Old Minds Are Fragile

Consider three design problems:

  • Case #1: First, you are asked to modify a stock car, making it into a truck to haul stuff. After you do that, you are asked to create a race car. Which would you rather start from for this second task, another stock car, or the stock car that you turned into a truck?
  • Case #2: A species of beetle lives in a varied and changing environment, and so has a rather simple and basic design. Some of these beetles invade a different and more stable environment, and acquire adaptations specific to that environment. A third rather different but also stable environment opens up adjacent to both previous environments. Which beetle type’s descendants will likely fill this third environment?
  • Case #3: Over the last decade a group wrote software to do a certain task (e.g., print driver, web server, spreadsheet, etc.) This design of this software was matched to certain features of the problem environment, such as hardware, network speeds, etc. Today there is a need for software to do a similar task, except that the problem environment has changed. To write this new software, would you have your team modify this previous software, or start a new system mostly from scratch?

In all these cases, one makes a system to function in a given environment, and can either modify a complex system adapted to a different environment, or “start over” via modifying a simpler system less adapted to any specific environment. In general, the more different is the new environment from the old, the better it is to start over. Old  systems tend to be rigid, which makes them fragile, in that they break if you bend them too far.

This suggests that designed systems tend to get irreversibly fragile as they adapt to specific environments. When context changes greatly, it is usually easier to build new systems from “scratch,” than to un-adapt systems designed for other contexts. Software tends to “rot“, for example.

An empirical prediction here is that species occupying highly variable environments tend to have more descendant species in other environments, compared to species occupying less variable environments. I don’t know if this has been tested. It fits with the Innovator’s Dilemma though, where firms who serve the low end of a product line with simpler techs tend to creep up and displace those serving the high end; high end products tend to be more complex.

Today I’m focused on this being bad news for the feasibility of immortality, at least for human-like creatures. You see, our minds seem designed to adapt to the environment in which we grow up, via youthful plasticity transitioning to elderly rigidity. For example, we are great at learning languages when young, and terrible when old. We are similarly receptive when young to new ways to categorize and conceive of things, but once we have often used particular ways, we find it harder to understand and use alternatives.

The brains of most animals peak in functionality during their key reproductive years, and do worse both before and after. Short lived animals peak sooner than long lived animals. Some of the early rise is due to learning, and some of later decline is due to the decline of individual cells and connections. Some of this pattern may even be due to an explicit plan to turn up some dials on plasticity early on, and then turn down those dials later. But I think another important part of this rise and fall is due to a general robust tendency for adapted systems to slide from plasticity to rigidity.

Thus even if we succeed in creating emulations of whole human brains, “ems” which can use backups, body swaps, etc. to avoid bodily death and decay, we should expect such ems to decay by getting mentally rigid with subjective age. Even if we do not emulate any decline in individual cell and connection performance, nor any age-specific general plasticity dial settings, the mind itself may well decay with subjective experience, because such decay is just intrinsic to mind design.

Now in software design one can often slow a slide to rigidity by refactoring code, such as by looking for better abstractions to achieve modularity. But the brain probably already has some analogues to refractoring, such as in its ways to reorganize concepts. And even with large refactoring efforts, most designed software eventually gets rigid, so that when environments change enough such software is replaced wholesale by new systems built from scratch.

Similarly, em workers who start out subjectively young, and then learn how to work in a stable environment, may become increasingly productive in that environment, even after thousands of years of subjective experience. But when a new quite different work environment appears, one can probably gain more work productivity by training subjectively young ems for it, rather than trying to change ems who had spend thousands of subjective years adapting to a very different environment.

Today most houses and cars are in principle immortal, in the sense that enough maintenance can keep them functioning indefinitely. Yet most houses and cars are not immortal in practice, because those maintenance costs keep rising to the point where it is cheaper to build new houses and cars. Similarly it might be possible to keep very old ems around, even when they have become much less productive because relevant environments have changed. Someone, however, would have to pay that cost, relative to the option of using more productive younger ems. And as with houses and cars today, maybe few will pay.

If you personally hope to become an em with an especially long productive subjective life, it is probably important to stay general and flexible for as long as you can. Prefer to acquire habits and insights that are widely applicable, and whose value is likely to long continue. Prefer to write, deal with people, and manage complexity, rather than learning the detailed layout of a city or how best to write in a particular new programming language.

Eventually we may find mind designs with a much weaker tendency toward rigidity with age. And we may find ways to transfer some important elements of once-human minds, such as their memory and personality, into this alternative framework. But even then there should be some aging. And it gets even less clear if you’d want to think of such a changed creature as you.

Even more eventually, the universe should get a lot more stable, and with it the environments where minds function. Then there will be a lot more scope for very long lived human-like minds. If there are any human-like minds left at that point.

Added: Stem cells fit this; bodies usually make cells designed for specific places from general simpler stem cells, not by changing other specific cells.

GD Star Rating
loading...
Tagged as: , ,
Trackback URL:
  • rs

    I wonder how specializing in pure mathematics would fare in the long term mind viability spectrum between the specialization on technical knowledge with a large contingent component and the specialization on complex social interactions between present-day humans.

