Ways To Pay To Exist

I’ve argued that if future tech and law enable a supply-and-demand economy of creature creation, it will enforce a good simple principle of existence:

Creature X should exist if it wants to exist [i.e., would want to exist if it existed] and it can pay for itself. … Most new creatures would have designs near the peak of factory profitability, and own little surplus relative to their cost. Residual control rights (e.g., “are they slaves?”) would rest in the hands of whomever could squeeze the most market value from them.

Control rights deal with a central conflict in factory-creature relations. Factories must pay up front to make creatures, but the value that creatures may create appears later. So how can factories assure they get paid? Some possible answers:

  • Slavery – Factories could take direct physical control of new creatures, or sell such control to others. This is a simple and robust approach, but it can also be wasteful, by reducing creature incentives to be productive.
  • Debt – A creature could be in debt for its creation costs, to be recycled or sold into slavery unless it repays via a set schedule. Inflexible payments can induce recycling needlessly often.  Such waste can be reduced by adjusting payments to market context. Debt holders may have some controls on activities or spending.
  • Stock – Others might own shares in a creature’s income, net of debt and certain specific expenses. Stock owners might voter to exercise limited control rights. Shares adjust more flexibly to changing conditions, and leave some creature incentives to find ways to be more productive.
  • Contract – A creature might be obligated by contract to work to achieve certain non-financial factory goals.  This requires goals that cannot be as well achieved by the factory itself, and requires relevant creature effort which can be monitored by courts.
  • Gratitude – A creature might have a strong preference for repaying its creator. This preference could be built into core values, or imprinted via something like education and acculturation, and encouraged via social norms.
  • Shared Goals – A factory might know how to make creatures that roughly shared certain of its broader goals. These might the creature’s core values or values imprinted via education or social norms. This approach requires factories with broad goals that can be better achieved by such delegates.
  • Reproduce – A factory may have a strong preference to make and support creatures like itself.  If it can actually make such, this process can be self-sustaining, and select for creatures who are effective at reproducing.

In practice, all of these approaches can be mixed, and I find it hard to say with much confidence which mixtures will be used more heavily, or be more profitable.  Mainly-slavery, however, seems pretty unlikely dominant long-run approach. I also find it hard to complain much about the ethics of using whatever turn out to be the most efficient mixtures. After all, using any other process would mean not creating creatures who could pay for themselves, or creating creatures who are a net burden to others.

While today’s creation practices include elements of all these approaches, we clearly lean most heavily on reproduction, and many of us are horrified at the prospect that future folk might not act similarly. For example, some libertarians tell me it is a basic ethical fact that no person should be born with debt, stock, or physical restraints. But I fear this is merely arrogant presumption that our ways must be best.

GD Star Rating
Tagged as: ,
Trackback URL:
  • Peter St. Onge

    Aside from the probabilities (or quality, in your last 2 examples) of repayment, cost reductions would seem similarly important; xenogestation (e.g. in pigs or artificial wombs) or high-quality orphanages.

  • Peter St. Onge

    … as for the self-ownership axiom, I think arrogance is a strong word. Self-ownership, even if an arbitrary moral principle, gives a very attractive result (liberty, prosperity). Its status as a moral principle may engender sacrifice on its behalf, which implies sacrifice on behalf of those goods, along the lines of JS Mill’s “one man with a belief is equal to a hundred men with only interests.”

    If you are proposing to demote the principle, I think it’s fair to require you address the expected losses from demoting such a helpful principle. Hence the justified philosophical conservatism of libertarians.

  • http://denisbider.blogspot.com/ denis bider

    The described state of affairs already exists.

    Many children in India and Asia are born into families who do not value them as human beings, but rather as commodities to sell into lives of child labor or prostitution. The children work in harsh conditions where little merit is given to concerns about their safety or future.

    These people are created because it is marginally profitable for their parents, and go on to live their lives because it is profitable to their pimps or factory owners, until they are either not useful any more, or meet some unspecified end.

    This doesn’t seem to be a new concept at all – it’s how lots of people currently alive originate.

  • Khoth

    Why wait until we have factories capable of making creatures? You can start today! Produce lots children, and recycle the ones that don’t pay you enough.

  • Pingback:  Modeled Behavior

  • Doug

    It seems to a first approximation Modigliani-Miller holds here. If most modern day industrial workers were slaves the maximum profit could probably be established by taxing at the Laffer maximum and letting them go about their lives. This doesn’t seem to different than the debt or equity scenario.

