Future Discounts

In few billion years our descendants may spread across billions of galaxies. Even so, if they do not drastically change the structure of space-time, then within a trillion years they will fragment into billions of isolated galaxy-sized “universes”.  Standard physics, you see, says that in a trillion or two years all the galaxies near the Milky Way will merge into one big galaxy, and other galaxies will be too distant to see in any way.  For all practical purposes, that merged galaxy will be a separate universe.

If we do nothing to change the situation, then within ten or so trillion years, all current stars will be dead (degenerate), and no more stars will form.  Over the next billion trillion years, stars will occasionally smash in a flash, or pass close enough to each other to throw one out of the galaxy; in the end 1-10% remain in a central black hole.

What if we change the situation?  Most useful resources, such as hydrogen to turn into lead, or mass not yet dropped into the central black hole, will likely be identified and claimed within a few million years.  How fast will folks use up these resources?

In principle, most everything might be burned quickly in a few million years of party-hardy gluttony, or most might be saved to use steadily over the billion trillion trillion years or more before protons decay.  How fast resources are actually used would be determined by the discount factors of the creatures who control resources.  But what would those be?

If unused resources were completely stable and if property rights in resources were completely secure, then we’d mainly have a selection effect in discount rates.  Agents who discount fast would dominate early activity, while those who discount slowly would dominate late activity.  Even if initially only a tiny fraction of agents cared about activity in a billion trillion trillion years, those agents would dominate such late activity.

Any natural rate at which resources decay would set an upper limit on discounting.  There is no point in planning to use resources long after you expect them to decay.  Similarly, insecure property rights would increase discount rates. If you expect a 1% chance that your property will be stolen every million years, you won’t expect to still have much after a billion years, so you might as well plan to use most of it before then.   The same holds if your property is never stolen, but you have to spend 1% of your resources every million years to ensure that fact.

“Switzerlands,” from which theft is naturally harder, might be the last locations of activity in each galaxy.  These might be matter sent on very long secret orbits, to return back to galaxy central after a very long time.  Similarly, resources which simply could not be physically used until a long delay might ensure some late universe activity.

The inhabitants of a galaxy-universe could have different degrees of central coordination; some might have a strong central government, while others lived in anarchy.  With a strong central government, long term activity seems strongly influenced by the discount rate of that government. If this government taxed 1% of resources every million years, and didn’t invest those resources for the long run, then there would be little point in planning to use your resources after a billion years.  No obvious selection effect ensures that galaxy governments take a long view.

Physics may set the ultimate limits on how long resources, and life, can last, but governments and property rights will determine when they are actually used. Resources, and life, will likely die long before their physical expiration dates.

GD Star Rating
Tagged as: , ,
Trackback URL:
  • Neo

    I would be happy if we survive next 20 years.


  • Mike

    Isn’t there one flaw in logic here? We have to suspend some disbelief to consider that human descendants survive long enough to spread to “billions” of galaxies, yet the idea that we can control matter at that point isn’t brought into the equation.

    With computing technology in its infancy we can already create new elements and manipulate matter to some degree. If our kind survives long enough to spread further than to just a few galaxies, wouldn’t it be safe to assume we can control elements in such a way that resource scarcity isn’t much of an issue?

    • David C

      Robin Hanson’s argument assumes the laws of thermodynamics are intractable. I don’t think anybody has ever seen any evidence that his assumption is incorrect.

  • Proper Dave

    But what if you reach the fundamental limits on efficiency? In that case you how “fast” or “slow” you use your resources are meaningless.
    What is the value of time when it is replaced by a subjective tau?

    • kentucky

      But what if you reach the fundamental limits on efficiency? In that case you how “fast” or “slow” you use your resources are meaningless.

      No it isn’t. If you use the resources quickly, you may enjoy a high standard of living, but you’ll run out sooner. If you conserve them via low population growth/low consumption, you’ll last longer before starving to death.