    • Hook

      Success in pure mathematics is aided by complex social interactions, and I’d hesitate to underestimate its contingent component.  That said, as long as one pursues a social form of mathematics study where one learns to combine the strengths of others, it doesn’t seem like a bad idea.

  • Mark M

    Fortunately, once immortality becomes generally available everyone else will have that choice as well.  As we get older we’ll have the help and comfort of those who became immortal with us – people our own age.

    While the infrastructure that underlies transportation and communication will change over time we’ll keep the things that make operating in that structure familiar.  (We’ll keep the same or similar user interface). 

    It would not be wise to ignore the market potential of the oldest of the immortals. 

    AND GET OFF MY LAWN!!  Kids today, sheesh…

    • http://overcomingbias.com RobinHanson

      Yes, the longer that people can live, the more incentive to not change user interfaces old people are used to. But I doubt that effect will be enough to keep old minds productive indefinitely.

  • http://profiles.google.com/slaanesh Brandon Reinhart

    What is the quality of the evidence for declining neuroplasticity with age? I had recently read that brain rigidity claims are questionable and that the brain more easily adapts with age than previously thought. I have seen it said (and need to find references and get a better grasp on the quality of the claims) that there are a host of reasons why we might observe fewer adults picking up new skills or belief-sets. In particular, the time involved in practice and training or research and study. As well as things like the social costs of switching position on long held beliefs. I don’t know what story most closely matches the facts of the matter.

    I’m not sure if far-future meat brains are that much different than far-future software brains in this regard. Both could be more or less plastic based on the availability of tools to address the causes of rigidity. It seems like em-builders would want their ems to be rigid in cases where those ems are assigned specific, well defined tasks (ems booted up from em-candidate minds that are already rigid thinkers well trained in a particular method). Although we might desire to think of us wanting to employ flexibly minded workers, it seems in reality many companies want easily managed, rigid, focused workers. If companies don’t see innovation as originating within the ranks, but something that the leadership is tasked with providing, then there is no incentive to create ems that need to learn much beyond their origin state. Also if the cost of creating new ems is low, you would spin up a new em for the new task (which is what I think your post is about in part). Also, there may be incentives in having ems who do not or are slower to adapt hedonically.

    So, whether or not there is declining neuroplasticity in human meat-brains, there is an argument for not having much neuroplasticity in emulated human brains. If you think the future will be one of large numbers of emulated humans, then you might expect to see a future with low neuroplasticity.

    • http://overcomingbias.com RobinHanson

      Yes, once an em has been trained to do a particular task there may not be much need after that for flexibility. But new kinds of tasks will continue to arise, and ems will need to be trained to do them, and will need flexibility during such training.

      On who chooses, labor markers are an intersection of supply and demand – the preferences of employers who want workers and pf people willing to work both matter.

  • Dave 944

    Right now the problems you are talking about are solved by death,programed death for the most part  these days.

    I have never been able to conceive how we could be able to segregate your described socially pertinent brain features,which can be encoded in words,such a memory and personality from the non-word realities that are personal and  noncommunicable, or are so socially trivial that they remain unspoken and unremembered except momentarily. Seemingly your ems would only manifest things that are socially important to you as an intellectual. But they could retain so much extra cerebral data receiving capacity,they would be  creatures in a state of sensory isolation. Surely this would be a hellish state.

    Most likely your ems would never experience anything that you could not encode verbally or symbolically.They would more resemble libraries than beings.Ok now that we are in the computer age,they would resemble computers. So they could simulate beings when viewed from the outside.
    I don’t want to be an em.  

  • Ely Spears

    The analogy I prefer is that of simulated annealing. To find a reasonable local optimum in the space of preferences, social habits, skills, etc., requires a lot of exploration of solution space (I prefer to call this exploration “playing” or “pretending”). We seem to solve it heuristically, proposing new mental configurations, testing them out, and only ever randomly accepting inferior mental poses when other circumstances are right. Our tolerance for such random acceptance of inferior poses diminishes with age, much like a cooling term in simulated annealing.

    I fear these changes in myself. I often wonder how much is brain plasticity and how much is lack of playing in adulthood. Are elderly parents’ minds more plastic than childless peers? I wonder if a lot of it isn’t social. In my hometown, for example, trying new things and having a high dynamic range in idea space are viewed as socially wrong, even for people in their mid twenties. I’m curious about whether these casual factors play a role. Are urban elderly more plastic than rural elderly? By how much?

    • http://daedalus2u.blogspot.com/ daedalus2u

       I think the problem with using a simulated annealing model is that the “solution” has to already exist as a possible state that the system can enter.  

      Before the solution is known, it is unknown how to instantiate that solution as a state in the system seeking the answer.  Unless the solution is already present in the solution space being interrogated, a simulated annealing-type approach will fail.  In other words, if the solution is orthogonal to all existing or reachable mental states, that entity will never be able to reach it.  

      The problem you pose of certain social groups being unable to entertain ideas outside their comfort zone is a real and very large problem.  It is probably the main problem that is impeding progress in all fields today.  Real potential solutions simply cannot be considered because the Very Serious People (to use Krugman’s term) simply cannot conceive of them being correct, no matter what the facts are.  Throw a little fraud, motivated reasoning and some signaling via faux agreement as in the Emperor’s new clothes, and real solutions become impossible. 