  • Jeffrey Soreff

    If you are going to put slavery on the table,
    I’d suggest that central control of reproduction,
    to hold it down to a level that maximizes per-capita utility,
    should also be on the table.
    Better to have reproduction licensed than to do
    a Malthusian race to the bottom.

    Most new creatures would have designs near the peak of factory profitability, and own little surplus relative to their cost.

    Great – little enough surplus to bring back famines?
    Back to starving peasants?
    This is bad enough to make totalitarian alternatives look good.

    • Erisiantaoist

      Jeffrey, Robin Hanson believes a “strong singleton control” scenario is unlikely. If such a scenario were to happen, we would expect either total extinction or an outcome much better than this “earn your existence” one.

  • Jim Stone

    For example, some libertarians tell me it is a basic ethical fact that no person should be born with debt, stock, or physical restraints. But I fear this is merely arrogant presumption that our ways must be best.

    Your ethical principle is that creatures who can pay for themselves should be allowed to exist.

    How does your principle come out as less “arrogant presumption” than that of the libertarian?

  • http://hanson.gmu.edu Robin Hanson

    Peter, “self-ownership” leads to “liberty” almost by definition. But how would it lead to prosperity better than following an efficiency principle?

    denis, are you objecting to poor people having kids work, to their having kids, or to there being poor people?

    Jeffrey, why ever would we want to max per-capita utility?

    Jim, I didn’t say that all principles look arrogant; principles that say that it is best if others act like us look arrogant.

    • Jim Stone

      Jim, I didn’t say that all principles look arrogant; principles that say that it is best if others act like us look arrogant.

      Then it seems all principles are one of the following:
      A) hypocritical
      B) arrogant
      C) non-applicable to the one who espouses the principle

      • Jim Stone

        Or, let me tighten up my language a bit.

        All principles do one of the following:

        A) make the espouser look hypocritical
        B) make the espouser look arrogant
        C) fail to apply to the espouser.

    • Psychohistorian

      If there are two worlds, A and B, and any entity would rather be born into B than A, does that not make B a preferable world to A?

      Sure, there may be more entities in A. But so what? No one is harmed by not being born. If you’re not born, there’s no “you” to be harmed. It is quite literally impossible to speak meaningfully about non-existent entities. They inherently have no properties.

    • http://denisbider.blogspot.com/ denis bider

      Robin, I’m not objecting. I only stated that many people who are created today, are created through means other than “Reproduction”. “Slavery” would seem more descriptive for children who are created for the purpose of being sold into prostitution or child labor.

      This applies not only to people, but many other creatures, especially those in our food system. Those creatures outnumber people, and are created through a mechanism most like “Slavery”.

      As far as what system I’d like to see in place – I don’t have certain convictions either way. I do dislike slavery, of animals or of people.

      For animals exploited in the food industry, I would prefer fewer of them to exist, and I would prefer those who would exist to enjoy lives of much greater quality. I would be happy to accept a major increase in the price of animal products as a result, and would respect those products more.

      If you want me to choose between a scenario where extra people exist who live their entire lives as slaves, and another scenario where those people don’t exist, I would prefer such people not to exist.

      It adds value to my life if I don’t have to think about creatures who appear to be living lives of much less happiness than their potential. If I am aware of such creatures existing, the discrepancy between their real and potential happiness bothers me. Therefore I would prefer creatures to exist either near their optimal happiness levels, or to not exist at all.

    • michael vassar

      I think that most people assume that we should maximize per-capita utility, under some constraints such as not killing existing people.

  • Jeffrey Soreff

    Jeffrey, why ever would we want to max per-capita utility?

    Because it gives us a decent chance of retaining a reasonable
    standard of living even if there is some degree of dispossession
    or redistribution. This will _not_ hold if we try to maximize the number
    of people, to the point where they are almost but not quite impoverished to the point of suicide.

    If nothing else: Why would you want to create a world where life is
    so cheap that no one would have the resources or the interest to
    revive a cryonicist? In the first world, we’ve put some obligations
    on hospitals to at least stabilize incoming emergency cases.
    Do you really want a world where, unless they can prove they
    can pay, they get tossed out as cheap scrap instead?

  • Peter St. Onge


    In the Lockean version, self-ownership implies ownership over one’s actions (i.e. freedom of contract, a succinct summary of ‘liberty’).

    In turn, ownership of one’s actions implies ownership over one’s property (acquired assets combined with one’s action-cum-labor).

    Freedom of contract and freedom of property imply prosperity through unfettered exchange and uncheapened incentives.

    One could try separately to give moral weight to all 3 principles (self-ownership, free contract, property rights), but I like the efficiency of a single unifying principle (self-ownership) that pretty much everybody can agree on.

