Instructions: To add a question/comment to a specific line, equation, table or graph simply click on it.
Click on the annotations on the left side of the paper to read and reply to the questions and comments.
In this paper, Bostrom describes his **"Simulation Argument"**. Thi...
Nick Bostrom is a Swedish philosopher at the University of Oxford. ...
Here is an interview with with Nick Bostrom at the Future of Humani...
This subject gained attention recently due to an interview with Elo...
There are mentions to simulated worlds in pop culture that date as ...
*Thinking is a function of man's immortal soul [they say.] God has ...
**Substrate-Independence:** Substrate-Independence is a term used t...
There is a lot of debate about whether we will be able to simulate ...
Ray Kurzweil predicts that by 2029 a personal computer will be 1,00...
The simulation argument is time independent. It does not matter tha...
The author invokes new physical phenomena yet to be discovered and ...
There are $10^{11}$ neurons in a human brain operate by sending el...
In the late 1980s Thomas K. Landauer performed a study in order to ...
We can easily compute this number of oprations. Let's assume th...
As you go deeper into the simulation and you consider multi-level s...
The author develops a probabilistic consideration of the argument o...
**(1)** $f_p \approx 0 \Rightarrow$ *"The fraction of human-level...
**The Brain in a Vat** thought-experiment is most commonly used to ...
We could go extinct trying to reach posthuman stage. Here is a vide...
Running ancestor simulations is either banned by future civilizatio...
Humans experience the world through their senses. What we call "rea...
If we believe that we are in a simulation and that our simulators a...
This is an interesting argument against the multi-level hypothesis....
The author makes a good analogy with religion. **The creators of an...
**The simulation hypothesis cannot be proven or disproven**. It als...
1
AREYOULIVINGINACOMPUTERSIMULATION?
BYNICKBOSTROM
[PublishedinPhilosophicalQuarterly(2003)Vol.53,No.21 1,pp.243255.(Firstversion:2001)]
Thispaperarguesthatatleastoneofthefollowingpropositionsistrue:(1)
the human species is very likely to go extinct before reaching a
“posthuman” stage; (2) any posthuman civilization is extremely unlikely
torunasignificantnumberofsimulationsoftheirevolutionaryhistory(or
variations thereof); (3)
we are almost certainly living in a computer
simulation.It follows thatthebeliefthatthereisa significant chance that
we will one day become posthumans who run ancestorsimulations is
false, unless we are currently living in a simulation. A number of other
consequencesofthisresultare
alsodiscussed.
I.INTRODUCTION
Manyworksofsciencefictionaswellassomeforecastsbyserioustechnologists
and futurologists predict that enormous amounts of computing power will be
available in the future. Let us suppose for a moment that these predictions are
correct. One thing that later generations might
do with their superpowerful
computers is run detailed simulations of their forebears or of people like their
forebears.Becausetheircomputerswouldbesopowerful,theycouldrunagreat
many such simulations. Suppose that these simulated people are conscious (as
they would be if the simulations were sufficiently fine
grained and if a certain
quite widely accepted position in the philosophy of mind is correct). Then it
could be the casethatthevast majority of minds like ours do not belong tothe
original race but rather to people simulated bythe advanced descendants of an
originalrace.
Itisthenpossibletoarguethat,ifthiswerethecase,wewouldbe
rational to think that we are likely among the simulated minds rather than
among the original biological ones. Therefore, if we don’t think that we are
currentlylivingina computersimulation,wearenotentitled
tobelievethatwe
will have descendants who will run lots of such simulations of their forebears.
Thatisthebasicidea.Therestofthispaperwillspellitoutmorecarefully.
2
Apartformtheinterestthisthesismayholdforthosewhoareengagedin
futuristic speculation, there are also more purely theoretical rewards. The
argument provides a stimulus for formulating some methodological and
metaphysical questions, and it suggests naturalistic analogies to certain
traditional religious conceptions, which some may find
amusing or thought
provoking.
thatweneedtoimportfromthephilosophyofmindinordertogettheargument
started. Second, we consider some empirical reasons for thinking that running
vastly many simulations of human minds
would be within the capability of a
future civilization that has developed many of those technologies that can
already be shown to be compatible with known physical laws and engineering
constraints.Thispartisnotphilosophicallynecessarybutitprovidesanincentive
for paying attention to the rest. Then follows the
core of the argument, which
makes use of some simple probability theory, and a section providing support
foraweak indifference principle thattheargument employs.Lastly, wediscuss
someinterpretationsofthedisjunction,mentionedintheabstract,thatformsthe
conclusionofthesimulationargument.