      • Ely Spears

        Simulated annealing would not “fail.” It would simply converge on some local optimum in a space of possible states. If the global optimum isn’t in that space, oh well, who cares? No one would seriously claim that human preferences, skills, or social acclimations are “globally optimal.” and since the environment is always changing, it’s a moving target anyway.

        The analogy here is with the proposal function, and willingness to move away from what you currently consider best to possibly get somewhere better. Less willingness to do that is like being in a cold tail of a Markov chain of tried options. “Set in your ways” so to speak.

  • Robert Koslover

    This was a depressing but insightful post.  Achieving immortality would seem to have many obstacles, and your new post now strongly suggests that it may not be particularly enjoyable even if it can be achieved.  Sigh.  I am once again reminded of something my aging mother said in regard to our rapidly-changing world:  “Thank heavens I’m not living nowadays!”

    • http://overcomingbias.com RobinHanson

      I’m pretty sure that such a long life would be enjoyable at most moments. It is in far mode, looking on the whole thing, where it seems less impressive and hence less worthy.

      • Dent

        “I’m pretty sure that such a long life would be enjoyable at most moments.”

        Why is that? Do you think the societal organization will humanely offset the ill-adjustment of aging ems by giving them free resources, or do you think the ems will simply amp up their pleasure pathways to offset the harshness of their ill-adjustment?

  • mjgeddes

    A mind is only fragile if its not a true general purpose intelligence
    once we know what the most general powerful abstractions are we simply program these as fixed classes in our domain model and no further top level refactoring is needed.
    Thats why years ago i realized that ontology is the key to true AGI and i sought to identify the deepest most general categorizations of thought.
    After years of deep thought I identified 27 ontological primitives that appear to be irreducible
    I guess a mind design that incorporates some finite set of sufficiently powerful ontological prims in the top level domain model is immune to mind aging and remains flexible forever

    • Awoodside

      Do you have these 27 primitives written anywhere? I’d like to read them.

      • mjgeddes

        detailed info is classified at this time but look here 
         http://zarzuelazen.com/27categories.html

  • http://profiles.google.com/rjvg50 Kirk Holden

    I have used “The Argument from ARM” in the fashion of your three examples. Roughly, the Intel Atom is a culmination of a line of complex machines built to run Windows and MS Office correctly on notebook PC’s with keyboard, mouse and large, spinning hard-drives. ARM builds machines that run Linux with the smallest collection of simple circuit modules. There are versions of ARM that go into cell phones, versions that go into internet toasters and versions that (now) support blade servers. Intel will continue to start with a Pentium to build a cell phone or a server and will not even try to build a cell phone (actually they will try and try and try – they just will not succeed). If I start with an ARM7 from a cell phone call processor (the oldest function is to turn 7 digit keypad input into voice) and add the enhancements to run a blade server I will end up with basically the same CPU that runs my Samsung Tab 10.1 — there are generally only so many good tricks in computer science. But if I start with an Atom and build a blade server the end result will not be a Xeon blade server — there are an infinite number of ways to solve this complex series of transformations with bubble gum and bailing wire. So I plan to put my EM on an ARM quantum computing platform and live forever running a standard release of Linux.

  • Arch1

    I’m not an economist so this is probably overly simplistic, but our current society constrains (in a somewhat hit-or-miss fashion) economic growth in the service of other things we value (quality of life, freedoms).

    If ems generally value certain things (e.g. non-obsolescence of old ems) more than the incremental growth lost by enabling those things, couldn’t they do the same?

  • V V

    … To write this new software, would you modify this previous software, or have a team start a new system mostly from scratch?

    Modify the existing software.

  • V V

    I don’t see your point.

    Yes, humans learn faster when they are young. Do we euthanize people as they turn 40?

    Why would a society of brain uploads do that?

    I don’t think that immortality is possible, at least not for anybody who is currently alive, but I don’t find your argument compelling.

    • daedalus2u

      It isn’t the result of a society of brain uploads, it is the result of a society that values property rights above all else.

      If you can’t afford what to pay what someone with property wants to sell that property for, then you do without.  If you need that property to survive and can’t buy it, you don’t survive. 

      In such a society, when ever an em can’t pay the going rate for information storage, they will be erased to make space for an em who can. 

      Uploaded brains will probably be among the first to go because they have such a large overhead. 

      • V V

         It seems that Hanson dystopia essentially entails a society where everybody slaves at subsistence wage to produce stuff (which is supposed to be bought by somebody, but I don’t get who they would be, since everybody is a subsistence earner).
        Anybody who hasn’t top performance just starves to death.

        That’s obviously not what happens in our current, first-world societies, and I don’t see how transitioning from meat bodies to silicon bodies would make that scenario likely.

      • daedalus2u

         That is not what has happened yet, but in the race to the bottom, the wealthy are trying to make it happen. 

        What happens to people who can’t afford health insurance? 

        That is essentially what happened in Ireland just before the Great Potato Famine, which is why so many died.  They were too poor to buy the food they were growing for the land owners. 

