In April 2017, Anders Sandberg, Stuart Armstrong, and Milan Cirkovic released this paper: If a civilization wants to maximize computation it appears rational to aestivate until the far future in order to exploit the low temperature environment: This can produce a 1030 multiplier of achievable computation. We hence suggest the “aestivation hypothesis”: The reason we are not observing manifestations of alien civilizations is that they are currently (mostly) inactive, patiently waiting for future cosmic eras. This paper analyses the assumptions going into the hypothesis and how physical law and observational evidence constrain the motivations of aliens compatible with the hypothesis. (
I haven’t read the paper, but what about time discounting? Ie isn’t the computing more valuable sooner rather than later? Risks from attacks etc can also be built into the discount rate.
Has anyone pointed out that a supermassive black hole's Schwarzschild surface is extremely cold, many orders of magnitude colder than the CMB? Couldn't such a black hole be used as a heat sink?
If the aliens want to maximize computation, they should be spreading across the universe and gathering energy up for use later. You can get more oomph from a tank of gas in a colder world, but the universes gas tank leaks.
Gather your hydrogen now. Use it later. Still visible.
Our disagreement with Sandberg et al. takes place within the framework of thermodynamics, where atoms are assumed to be infinitely small and energy can be exchanged in arbitrarily tiny amounts. There's a larger, more general framework of statistical mechanics where finite-size effects can be addressed, and we are just noting that that's out of our scope.
In the final section of your paper you say that among the things not addressed in your analysis is complications arising from the finite size of atoms. What complications?
I had thought you were the one deviating from the conventional definition, since it had seemed to me that you were arguing that the thermodynamically irreversible step of deleting information was actually reversible in a literal sense if done in a closed system.
It seems I had misunderstood your argument and that your actual argument is that the scarce resource is negentropy, not energy, and that deleting information uses the same amount of this scarce resource at any time.
"Reversible" has a particular technical definition in thermodynamics. There is no aestivation bonus for erasing bits at high temperature; that's the main point of our paper.
I'm not a fan of the aestivation hypothesis, for one thing since there is a lot of energy that could be harnessed now but will not plausibly be recoverable later by the same civilization (e.g. starlight).
But I don't find your arguments very convincing here.
For the "reversibility of cooling" thing at the end - it's not actually reversible, since dumping entropy into a heat reservoir having a temperature (i.e in thermal equilibrium) will necessarily raise the energy of the reservoir. You gain nothing over erasing bits at high temperature.
For the earlier argument, running your reversible computer and outputting into a closed system is "efficient"in the sense of being maximally efficient given the systems you are working with - but they are at high temperature and you are forgoing the aestivation bonus.
It can't really be reversed, because there is always some leakage. So your erasure is permanent and at non-optimally high temperature.
Even if you could really reverse it, there would be no point in not just delaying it anyway, since making it truly reversible would require not acting on the world.
Of course, the existence of thermal disequilibrium, and the fact that it will naturally tend to equilibrium, is an incentive for aestivators to first stabilize that negentropy. So it is still supports an argument against aestivation, just not the one you and your coauthors make in the paper.
I haven’t read the paper, but what about time discounting? Ie isn’t the computing more valuable sooner rather than later? Risks from attacks etc can also be built into the discount rate.
Has anyone pointed out that a supermassive black hole's Schwarzschild surface is extremely cold, many orders of magnitude colder than the CMB? Couldn't such a black hole be used as a heat sink?
*delayed rimshot for added effect*
Yes, of course.
If the aliens want to maximize computation, they should be spreading across the universe and gathering energy up for use later. You can get more oomph from a tank of gas in a colder world, but the universes gas tank leaks.
Gather your hydrogen now. Use it later. Still visible.
Thanks.
Had you waited one hour extra, you wouldn't have needed to say this.
Our disagreement with Sandberg et al. takes place within the framework of thermodynamics, where atoms are assumed to be infinitely small and energy can be exchanged in arbitrarily tiny amounts. There's a larger, more general framework of statistical mechanics where finite-size effects can be addressed, and we are just noting that that's out of our scope.
I didn't put in those words, so you'll have to ask Jess.
In the final section of your paper you say that among the things not addressed in your analysis is complications arising from the finite size of atoms. What complications?
I had thought you were the one deviating from the conventional definition, since it had seemed to me that you were arguing that the thermodynamically irreversible step of deleting information was actually reversible in a literal sense if done in a closed system.
It seems I had misunderstood your argument and that your actual argument is that the scarce resource is negentropy, not energy, and that deleting information uses the same amount of this scarce resource at any time.
"Reversible" has a particular technical definition in thermodynamics. There is no aestivation bonus for erasing bits at high temperature; that's the main point of our paper.
I'm not a fan of the aestivation hypothesis, for one thing since there is a lot of energy that could be harnessed now but will not plausibly be recoverable later by the same civilization (e.g. starlight).
But I don't find your arguments very convincing here.
For the "reversibility of cooling" thing at the end - it's not actually reversible, since dumping entropy into a heat reservoir having a temperature (i.e in thermal equilibrium) will necessarily raise the energy of the reservoir. You gain nothing over erasing bits at high temperature.
For the earlier argument, running your reversible computer and outputting into a closed system is "efficient"in the sense of being maximally efficient given the systems you are working with - but they are at high temperature and you are forgoing the aestivation bonus.
It can't really be reversed, because there is always some leakage. So your erasure is permanent and at non-optimally high temperature.
Even if you could really reverse it, there would be no point in not just delaying it anyway, since making it truly reversible would require not acting on the world.
Of course, the existence of thermal disequilibrium, and the fact that it will naturally tend to equilibrium, is an incentive for aestivators to first stabilize that negentropy. So it is still supports an argument against aestivation, just not the one you and your coauthors make in the paper.