II.THEASSUMPTIONOF
SUBSTRATEINDEPENDENCE
A common assumption in the philosophy of mind is that of substrate
of physical substrates. Provided a system implements the right sort of
computational structures and processes, it can be associated with
conscious
experiences. It is not an essential property of consciousness that it is
implemented on carbonbased biological neural networks inside a cranium:
siliconbasedprocessorsinsideacomputercouldinprincipledothetrickaswell.
Argumentsfor thisthesishavebeengiven intheliterature,andalthough
itis
notentirelyuncontroversial,weshallheretakeitasagiven.
The argument we shall present does not, however, depend on any very
strong version of functionalism or computationalism. For example,weneednot
assume that the thesis of substrateindependence is necessarily true (either
analyticallyormetaphysically)just
that,infact,acomputerrunningasuitable
program would be conscious. Moreover, we need not assume that in order to
createamindon a computer it would be sufficient toprogram it in such away
that it behaves like a human in all situations, including passing the Turing
test
etc.Weneedonlytheweakerassumptionthatitwouldsufficeforthegeneration
ofsubjectiveexperiences that the computational processesof ahumanbrainare
structurally replicated in suitably finegrained detail, such as on the level of
3
individual synapses. This attenuated version of substrateindependenceis quite
widelyaccepted.
Neurotransmitters, nerve growth factors, and other chemicals that are
smallerthanasynapseclearly playa roleinhumancognitionandlearning.The
substrateindependencethesis isnotthattheeffectsofthese chemicalsaresmall
orirrelevant,
butratherthattheyaffectsubjectiveexperienceonlyviatheirdirect
orindirectinfluenceoncomputationalactivities.Forexample,iftherecanbeno
difference in subjective experience without there also being a difference in
synapticdischarges,thentherequisitedetailofsimulationisatthesynapticlevel
(orhigher).
III.THETECHNOLOGICALLIMITSOFCOMPUTATION
At our current stage of technologicaldevelopment, we have neither sufficiently
powerful hardware nor the requisite software to create conscious minds in
computers. But persuasive arguments have been given to the effect that if
technological progress continues unabated then these shortcomings will
eventually be
overcome. Some authors argue that this stage may be only a few
1
Yet present purposes require no assumptions about the time
scale. The simulation argument works equally well for those who think that it
will take hundreds of thousands of years to reach a “posthuman” stage of
civilization, where humankind has acquired most of the technological
capabilities that one cancurrently show
tobeconsistentwithphysicallawsand
withmaterialandenergyconstraints.
Suchamaturestageoftechnologicaldevelopmentwillmakeitpossibleto
convert planets and other astronomical resources into enormously powerful
computers. It is currently hard to be confident in any upper bound on the
computing power that may
be available to posthuman civilizations. As we are
still lacking a “theory of everything”, we cannot rule out the possibility that
novel physical phenomena,notallowedforincurrentphysicaltheories,maybe
utilizedtotranscendthoseconstraints
2
thatinourcurrentunderstandingimpose
1
See e.g. K. E. Drexler, Engines of Creation: The Coming Era of Nanotechnology, London, Forth
Estate, 1985; N. Bostrom, “How Long Before Superintelligence?” International Journal of Futures
Studies, vol. 2, (1998); R. Kurzweil, The Age of Spiritual Machines: When computers exceed human
intelligence,NewYork,VikingPress,1999;H.
Moravec,Robot:MereMachinetoTranscendentMind,
OxfordUniversityPress,1999.
2
Such as the BremermannBekenstein bound and the black hole limit (H. J. Bremermann,
“Minimumenergyrequirements of information transfer and computing.” InternationalJournalof
TheoreticalPhysics 21: 203217(1982);J. D. Bekenstein, “Entropy contentand informationflowin
systemswithlimitedenergy.”PhysicalReviewD30:16691679
(1984);A.Sandberg,“ThePhysics
4
theoretical limits on the information processing attainable in a given lump of
matter. We can with much greater confidence establish lower bounds on
posthuman computation, by assuming only mechani sms that are already
of a sugar cube
(excluding cooling and power supply) that would perform 10
21
instructionspersecond.