      • http://entitledtoanopinion.wordpress.com TGGP

        The idea is that it would be incredibly cheap to replicate ems and there would be great opportunity to do it. The population would explode until Malthusian limits are reached. Then once life is cheap, death will be cheap. My impression from Hanson is that the result is pretty much invariant with respect to property rights (unless there is a Singleton which would be capable of enforcing some other plan on all agents), it’s a matter of more primitive Darwinism.

      • V V

        Most people in developed and developing (e.g. China) countries actively limit their fertility, and this has avoided the Malthusian catastrophe so far.

        It’s true that there are countries where people live near Malthusian limits because they don’t limit their fertility, for a variety of reasons (unavailability of reliable birth control and relevant information, irrational religious or ideological beliefs in favour of fertility, rational arguments in “Tragedy of the commons” scenarios). Note that in these countries virtualy everybody is employed in food production.

        It is possible that extreme economic crisis extends this Malthusian subsistence economy to the first-world, but in a scenario where there is no economic collapse and brain upload technology is developed, I don’t see how this is a likely outcome.

      • David

        V V, you mostly underestimate the consequences of person-copying technology on population dynamics.

        It is astonishing how many people don’t see that we are living in a bubble due to excess resources from industrial technology and fossil fuels and limited fertility from contraception. This bubble is both historically unique and unsustainable.

        It is possible to restrict reproduction in a system of complete surveillance. But it is hard to argue that a person who can self-copy, wants to self-copy and is productive enough to pay for it should not have the right to self-copy.

        Since physical conflict is *very* costly in a world of dense mind clusters, agreements over property rights will probably be strong, because no major faction will have an incentive to bear the cost of large-scale physical violence.

        I predict that competition results in

        a) massive innovation and expansion
        b) rapid evolution and self-modifying of digital minds
        c) large clusters of ems who own resource shares and run mostly themselves without evolving or buying anything from others
        d) parasitic algorithms trying to hack the “old” folk in c), rather than physically attack them

      • daedalus2u

         Yes, exactly.  Women do limit their fertility when they have the opportunity, but greedy people do not limit their voracious acquisition of property.  

        A Malthusian end state can come from either case, no limits on population growth other than resource limitations, or actors that use resources to acquire more resources instead of reproduction until the existing population becomes resource limited.    

        The high growth of the 1950′s and 1960′s was due to the high tax rates thwarting Malthusian property acquisition.  Now that the property acquirers have latched onto control of government and are thwarting limits to their property acquisition, we are approaching a Malthusian end state from the other end, from property being owned by a few rather than too high a population.  

        I suspect that it was the Black Death that broke the property acquisition monopoly of the feudal lords and allowed non-Malthusian growth.  World War II did the same.  WWII required such a large effort that the wealthy had to be taxed at high rates and that generated surplus capacity.  That changed the dynamic and then fueled the economic expansion that followed.  It wasn’t so much that there was surplus production capacity, as it was that the greedy lost control of enough resources to maintain control of the rest.  It is pretty easy to seen in this graph by Krugman.

        http://krugman.blogs.nytimes.com/2012/07/25/zombie-straw-men/

        It is the financial sector that is growing at the expense of everything else.  

        http://krugman.blogs.nytimes.com/2012/07/17/finance-capitalism/?smid=go-share

        Which makes no technical sense.  Finance is the system that makes the connection between people with capital to invest and those who need capital to make productive investments.  Computers make finance much easier and cheaper to do.  Unfortunately finance has been taken over by people who are greedy and don’t mind screwing everyone else.  

        http://www.economonitor.com/lrwray/2012/07/23/why-were-screwed/

        If ems acquire the same rights as biological humans while maintaining the same compulsion for property acquisition of the greedy, then biological humans will go extinct pretty quickly due to resource competition.  

        It will be like BSG with the ems playing the role of the Cylons.  But we probably won’t make it to an off-Earth civilization.  Probably not even to where ems are even possible.  The 0.1% are already the Cylons. 

      • V V

        @bc10f5a14cefc0eb3a4eccbd579f918d:disqus

        you mostly underestimate the consequences of person-copying technology on population dynamics.

        Why would person-copying technology imply that we will copy ourselves up to the environment carrying capacity like bacteria in a flask?
        Why would I make many copies of myself if I have to share resources with them?

        It is astonishing how many people don’t see that we are living in a
        bubble due to excess resources from industrial technology and fossil
        fuels and limited fertility from contraception. This bubble is both
        historically unique and unsustainable.

        Barring some major technological regression, birth control is not going to become unavailable.
        Fossil fuels and other natural resources are indeed dwindling, and it’s quite likely that the current human population size is unsustainable. This doesn’t mean that we are destined to live in a Malthusian subsistence economy at carrying capacity. We could well settle at a lower population size and relative wealth.

        Anyway, if the Malthusian scenario happens, I don’t think there will be any brain uploads or AIs at all.
        (I think they are unlikely, at least in the foreseable future, even in non-Malthusian scenarios).

      • David

        “Why would I make many copies of myself if I have to share resources with them?”

        This is not about you. Ask yourself: Is there anybody who would be motivated to do so? If there are at least small percentages of people who would, they will quickly grow to encompass most of the population.