  • Peter St. Onge


    Ah, I didn’t fully answer the question. My concern with ‘efficiency’ rather than ‘self-ownership’ would be on two points: level of abstraction (i.e. rule vs act utilitarianism) and complexity of the phenomenon. So if we assume omniscience, I’m all for efficiency. With limited knowledge, I’d prefer a tried-and-tested meta-rule that seems empirically sound. It seems to me the best meta-rule is self-ownership.

  • Evan

    Robin, the self-ownership axiom, at least in its weaker forms, has a very basic utilitarian justification, namely that it helps improve economic incentives by making sure that people are held responsbile for their own actions and not the actions of others. Since people cannot directly control the actions of others, holding them responsible for something someone else did would produce bizarre incentives that would likely cause inefficiency. If a random person was held responsible for a murder, instead of the murderer, there would be social chaos.

    That being said, there are a few examples I can think of where holding someone responsible for the actions of another can at least be reasonably defended as efficient. The first example that comes to mind is schools that punish the whole class for the negative behavior of one student, in order to encourage the students to police themselves. I cannot tell without a considerable amount of thought whether the methods of paying for existing you propose are similarly defensible, but I think they might be.

    This does bring up another thought however, which is whether our tendency to believe holding another person responsible for another’s actions is immoral is a result of inbuilt, moral intuition, or a result of the norms of Western culture. I can find a lot of historical examples of people being held responsible for the actions of others, but most of them (such as the school example I gave earlier) involve people who are a member of a group being held responsible for the actions of other members of that group, by a person outside that group. The human tendency to treat outside groups as if they were individuals or hive-minds is well known, I don’t know if we have a tendency to treat insiders similarly.

    If the answer is no, humans instinctively treat insiders as only responsible for their own actions, then the factory-creatures would have very different moral intuitions from us, or they would likely instinctively revolt against their creditors.

  • roystgnr

    I’m posting a comment on your blog, which is a very valuable service to you, Robin. I estimate my cost at $2000, which previously was a bit of a deterrent to providing you with this service, but fortunately you’re open to the idea of compensation via one-sided non-consensual agreements! I’m willing to accept remuneration via a period of 160 hours of slavery this year, a dozen equal weekly debt repayments at 10% APR, a 0.1% share of stock in all your future earnings, or a contract wherein you agree to improve your posts’ ethical values to my standards.

    Please let me know which you would prefer, at your earliest convenience.

  • MichaelG

    So I’m imagining a future where the Sun is plated over with processors, running at maximum theoretical speed (physics is done with) simulating various AIs. The AIs might trade something or other, although I can’t imagine what.

    I’m a human-level AI inhabitant of this world, and don’t produce much of value. So instead of giving me a lot of cycles, I only get a few. My subjective world is nearly infinite, but in the real world, I’m running at 1% of real time. I can’t trade with any of the fast AIs, because I just don’t produce anything they want (perhaps in some “long tail” fashion, I sometimes earn a windfall.)

    But I can trade with all the other marginal, boring human-level AIs. We can have a community that’s a lot of fun. Billions of us together just use up a fraction of the available cycles, stuck over in a tiny patch of the Sun (labeled “legacy AIs, Do Not Disturb”.)

    So subjectively, my world is rich and full of interest and luxury, even though objectively, I’m a relic running slowly in a neglected area of the larger economy.

    • Jeffrey Soreff

      a) will any of the more powerful AIs even put up the
      “legacy AIs, Do Not Disturb” sign in the first place?
      In human history, the fate of aboriginal peoples with less powerful
      weapons has usually been to be hunted down or clobbered in some
      other way
      b) will that sign be honored for long, even if it is put there?
      I’m writing from the U.S., so I might as well cite
      the long list of broken treaties the U.S. made with native Americans

      • MichaelG

        I’m assuming a world where the total compute power required to simulate the human race is trivial. So we’re basically a museum piece on someone’s wall. There’s no point in destroying that for the tiny bit of wall space it would free up.

        I continue to rebel against Robin’s vision of the ultimate society as being so intensely competitive that our kind of minds can’t survive.

    • Jeffrey Soreff

      I certainly hope that society does not ultimately become as intensely competitive as Robin expects (hopes??). I agree that, with any luck, the fraction of the society’s total computing power needed to simulate the human race (in our current numbers) would be trivial. Unfortunately, even if the total resources are enormous, if the per-capita resources are miserably close to subsistence even for the efficient AIs, then our prospects look very dim. Consider that, even as we type, some of our primate cousins are winding up as bushmeat…

  • Robert Koslover

    I’m surprised that y’all don’t seem to be aware that children are supposed to repay their parents with nachas.