      • Steve

        Proper Dave’s saying that “years of life remaining” ceases to have meaning if you’re, say, an emulated mind running on the most efficient hardware possible, and you don’t care much about keeping up with the scenery in realtime. In this case, your remaining lifespan is how many “processor cycles” you have remaining; and extending those processor cycles by using your resources slowly has no subjective effect.

  • Michael Howard

    Sounds like an excellent argument for a Friendly singleton that acts common interest, and if necessary, for that Friendly singleton to spin-off local versions for each “separate universe”.

  • Robert Koslover

    Ok now, you kinda lost me somewhere around your trillion-year projections mark (or perhaps slightly sooner?). So now you’re projecting the social behavior of beings billions of years in the future. Hmmm. Sort of like how the pre-viruses of the primordial ooze successfully predicted the 20th century rise of Rock and Roll? Robin, even your brain, impressive as it is, isn’t nearly big enough to predict that far. Not only that, but it is not even clear if timescales of a trillion years even have an actual meaning, even purely theoretically, when considering the intimacy of space and time if viewed on such scales and related cosmological considerations. So I can only assume you wrote this either as a joke (and hey, that’s perfectly fine) or perhaps while on some medication with an unexpected side effect (in which case you should seek medical assistance). 🙂

    • Norman

      I’m gonna have to go with Robert on this one. Physics simply doesn’t have enough data on the universe in general–and life in particular–to forecast that far out. The error term is several orders of magnitude greater than any contribution of information can counteract.

      I realize the point of the post was that institutions will matter at any conceivable future date, so presumably they matter today. But this seems like using the story of Icarus to argue for regular airplane maintenance. Misplaced fantasy.

    • Cosmologists talk regularly about the big bang 0.014 Trillion years into the past; why should 1.0 Trillion years into the future be any less real?

      • Jess Riedel

        You’re extrapolating 100-fold in terms of the age of the universe, and 10^6-fold in terms of the age of humanity. It’s not that it’s not real, it’s that we can’t have even minimal confidence in our predictions. Yes, we hope to be able to rely on the laws of physics, but it is greatly under-appreciated in the popular consciousness how unsure physicists are about things like black holes and dark matter, much less proton decay. Further, we have zero idea what out descendant will want/do (your Darwinian arguments notwithstanding). Seeing as that the dawn of (a) life and (b) consciousness happened within the last 4 billion years, how could you possibly think it likely that nothing as world-shattering will come along in a trillion years?

  • Tim Tyler

    Re: “Even if initially only a tiny fraction of agents cared about activity in a billion trillion trillion years, those agents would dominate such late activity.”

    Often, being concerned with far future events is counter-productive – since the survival challenges organisms face are in the here and now – and far-future fantasies just divert resources away from those – and so increase the chances of your germ line being obliterated via competition.

  • Randy

    Face = Melted

  • Just a question. If we could somehow substantiate ourselves as photon-photon interactions would that mean that we would never decay? If so, how does this jive with the inevitable heat death scenario?

    • Nick Tarleton

      I think in that case a negentropy source would still be required for error correction. Also, to avoid eternal recurrence, you have to keep increasing storage capacity, which means increasing in volume; it’s not clear how a computer made of photons could do that while holding onto all of its photons. Dark energy probably ruins it in other ways.

      • Hedonic Treader

        Why avoid eternal recurrence? If you find a lossless substrate, program a rich hedonic consciousness narrative into it, then loop. It doesn’t have to be daft like “one orgasm”; if you have enough storage capacity, program epic millennia of diverse bliss, then loop those.

  • The issue is oddly framed since loss of matter or loss of any type of matter isn’t the problem. The problem is loss of energy. If anything, people would be stealing electrons.

    Maybe by then they’ll figure out how to make more and more complex systems sustain themselves on less and less energy. Maybe they’ll figure out how to harvest energy from the now inaccessible 11th dimension. Maybe they’ll figure out exactly how many angels need to fit on a pin’s head in order for their to be a miraculous explosion of light which revitalizes all stars.