3
Anotherauthorgivesaroughestimateof10
42
operations
per second for a computerwitha mass on order of alarge planet.
4
(If wecould
createquantumcompu ters,orlearntobuildcomputersoutofnuclearmatteror
plasma, we could push closer to the theoretical limits. Seth Lloyd calculates an
upperboundfora1kg computerof5*10
50
logi cal operations persecondcarried
out on ~10
31
bits.
5
However, it suffices for our purposes to use the more
conservativeestimatethatpresupposesonlycurrentlyknowndesignprinciples.)
The amount of computing power needed to emulate a human mind can
likewise be roughly estimated. One estimate, based on how computationally
expensive it is to replicate the functionality of a
piece of nervous tissuethatwe
have already understood and whose functionality has been replicated in silico,
contrastenhancementintheretina,yieldsafigureof~10
14
operationspersecond
for the entire human brain.
6
An alternative estimate, based the number of
synapses in the brain and their firing frequency, gives a figure of ~10
16
10
17
operationspersecond.
7
Conceivably,evenmorecouldberequiredifwewantto
simulate in detail the internal workings of synapses and dendritic trees.
However,itislikelythatthehumancentralnervoussystemhasahighdegreeof
redundancy on the mircoscale to compensate forthe unreliability and noisiness
ofits neuronal
components.Onewouldtherefore expect a substantialefficiency
gainwhenusingmorereliableandversatilenonbiologicalprocessors.
Memory seems to be a no more stringent constraint than processing
power.
8
Moreover,sincethemaximumhumansensorybandwidthis~10
8
bitsper
second, simulating all sensory events incurs a negligible cost compared to
simulating the cortical activity. We can therefore use the processing power
of Information Processing Superobjects: The Daily Life among the Jupiter Brains.” Journal of
EvolutionandTechnology,vol.5(1999)).
3
K. E. Drexler, Nanosystems: Molecular Machinery, Manufacturing, and Computation, New York,
JohnWiley&Sons,Inc.,1992.
4
R. J. Bradbury, “Matrioshka Brains.” Working manuscript (2002),
5
S.Lloyd,“Ultimatephysicallimitstocomputation.”Nature406(31August):10471054(2000).
6
H.Moravec,MindChildren,HarvardUniversityPress(1989).
7
Bostrom(1998),op.cit.
8
Seereferencesinforegoingfootnotes.
5
required to simulate the central nervous system as an estimate of the total
computationalcostofsimulatingahumanmind.
If the environment is included in the simulation, this will require
of the simulation. Simulating the entire universe
down to the quantum level is
obviouslyinfeasible, unless radically new physics is discovered. But in order to
get a realistic simulation of human experience, much less is needed only
whateverisrequiredtoensurethatthesimulatedhumans,interactinginnormal
human ways with their simulated environment, don’t
notice any irregularities.
ThemicroscopicstructureoftheinsideoftheEarthcanbesafelyomitted.Distant
astronomicalobjectscanhavehighlycompressedrepresentations:verisimilitude
need extend to the narrow band of properties that we can observe from our
planetorsolarsystemspacecraft.OnthesurfaceofEarth,macroscopicobjects
in
inhabited areas may need to be continuously simulated, but microscopic
phenomena could likely be filled in ad hoc. What you see through an electron
microscope needs to look unsuspicious, but you usually have no way of
confirming its coherence with unobserved parts of the microscopic world.
Exceptions arise when
we deliberately design systems to harness unobserved
microscopicphenomenathatoperateinaccordancewithknownprinciplestoget
a computer. The simulation may therefore need to include a continuous
representationofcomputersdowntothelevel
ofindividuallogicelements.This
presents no problem, since our current computing power is negligible by
posthumanstandards.
Moreover,aposthumansimulatorwouldhaveenough computing power
to keep track of the detailed beliefstates in all human brains at all times.
Therefore, when it saw that a human was about
to make an observation of the
microscopic world, it could fill in sufficient detail in the simulation in the
appropriatedomainon an asneeded basis.Should anyerror occur, thedirector
couldeasilyeditthestates ofanybrainsthathavebecomeawareofananomaly
before it spoils
the simulation. Alternatively,the director could skip back a few
secondsandrerunthesimulationinawaythatavoidstheproblem.