      • daedalus2u

         
        David, It isn’t even necessary to
        increase population. If a subset of the population acquires property
        faster than growth of the whole economy, then a subset of the
        population is losing property and eventually will be unable to
        sustain itself.

        This is what is happening now, to what
        was once the middle class.

        When you couple that with unsustainable
        resource extraction (for example use of fossil fuels with no regard
        to AGW), it is an unsustainable economic bubble and when the bubble
        bursts, a lot of the population becomes unable to sustain themselves
        because the carrying capacity of the economy has suddenly shrunk.
        Unless there is redistribution, the have-nots are doomed.

        Physical conflict is only more
        expensive to those who are not going to die because of resource
        deprivation.

      • David

        “David, It isn’t even necessary to increase population. If a subset of the population acquires property faster than growth of the whole economy, then a subset of the population is losing property and eventually will be unable to sustain itself.”
        Deadalus, yes, there are many who already live on subsistence. There is some charity, but also inequality and a lot of waste. However, in a future em world, a lot of that waste may be used differently than now. For instance, the property owning ems can cheaply run many happy ems, including high-speed copies of themselves. The current rich don’t have this option. While inequal, there is a utopic element to it if you care about the happy copies. It’s also not clear that ‘starving’ ems suffer in the way current starving biohumans do. It may just be lower clockspeed, no room for copies etc.

        “Physical conflict is only more expensive to those who are not going to die because of resource deprivation.”

        This would actually limit inequality because the rich would then have to bribe the deprived not to use violence. But often, the deprived masses actually don’t have much violent competence to bargain with, and those who do are selectively hired to suppress the rest. This works well in totalitarian systems that don’t need to care about human rights politically.

        I think in an em future, it depends on how the clusters can decide about weapon use. Weapons of mass destruction would be *far* more dangerous with such wealth density.

      • http://entitledtoanopinion.wordpress.com TGGP

        To re-emphasize again how robust Hanson’s Malthusian prediction is to the political stuff being discussed above, here is his post from 2009. “As long as enough people are free to choose their fertility […] income per capital must fall over the long run, a fall whose only fundamental limit is subsistence.” The Darwinian explanation is that as long as there is some variation (say, in resistance to fitness-reducing birth-control) the more fit will replace the less fit. In evolutioanry terms, the demographic transition happened extremely recently and there hasn’t yet been much adaptation in response. This result doesn’t even rely on ems, although that would be expected to bring up such a result sooner (one of the most important things about ems is that they are faster).

        daedalus2u, this blog has earlier discussed the effect of the Black Death on serfdom here. Actually, you’ll have to click through to Paul Krugman’s original piece explaining that the common belief that the Black Death caused the end of serfdom is wrong. If you want to know what caused the end of the Malthusian era, you’ll have to read Greg Clark’s “A Farewell to Alms”. Economic growth only started outpacing population growth in the late 1700s with the industrial revolution.

        “Physical conflict is only more expensive to those who are not going to die because of resource deprivation.”
        Is that some kind of “what does not kill you only makes you stronger” kind of thing? Nietzche’s saying is false for both people who are permanently crippled and the Byzantine & Sassanid empires which exhausted themselves fighting each other, creating the perfect opportunity for Arab expansion. Resources one em (or its clan) uses up in conflict can’t be used for further replication or competition with the lucky ems who’ve sat out.

      • http://juridicalcoherence.blogspot.com/ Stephen R. Diamond

        “As long as enough people are free to choose their fertility […] income per capital must fall over the long run, a fall whose only fundamental limit is subsistence.” The Darwinian explanation is that as long as there is some variation (say, in resistance to fitness-reducing birth-control) the more fit will replace the less fit.

        If that really came to pass, I would anticipate state measures stronger than limiting fertility. A eugenics program would ensue, culling out of the population those disposed to high fertility. This might be made severe enough to outweigh the effects of the deviants who can’t be controlled.

        But I’m biased: I, like some others, consider Robin’s vision a dystopia. I’m not partial to eugenics, but given a choice… As II understand Robin’s utilitarianism, however, it’s a utopia consisting of trillions of centers of sentience at least marginally “happy.” So, both sides may be engaging in wishful thinking in this exercise in ultrafar thinking.

      • David

        “A eugenics program would ensue, culling out of the population those disposed to high fertility.”

        This implies genetic determinism of reproductive behavior. I actually think this is quite false, cultural aspects overwhelm this. Availability of contraception and social incentives play a larger role, though there are certainly genetic differences. I also dislike the jump to “culling” even though there are milder degrees of coercion such as taxing high fertility or subsidizing low fertility.

        Important: Memetic evolution is faster than genetic. For a new biological generation, there is a childhood stage of at least 12-14 years before pregnancy can commence. For ems, reproduction would be practically instantaneous. And all the memes are copied directly – the beliefs about reproduction, the behavioral habits, the personality traits etc. This is still a Darwinian explanation, just not on the genetic level.

        “That claim entails that r-selected reproductive strategies (“quantity over quality”) always dominate K-selected reproductive strategies (“quality over quantity”).”