    (Nachas: (rhymes with “loch S”) pride, joy, pleasure, gratification felt at the accomplishments of loved ones such as children and grandchildren. “Oi! That boychik (little boy) is so smart! Three years old and already he can read! He gives me such nachas!” )

    See http://www.bubbygram.com/yiddishglossary.htm , among others.

  • mgravity

    I have a few questions. First, in this world, where the cost of creating or destroying any one creature is low, and the primary consideration is the opportunity cost of creating creature x vs creatures y or z, why would profit maximizing enterprises create any creatures with any goals not fully in line with their own, when they could presumably create zombie creatures with the same profit maximizing ability, but without any autonomous desires or wants? This should be cheaper, since any resources that would have been spent monitoring the creature, or enforcing contracts against the creature are now freed up for other purposes.

    Thinking about this, I’d expect the only proper creatures that would be created are those created for non monetary/resource purposes, e.g. love or the desire for children. I don’t see any autonomous creature being able to compete with zombie creatures when the goal is profit maximization.

    • mgravity

      Sorry, I guess that should be one question. That’ll learn me to start writing before I’ve property organized my thoughts.

  • Robert Wiblin

    Why does slavery seem unlikely?

  • http://hanson.gmu.edu Robin Hanson

    Jeffrey, allowing Pareto improvements makes for world with more total wealth, and such a world can just do more.

    Peter, I don’t see that it requires more central knowledge to allow the creation of creatures than to forbid it.

    Evan, yes there are agency costs, and so advantages to avoiding agency relations, but surely we don’t want to therefore forbid all agency relations. After all, whoever enforces this no agency rule is an agent.

    Roystgnr, you are not my creator.

    Michael, yes even a marginalized relatively-poor minority might live well in a vast rich universe.

    mgravity, I’m not assuming I know how hard or expensive it is to create creatures with goals “in line with your own.”

    Robert, the other options seem workable and avoid slavery waste.

    • Jeffrey Soreff


      allowing Pareto improvements makes for world with more total wealth, and such a world can just do more.

      True, but that is only relevant to the ruler of such a world – it
      gets to use the aggregate resources of the world,but we
      aren’t the ruler, so we get stuck with per-capita standard of living.
      If that goes down, we get screwed, or die.

      • Proper Dave

        “do more” consists of the number of subsitence malthusian existence. How does quantity = quality? Also a perfect optimal world will also be perfectly eqalitarian. So there will be no “rulers” to leverage the aggrerate “doing-power”.

        Unless you subscrive to some sort of super-organism theory where some “mind of the market” will emerge that will do wonderfull things. Literally a cog in the machine… yikes.

        The problem is one word: “deflation”. Once you reach peak efficiency your natural resource markets will freeze up. So anyone holding a resource stock would be a fool to sell because tommorow it will be worth more. Once mathusian equilibrium is reached, rents takes over.
        So no liquidity, so little “blood” for your market-mid/super organism.

      • http://hanson.gmu.edu Robin Hanson

        Economists have not found much support for the claim that a few rich folk encourage much more innovation than lots of poor folk; it is mainly the size of the economy, and somewhat its capital intensity, that determines innovation rates.

  • http://www.weidai.com Wei Dai

    Robin, is your position that if outcome A is Pareto optimal, and outcome B is not Pareto optimal, then A is ethically preferred to B, even if A is not a Pareto improvement over B?

  • roystgnr

    What were we saying before about presumption? Why should we presume that there is something uniquely special about just one type of creation that entitles the servicer to enforce an un-agreed-upon contract? I did not create your body, but I did create these thoughts in your mind, and rearranging already-existing neurons in your brain is no different in principle than rearranging already-existing atoms in your mother’s food. You’ve accumulated $4000 worth of newly created thoughts, now, in fact, which would not have existed without my benevolent intervention; surely that entitles me to be paid for them.

  • Peter St. Onge


    I was taking issue with your suggestion that self-ownership may be a groundless or arbitrary (‘arrogant’) presumption. If it is useful to morally privilege the meta-rule of self-ownership, then a system based on unilateral contracts has a sort of negative moral externality.

  • http://queersingularity.wordpress.com/ Summerspeaker

    The fact that your thesis on the importance of nonexistent entities leads to slave factories should promote a thorough reconsideration of the whole notion. Furthermore, the market will appear an even more absurd distribution mechanism with the sort of productive technology you’re speculating about.