It thus seems plausible that the main computational cost in creating
simulations that are indistinguishable from physicalreality for human minds in
thesimulationresidesinsimulatingorganic
brainsdowntotheneuronalorsub
neuronallevel.
9
Whileitisnotpossibletogetaveryexactestimateofthecostofa
realistic simulation of human history, we can use ~10
33
‐10
36
operations as a
9
As webuild moreand faster computers,thecostofsimulating ourmachinesmighteventually
cometodominatethecostofsimulatingnervoussystems.
6
roughestimate
10
.Aswe gain moreexperience with virtual reality,we will get a
better grasp of the computational requirements for making such worlds appear
realistic to their visitors. But in any case, even if our estimate is off by several
ordersofmagnitude,thisdoesnotmattermuchforourargument.
Wenotedthat
a rough approximation of the computational power of a planetarymass
computer is 10
42
operations per second, and that assumes only already known
nanotechnologicaldesigns,whichareprobablyfarfromoptimal.Asinglesucha
computer could simulate the entire mental history of humankind (call this an
ancestorsimulation) by using less than one millionth of its processing power for
one second. A posthuman
civilization may eventually build an astronomical
numberofsuchcomputers.Wecanconcludethatthecomputingpoweravailable
to a posthuman civilization is sufficient to run a huge number of ancestor
simulations even it allocates only a minute fraction of its resources to that
purpose.Wecandrawthisconc lusioneven
whileleavingasubstantialmarginof
errorinallourestimates.
Posthuman civilizations would have enough computing power to run
hugelymanyancestorsimulationsevenwhileusingonlyatinyfractionof
theirresourcesforthatpurpose.
IV.THECOREOFTHESIMULATIONARGUMENT
Thebasicideaof
thispapercanbeexpressedroughlyasfollows:Iftherewerea
substantialchancethat ourcivilizationwillevergettotheposthumanstageand
run many ancestorsimulations, then how come you are not living in such a
simulation?
Weshalldevelopthisideaintoarigorousargument.Let
usintroducethe
followingnotation:
P
f
: Fraction of all humanleveltechnological civilizations that survive to
reachaposthumanstage
N
: Average number of ancestorsimulations run by a posthuman
civilization
H
:Averagenumberofindividualsthathavelivedinacivilizationbefore
itreachesaposthumanstage
10
100 billion humans
50 years/human
30 million secs/year
[10
14
, 10
17
] operations in each
humanbrainpersecond
[10
33
,10
36
]operations.
7
The actual fraction of all observers with humantype experiences that live in
simulationsisthen
HHNf
HNf
f
P
P
sim
)(

Writing
I
f for the fraction of posthuman civilizations that are interested in
running ancestorsimulations (or that contain at leastsome individuals whoare
interested in that and have sufficient resources to run a significant number of
such simulations), and
I
N for the average number of ancestorsimulations run
bysuchinterestedcivilizations,wehave
II
NfN
andthus:
1)(
IIP
IIP
sim
Nff
Nff
f
 (*)
Because of the immense computing power of posthuman civilizations,
I
N is
extremelylarge,aswesawintheprevioussection.Byinspecting(*)wecanthen
seethatatleastoneofthefollowingthreepropositionsmustbetrue:
(1)
0
P
f
(2)
0
I
f
(3)
1
sim
f
V.ABLANDINDIFFERENCEPRINCIPLE
Wecantakeafurtherstepandconcludethatconditionalonthetruthof(3),one’s
credence in the hypothesis that one is in a simulation should be close to unity.
More generally, if we knew that a fraction x of all observers with human
type
experiencesliveinsimulations,andwedon’thaveanyinformationthatindicate
that our own particular expe riences are any more or less likely than other
humantype experiences to have been implemented in vivo rather than in
machina,thenourcredencethatweareinasimulationshouldequal
x:
8
xxfSIMCr
sim
)|(
(#)
This stepis sanctioned bya veryweakindifference principle.Letusdistinguish
twocases.Thefirstcase,whichistheeasiest,iswhereallthemindsinquestion
are like your own in the sense that they are exactly qualitatively identical to
yours:theyhaveexactlythesame
informationandthesameexperiencesthatyou
have.Thesecondcaseiswherethemindsare“like”eachotheronlyintheloose
senseofbeingthesortofmindsthataretypicalofhumancreatures,but theyare
ofexperiences.