        No, it doesn’t. But it it striking that direct person-copying has much more in common with r-selection than K-selection. After all, K-selection is all about getting few long-lived offspring to be successful in competing for resources. And with digital minds, the copying of an experienced person is essentially free.

        “They would quickly reach near-subsistence level and then replicate much more slowly.”

        But near-subsistence level means that the copies compete for subsistence wages with the rest of the population, driving everyone else’s wages down. If the rest doesn’t find that life appealing, they won’t reproduce as much as the ems who accept it and spend all their available income on reproduction. The outcome is exactly the Malthusian one Hanson predicts. If property rights are preserved, this makes capital owners incredibly rich since they can rely on extremely cheap labor while owning natural resources or the physical means of production.

      • daedalus2u

         TGGP, what I meant about physical conflict not being too expensive is that desperate people will do desperate things.  If you have a 99.999% chance of dying of starvation, a 10% chance of success in physical conflict looks great in comparison.  

        I looked at the wikipedia page on Greg Clark’s and he gets the biology wrong.  The idea that the the children of the wealthy had some inherent genetic superiority is not correct.  Economic deprivation all by itself is sufficient to stunt the abilities of the poor through nutrition, lack of education, and epigenetic programming.  

        We are already at subsistence wages.  Minimum wage is not enough to live on.  People on minimum wage need food stamps and other governmental support to survive.  

        http://www.politicususa.com/minimum-wage-study-torpedoes-wing-talking-point.html

        This change (over what was happening 40 years ago) did not occur due to population increase.  It occurred due to a transfer of income from the middle class to the most wealthy.  

        The whole point of the food stamp program in the US was to provide adequate nutrition to the poor so that they would be suitable as draftees in the event of war.  War now depends more on the military industrial complex and less on soldiers, so food stamps are a luxury (as far as the military industrial complex is concerned).  

        If the poor are deprived of the resources they need to grow into competent adults, then there will be a permanent underclass of poor people.  The resources that are needed to become a competent adult include prenatal nutrition, early childhood nutrition, kindergarten, grade school, high school and college.  Thwart acquisition of these things and you prevent children from becoming competent adults.

      • V V

         @bc10f5a14cefc0eb3a4eccbd579f918d:disqus

        This is not about you. Ask yourself: Is there anybody who would be
        motivated to do so? If there are at least small percentages of people
        who would, they will quickly grow to encompass most of the population.

        No. They would quickly reach near-subsistence level and then replicate much more slowly.

      • V V

         @bc10f5a14cefc0eb3a4eccbd579f918d:disqus and @TGGP:disqus

        The Darwinian explanation is that as long as there is some variation
        (say, in resistance to fitness-reducing birth-control) the more fit will
        replace the less fit.

        Beware of naive “Darwinian” explanations.

        That claim entails that r-selected reproductive strategies (“quantity over quality”) always dominate K-selected reproductive strategies (“quality over quantity”).
        This is empirically false: both r-selected and K-selected stragies exist in nature. Humans, in particular, are an extremely K-selected species.

        http://en.wikipedia.org/wiki/R/K_selection_theory

      • http://entitledtoanopinion.wordpress.com TGGP

        Humans do emphasize “quality” over “quantity” relative to insects who have huge broods with most of their children not surviving. But post demographic-transition humans are not following a strategy that will be selected when they have fewer grandchildren than their grandparents did. Both r selection and K selection will eventually result in Malthusianism.

        Stephen Diamond, a scenario where some authority limits the choices of all entities (as in your eugenics example) is referred to as a “singleton”. Eliezer’s artificial intelligence could constitute a singleton, as it quickly grows more powerful than anything else in existence, which is why he fears it tiling the universe with smiley-faces. Robin thinks coordination is too hard for an effective world government to emerge and control fertility as you describe.

      • V V

         @TGGP:disqus

        Both r selection and K selection will eventually result in Malthusianism.

        That’s not obvious to me. If there is a small fraction of the population that uses a reproductive strategy that is suboptimal w.r.t. the average, they will not become dominant.

        Stephen Diamond, a scenario where some authority limits the choices of
        all entities (as in your eugenics example) is referred to as a
        “singleton”.

        Actually it is referred to as “government”.

        Robin thinks coordination is too hard for an effective world government to emerge and control fertility as you describe.

        The People’s Republic of China, which population-wise is the closest thing to a world government we have, seems quite successful at mandating fertility limitations.

        Anyway, I don’t think that in a society of wealthy, educated people with access to reliable birth control, mandated fertility limitations are neccessary to keep the population size in check.

      • http://entitledtoanopinion.wordpress.com TGGP

        V V, a Malthusian situation is one in which you can’t reproduce more (by which I mean having viable surviving offspring) because of resource limitations. K vs r selection is a matter of taking a gamble with lots of low-probability offspring or concentrating your eggs in fewer baskets (which individually have a higher chance of survival). If we trace out a population history of some K or r selected species, the difference is that a lot more of the latter never get to have offspring, although their siblings may. For both population keeps growing until there aren’t enough resources for further growth. And human population growth can be explosive, even if we are near the extreme end of small broods. Cochran, Harpending, Hawks & etc paper on the speedup in evolution since the invention of agriculture led to population growth is one example. Within colonial America there were some unbelievable fertility rates among the Yankee farmers who settled Greater New England.

