Easter and Amen, I guess.

bookmark_borderThe Great Happiness Filter


The universe is full of wandering celestial bodies covered in complex, superintelligent lifeforms who engage in eternal masturbation.

That’s it. That’s the Great Happiness Filter.


This idea has been discussed before, but it strikes me that people still struggle to find other solutions to the Fermi paradox, as if the Great Happiness Filter was not already explaining everything. Sure, synthetic biology going out of control, AI going out of control, cosmic superpredators going out of control, Earth-is-a-zoo, all make great film plots. But that is what they are: great film plots. We hear about them because they are on display in sci-fi films and books. You know what Great Filter hypothesis would make a terribly boring film? Aliens sitting in the dark with a wire plugged straight to the happiness center of their brains, doing nothing, while a combination of robots and nuclear fusion takes care of all the logistics1An earlier version of this post was about stimulating the pleasure centers of the brain. But stimulating happiness should work just as well (just replace dopamine with serotonine, something along this line). It also makes the idea a bit scarier, for some reason.. All they have to do is figure out which neurotransmitters and which parts of the brain are involved in the “I’m happy” sensation, and find a way to stimulate it. Save your philosophical arguments about whether they are really happy or if they just have the illusion of being happy, it does not matter for the following.

You might object, this is not optimal, eventually they will run out of whatever fuel they are using, or their sun will turn into a red giant, so they should still try to expand and obtain more energy. As Robin Hanson puts it:

After all, even navel-gazing virtual reality addicts will likely want more and more mass and energy (really negentropy) to build and run better computers, and should want to spread out to mitigate local disasters.

This rests on the assumption that intelligent civilizations will necessarily try to fully optimize masturbation. They won’t. Compare this to heroin addiction: all things considered, heroin addiction is far from being the ultimate hedonistic experience (quite the opposite), yet many people still get trapped into it. You don’t see heroin-addicts building Dyson spheres to make sure they have a sustainable feed of high-quality heroin forever in the future. This also applies to masturbation. For the Great Happiness Filter to occur, you don’t need a perfect self-sustaining planetary masturbation system. You just need to reach the threshold were masturbation is just the right amount of good, so it’s not worth working to fix the flaws of your current masturbation scheme, because you would need to stop masturbating in order to do that. Past this threshold, intelligent lifeforms will not try to improve their masturbatory experience anymore, and will just chose to masturbate instead. Maybe there will be a warning, like as little red icon on the lifeform’s internal brainscreen that will say “warning: your fusion reactors have almost turned everything into iron already, please plug the system to a new planet”, but who cares at that point? You can just ignore the warning and enjoy maximal sensation of fulfillment and satisfaction. This is basically a sink point.


The real difficulty with the Great Happiness Filter is the order in which the relevant technologies are discovered. If robotic servants and artificial general intelligence are developed before happiness-pods, then there is a chance that we get to stay in the pod while a robotic butler continues to improve the experience on our behalf. In that case, we are back to the Fermi paradox and space colonialism, because our butlers will try to maximize the energy we can spend on happiness-pods. Eventually, our civilization’s masturbation-maximizers might conflict with other civilizations’ paperclips-maximizers (not all civilizations can be as wise as ours), leading to cosmic-scale battles. On the other end, it is likely that happiness-pods are available to everyone before we get the appropriate energy source to sustain them. In that case, the intelligent lifeform will quickly go broke and possibly go extinct (as if the entire humanity got addicted to heroin at once). But between these two extremes, there is a large sweet stable spot where there is enough automation to power the happiness-pods and make sure everything is running well, but not enough to expand and reach for new planets. As a result, we get plenty of silent planets covered in happy masturbating lifeforms, traveling through space at speeds beyond imagination. Remember this when you look at the stars.