I maintain that even in the latter case, where the minds are qualitatively
different, the simulation argument still works, provided that you have no
information that bears on the question of which of the various minds are
simulatedandwhichareimplementedbiologically.
strongerprinciple,whichimpliestheabovestance
forbothcasesastrivialspecialinstances,hasbeengivenintheliterature.
11
Space
doesnotpermita recapitulationofthatdefensehere,but wecanbringoutoneof
the underlying intuitions by bringing to our attentionto an analogous situation
ofamorefamiliarkind.Supposethatx%ofthepopulationhasacertaingenetic
sequence Swithinthe partof
theirDNAcommonlydesignatedas“junkDNA”.
Suppose,further,thattherearenomanifestationsofS(shortofwhatwouldturn
up in a gene assay) and that there areno knowncorrelations between having S
DNA sequenced,
it is rational to assign a credence of x% to the hypothesis that
youhaveS.Andthisissoquiteirrespectiveofthefactthatthepeoplewhohave
Shavequalitativelydifferentmindsandexperiencesfromthepeoplewhodon’t
haveS.(Theyaredifferentsimplybecause
allhumanshavedifferentexperiences
from one another, not because of any known link between S and what kind of
experiencesonehas.)
The same reasoning holds if S is not the property of having a certain
genetic sequence but instead the property of being in a simulation, assuming
only that
we have no information that enables us to predict any differences
betweentheexperiences ofsimulatedminds andthoseof the originalbiological
minds.
Itshouldbestressedthattheblandindifferenceprincipleexpressedby(#)
prescribes indifference only between hypotheses aboutwhich observer you are,
when you have no
informationabout which of these observers you are. It does
11
++
,andQuantumJoe.”Synthese
127(3): 359387 (2001); and most fully in my book Anthropic Bias: Observation Selection Effects in
ScienceandPhilosophy,Routledge,NewYork,2002.
9
notingeneralprescribeindifferencebetweenhypotheseswhenyoulackspecific
informationabout whichofthehypothesesistrue.Incontrast to Laplacean and
other more ambitious principles of indifference, it is therefore immune to
Bertrand’s paradox and similar predicaments that tend to plague indifference
principlesofunrestrictedscope.
with the Doomsday argument
12
may worry that the
bland principle of indifference invoked here is the same assumption that is
responsible for getting the Doomsday argument off the ground, and that the
counterintuitivenessofsomeoftheimplicationsofthelatterincriminatesorcasts
doubtonthevalidityoftheformer.Thisisnot
so.TheDoomsdayargumentrests
on a much stronger and more controversial premiss, namely that one should
reasonas ifone werea randomsample from the set of allpeoplewhowill ever
havelived(past,present,andfuture)eventhoughweknowthatwearelivinginthe
early
twentyfirst century rather than at some point in the distant past or the
future.Theblandindifferenceprinciple,bycontrast,a ppliesonlytocaseswhere
If betting odds provide some guidance to rational belief, it may also
be
worthto ponder that if everybody wereto placea bet on whether they are in a
simulation or not, then if people use the bland principle of indifference, and
consequentlyplacetheirmoneyonbeinginasimulationiftheyknowthatthat’s
wherealmostallpeopleare,then
almosteveryonewillwintheirbets.Iftheybet
onnotbeinginasimulation,thenalmosteveryonewilllose.Itseemsbetterthat
theblandindifferenceprinciplebeheeded.
Further, one can consider a sequence of possible situations in which an
increasing fraction of all people live in simulations:
98%, 99%, 99.9%, 99.9999%,
and so on. As one approaches the limiting case in which everybody is in a
simulation (from which one can deductively infer that one is in a simulation
oneself), it is plausible to require that the credence one assigns to being in a
simulation gradually approach the
limiting case of complete certainty in a
matchingmanner.
VI.INTERPRETATION
The possibility represented by proposition (1) is fairly straightforward. If (1) is
true, then humankind will almost certainly fail to reach a posthuman level; for
virtually no species at our level of development become posthuman, and it is
hardtoseeany justification for thinkingthat our ownspecies will be especially
privileged or protected from future disasters. Conditional on (1), therefore, we
12
Seee.g.J.Leslie,“IstheEndoftheWorldNigh?PhilosophicalQuarterly40,158:6572(1990).