        Governments aren’t omnipotent even within their jurisdictions. Nor is there a world-government yet. The term “singleton” refers to something much more potent than anything currently existing. A world government could potentially be a singleton, scenarios of that sort often involve it having some kind of powerful A.I.

        It is currently the case that fertility has dropped a lot in much of the first world (this is a relatively recent event evolution has not yet had time to overturn). But all that’s necessary for the Malthusian outcome is for some subset to be different. There are people in the U.S who are simply indifferent, don’t bother using birth control, and have repeated unplanned births. There are people who explicitly try to have lots of kids as part of a larger ideology, as with the “quiverfull” or even Bryan Caplan (to a lesser extent). There are subcultures that close themselves from most of the fitness-reducing effects of modernity, like the Amish or ultra-orthodox Jews. This should not at all be a Darwinian puzzle.

      • David

        TGGP:

        “Governments aren’t omnipotent even within their jurisdictions. Nor is there a world-government yet.”

        This is not necessary for reducing reproduction to above-subsistence. All you need is a government that is functional enough to steer fertility within its own borders, defend those borders against immigration and invasion, and allocate its resources sustainably and socially enough to maintain a high per-capita wealth.

        “There are people in the U.S who are simply indifferent, don’t bother using birth control, and have repeated unplanned births. … There are subcultures that close themselves from most of the fitness-reducing effects of modernity”

        The question is how stable the heritability of these traits is in the long run. People being indifferent or negatively inclined to use birth control may see their children do the exact opposite as culture and incentives change, which can be very rapid. Isolationist pro-fertility societies are not necessarily scalable. And they aren’t competitive enough against outside pressure. Think what religion does to science. Or what an all-Amish USA could do to stave off drone-using invaders. It’s a joke.

        I think the game-changer is person-copying technology. Then you copy the memes directly into the offspring, which will not undergo a childhood stage before it becomes functional and (re-)productive.

      • http://entitledtoanopinion.wordpress.com TGGP

        David, we can imagine an isolated society which is self-sufficient with respect to resources, perhaps in a biodome or something. That might possibly persist apart from the rest of the world, while most of humanity returned to Malthusian existence. But if they required some outside resource (oil, or maybe even sunshine if the planet got covered) it would be hard to compete with all those outsiders.

        Dawkins came up with the term “meme” in an analogy to genes. It’s rather obvious that a gene encouraging reproduction will spread while genes for infertility will go extinct. As long as children’s memes are positively correlated with their parents, we should similarly expect a spread of pro-natalist memes over the long term. If there are prevalent less-natalist memes, we could even expect the spread of genes resulting in resistance to such memes (the movie Idiocracy humorously suggests that genes for lower IQ will be selected).

      • V V

         Actually it’s far from obvious that trying to reproduce as much as possible actually yields increased evolutionary fitness, particularly at group level.

        Consider the Anabaptist (Amish, Mennonite and Hutterite) groups in North America. They marry young, follow traditional “farmer” values, and until recently, they didn’t use any birth control.Despite their apparent rejection of modern technology, they do enjoy some benefits of the technological society they are embedded into, in terms of increased food production and reduced infant mortality. In principle this would allow for fertility rates in excess of 12 children per woman, and most of these children would reach adulthood.Indeed, in the early decades of the 20th they had this spectacular evolutionary fitness, but their fertility rates had fallen since then they are now comparable to those of low-class Americans and Canadians.We definitely don’t expect these group to ever become anything but a negligible fraction of the general population.

      • http://entitledtoanopinion.wordpress.com TGGP

        I was under the impression that they still had very high fertility rates and have been expanding their territory because there wasn’t enough farmland in Lancaster. I know that there are still very high fertility rates in Kiryas Joel (Satmar Hasidic rather than Amish), which officially has the highest poverty rate in America. Could you provide a link on those changes in Amish fertility?

      • V V

        I was referring to the Wikipedia article:

         http://en.wikipedia.org/wiki/Amish#Population_and_distribution

        but I was mislead by the table, because it reports the increment every eight years except the last one which refers to a two years interval.

        So, according to that table, their yearly growth rate for the 2008-2010 was 6.15%, while for the previous eight year period it was 3.64%.
        However, the article mentions that the figures are from two different sources, and that there are methodological difficulties in the estimation of the Amish population size.

        This article: http://phys.org/news/2012-07-amish-founded-weeks.html
         cites a 2012 census that reports an estimate of about 251,000 Amish, which compared to the 2008 estimate yields a yearly growth rate of 3.23%, which is consistent with the 21 – 22 years doubling time reported in the same article.

        By contrast, in the 1984-1992 period, they had a growth rate of 5.09%, corresponding to a doubling time of about 14 years.

        Thus, their fertility did indeed drop, though not as dramatically as I thought before.

        Their growth model of splintering off new communities using more and more land is clearly unsustainable. They might keep growing up to  Malthusian levels, but we don’t expect them to push to Malthusian levels the surrounding population. At some point, they will be unable to buy additional land, and military conquest is certainly not an option.