  • I said their instead of there. That really bothers me.

  • ShardPhoenix

    feministx: electrons are matter. Also, “energy” is really just a way of talking about the properties of various configurations of matter.

  • mjgeddes

    Reply from trillion year old beings:

    Won’t happen; the obvious reason is that when quallitropping the ghiklenoot field, the resulting horstthrof effect is sufficient, interdimensionally speaking, to project matter and energy indefinitely into the jukilovook highway. As to terms such as ‘discounting’ and ‘property rights’ we have no idea why you think these meaningless anarchisms from a 19th century human philosophy would have any relevence, it should obvious that our social organization is based on stryoloopic mind melding and fretalayic weaving of cognitive youlippic star points.

    • libertarian

      Anarchisms? Anachronisms maybe?

    • Steve

      As long as there are resources, and more than one entity-with-preferences which is made out of resources and requires resources, I think most of Robin’s very basic assumptions hold water.

  • I agree with Michael Howard. It does seem like a good argument for a Friendly singleton.

    A possible alternative to letting those other galaxies split into their own isolated “universes” is to send out robots to convert all of their matter into photons (using the Hawking radiation of small black holes), and then send those photons back to us before the universe expands too much.

    These might be matter sent on very long secret orbits, to return back to galaxy central after a very long time.

    This is a cool idea for the scenarios without strong property rights. Presumably, you want to come back as late as possible, because the central black hole will grow bigger and colder with time, but you’re constrained by how long you can keep yourself alive while you’re out on the secret orbit.

  • RobbL

    Isn’t the real problem with your line of thinking that it is almost certainly wrong? We have the history of billions of years of the universe to examine and this idea that some organism spreads all of the place just doesn’t happen. Of course the past is not completely predictive of the future, but it is a pretty good guide after all these years. Most likely we won’t be around to spread or we will lose the energy somehow.

  • Jack (who uses this name at LW)

    It isn’t obvious to me what the incentive is to post-pone resource consumption, at least if resource consumption scales to faster computing speed. If agents prefer a subjectively long lifespan the best plan is to use all their resources immediately to run their software as fast as possible. Agents get a long, fulfilling life doing whatever in simulated environments while minimizing external threats (using their resources before anyone has a chance to steal or tax them). Maybe some would have a quixotic desire to be around in the end or aspire to escape heat death but in general “slowing down the universe” seems like a better strategy than “lasting longer”.

    • If you can gain twice of what you want in any given time by burning resources twice as fast, then you will conclude you should burn resources as fast as possible. Such agents will burn fast and bright and then be gone, leaving other agents to sift among their ashes.

  • Mike

    The sweep here is breathtaking. I wonder if it is performance art, or perhaps a modern parable tweaking our own deeply loved beliefs? Lovers of this blog enjoy and believe in the broad sweep of reason beyond the bonds of knowledge. Why SHOULD there be a limit?

  • gregorylent

    what, you think there is a universe external to yourself? how silly ..

    tis all a projection of consciousness, lasts as long as there is consciousness … and that is eternal … not infinite, eternal ..

    enjoy being

  • David Strauss

    As long as we’re speculating about a trillion years into the future and the entropic iron grip of the Second Law, I have to express one doubt based on our existence.

    To put this consideration into the context of the Second Law, “everything” (in the broadest sense) is, by definition, a closed system. Yet, there exists order in our universe, as evidenced by our own existence. Yet, the Second Law appears to forbid going from nothingness in a closed system to order.

    So, one of these must be true:
    (1) I’m unreasonably assuming that nothingness is the original state of everything.
    (2) My interpretation of the Second Law is flawed; order from nothingness is possible.
    (3) The Second Law isn’t universally true; there are secrets to entropy we have yet to unlock.

    If #3 is the case, there is reason to doubt a conclusion of inevitable, extinction-causing scarcity on a galactic scale.