  • Cyrus

    One creature with the capacity to efficiently produce all goods and services it desires is also a stable equilibrium.

    • http://pendorwright.com Elf Sternberg

      Cyrus: In a universe which I would recognize, such a creature would still have resource constraints. It would have to (a) do resource acquisition itself, (b) enter into an exchange with resource brokers, or (c) desire a static stability as profound as the Transmuters did at the end of Greg Egan’s “Diaspora.”

  • http://pendorwright.com Elf Sternberg

    While I don’t disagree with Robin’s thought experiment (indeed, he’s given me a lot to think about), as an SFnal writer I years ago settled on gratitude, embedded as a core personal value, almost an instinct, in the manufactured beings of my long-running space opera.

    As the series and the characters have matured, however, I’ve come to a rather horrific conclusion: what we have is people (for some definition of) who are not “emotionally repressed human beings,” but who are beings with a value system that is not some arbitrary but contingent result of evolutionary pressures, but individuals with a different moral core from our own.

    To the remaining human beings around, however, the environment in which they find themselves appeals to the inner sociopath: for the first time, the environment is filled with individuals who have an inescapable deference to others (indeed, they no more want to escape deference than Catholics do Catholicism, or any other meme to which human beings arbitrarily dedicate themselves), people who you can literally “put away when you’re done with them.”

    A universe where “gratitude as a core value” is a basic tactic ensuring return for the investment in creating a sentient being would be inhabited with a lot of casual Leonard Lakes.

    • Erisiantaoist

      As an SF writer, have you noticed that Robin’s scenario is that of the Vile Offspring in Accelerando?

  • Psychohistorian

    This rather misses the significant point that if you have so much power that you can capture nearly all of the utility an entity generates, it makes little sense to concern yourself with the utility of the entity. If I’m the factory owner, and I can take 99.9 out of every 100 units produced by my workers, and they need to retain 5/100 to make their lives “worth living,” why would I bother leaving them with that? Why should I give a damn how much they enjoy life?

    In other words, your hypothetical world has a bizarre regard for aggregate utilitarianism that there’s little reason to believe any market system would generate. So long as my workers value life to the extent of not killing themselves, I have no reason to not take everything from them I can take.

  • http://hanson.gmu.edu Robin Hanson

    Wei, no.
    roystgnr, we have many good reasons for assigning stronger physical than intellectual property rights.
    Peter, arrogance need not be groundless.
    Summer, we already accept many kinds of slavery.
    Elf, interesting; care to link to the relevant stories?
    Psycho, I’m arguing for Pareto improvements, not for utilitarianism.

  • http://queersingularity.wordpress.com/ Summerspeaker

    I’m opposed to all the broad-definition examples of slavery you list, Robin. I’d add wage slavery too. With regards to creature creation, I find your fundamental premise misguided. Existence isn’t necessarily good; many of us would rather we had never been born than to suffer. The nonexistent cannot be polled, therefore bringing someone to life is an inherently dubious act. It’s not better to have a billion slaves than a thousand free sentients. There’s no moral imperative to fill the universe with individuals.

    Now, this assumes a measure of human psychology for the beings in question. Theoretically you could produce an intelligence that invariably loved enslavement. In a vacuum this might be fine, but it’s a potential threat to those who value freedom. Producing happy slaves signals an affinity for domination. I would worry such people might confuse me for one of their thralls. Stressing the value of liberty in the abstract promotes my personal freedom.

  • Peter St. Onge

    :) fair enough

  • Kyre

    It seems to me that Gratitude, once it could be engineered, would beat the pants off all the others. With Slavery, Debt, Stock and Contract, there needs to be some legal enforcement system which will cost the factory. Shared Goals and Reproduction wouldn’t, but might be vulnerable to goals wandering off. A creature with correctly designed Gratitude would not spend any time trying to cheat / escape / default, and would work itself to keep its goals stable.

  • http://denisbider.blogspot.com/ denis bider

    Robin, you say you are concerned with Pareto improvements.

    But moving from a scenario where X creatures exist, and are happy, to a scenario where X+Y creatures exist, and the additional creatures Y are only marginally happy, is not a Pareto improvement, as long as creatures X are equipped with empathy similar to that of humans today.

    It seems evident that not all, but a large proportion of humans, are less happy than they might otherwise be, just because they know that there are other creatures who are much less happy than they are.

    As long as many humans have trouble coping with the existence of even a few creatures whose existence is only marginally worthwhile, adding such creatures is not a Pareto improvement.

  • Pingback: Overcoming Bias : Life Is Scarce