    • Robin Hanson

      Saving up during 40 years of working to pay for 15 years of lower productivity retirement isn’t remotely like saving up to pay for a *forever* of lower productivity retirement.

      • V V

        I wasn’t talking about 40 years of work.

  • mjgeddes

    another analogy here could be with stem cells in the human body.

    stem cells are sufficiently general and simple that they can be transformed into any more complex and specialized cell type.  stem cells are ageless, whereas cells that have specialized breakdown eventually.
    the software analogy is the need for ontological primitives capable of serving as prototypes for *any* new concepts that could  *ever* be invented.
    remember in object oriented programming , classes are models of concepts. so without `ontological stem cells` or `super classes` representing the most basic and general concepts logically possible, effective cross domain reasoning will become impossible due to increasing inefficiencies in refactoring classes as the complexity of the software increases.
    unless the ultimate base concepts needed to model reality are hard coded into the mind design from the start, the software system cant properly reorganize old concepts as new ones are added and ontological crises will eventually choke the mind.
    ontological entropy can age minds as surely as physical entropy can fog brains .

    • Robin Hanson

      Good point about the stem cells.

  • Douglas Knight

     For example, we are great at learning languages when young, and terrible when old.

    While we are better at learning to distinguish and produce sounds when we are children, in all other aspects of language learning that have been measured, adults learn faster than children. a survey

    • Robin Hanson

      That paper has three main points, the third of which is “Acquirers who begin natural exposure to second languages during childhood generally achieve higher second language proficiency than those beginning as adults.” So overall kids learn language faster.

      • Douglas Knight

        Learning more is not learning faster. 

        Adults give up on learning sooner, probably because they are satisfied with communicating with other expats. It is believed that isolated expats make more progress, but it is hard to measure.

  • Michael Wengler

    The amount we will need to know in order to make ems is pretty large.  We are already able to emulate 1000-neuron networks (not the traditional but not very biological “neural nets”) and we know how to turn up and turn down their learning, how to stop their learning.  If we are able to build a reliably functioning em, we will know how to emulate the 2 year old brain, the 10 year old brain and the 50 year old brain.  We will know in enough detail, and likely have enough modularity, that an em will probably be able to choose which memories to keep accurately, and which bodies of expertise to “turn up” the innovation rate and which to keep it fixed.  Further, there will probably be enough modularity that earlier states of expertise can be saved and brought back out if/when needed in consciousness.  

    An em which has all the limitations of the em running on meat is unlikely to be cost effective.  We will have to know enough to do fine tuning, and what entity would pay extra to emulate all the limitations that come from evolved meat?

  • http://profiles.google.com/khallow Karl Hallowell

    For this, I have a couple of observations. First, if one knows in advance that redesign is likely, one can make a design more flexible and capable of redesign. For example, one could maintain a network structure of usual design states. The “stock car” of the car example above could be a base state, with a truck and race car state as well. If there was a lot of demand for shifting between the truck and race car, then one could develop direct transitions. Else one could transition back to the stock car and then to the new state.

    Second, we also need to keep in mind that a system that can change and has recently done so (that is, hasn’t settled down as evolutionary systems tend to do), is probably more likely to be changeable again than a system that hasn’t changed in a long time. The critter that evolved into a new niche may well have tricks that allow it to evolve quicker to a new niche than the old critter.

    Looking at the ems discussion about the declining value of ems “labor”, I have to say that capital is the big missing factor. Not only can having the right capital make your existence cheaper and easier, but it can be used to generate passive income on its own or to leverage your ability to make income.

    I find a good analogy here is to the gaming ideas of “grind” income versus “passive” income to be quite relevant here. Assume that you have a game with some sort of economic aspect to it. Grind income requires that you sit at the keyboard/console and actually play the game in order to get income or something of value. Passive income means that you don’t need to be playing in order to rake it in. It turns out that successfully creating passive income (well in games that have such) usually differentiates the wealthiest players from everyone else. Another path is via cooperative behavior. Leaders in a group or “guild” can earn income well over what they could earn on their own, just through the efforts of the teammates they coordinate.

    Frankly, a lot of massively multiplayer online games are very close to ems in character. You can run many characters and cooperate between them (it’s common that cooperative behavior results in better grind and passive income in such games). So it is possible just to run many characters and harvest the income yourself. That’s a bit different from ems since you have a scarce resource, your time at the keyboard. So there’s only so much time you can spend running your characters.

    Anyway, I think capital is the missing ingredient in this soup. If you can create ems without creating corresponding support capital (so that they have to work work work in order to continue to exist), then eventually the value of their “labor” is going to drop to near subsidence. Some factors can aggravate this, such as a tax on capital that transfers wealth from capital owners to poor ems. Other policies could work differently, such as requiring a gift of capital to every created em either from society or from the ems creator. It’ll be much harder to flood the system with ems, if one has to sacrifice capital everytime an em is created.

  • http://www.facebook.com/florinclapa Florin Clapa

    In a worst case scenario, immortalists might avoid competition with pro-growth ems by leaving the solar system and making sure that they’re one step ahead of any expanding wave of pro-growth em colonization.

  • Pingback: Stares at the World » A Culture of Death