10
must give a high credence to DOOM, the hypothesis that humankind will go
extinctbeforereachingaposthumanlevel:
1)0|(
P
fDOOMCr
One can imagine hypothetical situations were we have such evidence as
would trump knowledge of
P
f
. For example, if we discovered that we were
about to be hit by a giant meteor, this might suggest that we had been
exceptionally unlucky. We could then assign a credence to DOOM larger than
our expectation of the fraction of humanlevel civilizations that fail to reach
posthumanity.In
theactualcase,however,weseemtolackevidenceforthinking
thatwearespecialinthisregard,forbetterorworse.
Proposition(1)doesn’tbyitselfimplythatwearelikelytogoextinctsoon,
only that we are unlikely to reach a posthuman stage. This possibility is
compatible
with us remaining at, or somewhat above, our current level of
technologicaldevelopmentfor alongtimebeforegoingextinct.Anotherwayfor
(1)tobetrueisifitislikelythattechnologicalcivilizationwillcollapse.Primitive
humansocietiesmightthenremainonEarthindefinitely.
There are many ways
in which humanity could become extinct before
reachingposthumanity.Perhapsthemostnaturalinterpretationof(1)isthatwe
are likely to go extinct as a result of the development of some powerful but
dangerous technology.
13
One candidate is molecular nanotechnology, which in
its mature stage would enable the construction of selfreplicating nanobots
capable of feeding on dirt and organic matter a kind of mechanical bacteria.
Suchnanobots,designedformaliciousends,couldcausetheextinctionofalllife
onourplanet.
14
Thesecondalternativeinthesimulationargument’sconclusionisthat the
fraction of posthuman civilizations that are interested in running ancestor
simulationisnegligiblysmall.Inorder for (2) tobetrue,theremustbe a strong
convergence among the courses of advanced civilizations. If the number of
ancestorsimulations
createdbytheinterestedcivilizationsisextremelylarge,the
rarity of such civilizations must be correspondingly extreme. Virtually no
posthuman civilizations decide to use their resources to run large numbers of
ancestorsimulations. Furthermore, virtually all posthuman civilizations lack
13
Seemypaper“ExistentialRisks:AnalyzingHumanExtinctionScenariosandRelatedHazards.”
Journal of Evolution and Technology, vol. 9 (2001) for a survey and analysis of the present and
anticipatedfuturethreatstohumansurvival.
14
See e.g. Drexler (1985) op cit., and R. A. Freitas Jr., “Some Limits to Global Ecophagy by
Biovorous Nanoreplicators, with Public Policy Recommendations.” Zyvex preprint April (2000),
http://www.foresight.org/NanoRev/Ecophagy.html.
11
individuals who have sufficient resources and interest to run ancestor
simulations; or else they have reliably enforced laws that prevent such
individualsfromactingontheirdesires.
What force could bring about such convergence? One can speculate that
of an ethical prohibition against running ancestorsimulations because of the
sufferingthatisinflictedontheinhabitantsofthesimulation.However,fromour
present point of view, it is not clear that creating a human race is immoral. On
the contrary, we tend to view the existence of our race
as constituting a great
ethical value. Moreover, convergence on an ethical view of the immorality of
running ancestorsimulations is not enough: it must be combined with
convergence on a civilizationwide social structure that enables activities
consideredimmoraltobeeffectivelybanned.
Another possible convergence point is that almost
all individual
theylosetheirdesirestorunancestorsimulations.Thiswouldrequiresignificant
changes to the motivations driving their human predecessors, for there are
certainlymanyhumanswhowouldliketorunancestorsimulationsiftheycould
afford
todoso.Butperhapsmanyofourhumandesireswillberegardedassilly
by anyone who becomes a posthuman. Maybe the scientific value of ancestor
simulations to a posthuman civilization is negligible (which is not too
implausible given its unfathomable intellectual superiority), and maybe
posthumans regard recreational activities as
merely a very inefficient way of
getting pleasure which can be obtained much more cheaply by direct
stimulationofthebrain’srewardcenters.Oneconclusionthatfollowsfrom(2)is
that posthuman societies will be very different from human societies: they will
not contain relatively wealthy independent agents who
have the full gamut of
humanlikedesiresandarefreetoactonthem.
The possibility expressed by alternative (3) is the conceptually most
intriguing one. If we are living in a simulation, then the cosmos that we are
observing isjust atinypiece ofthetotalityof
physicalexistence. The physics in
theuniversewherethecomputerissituatedthat is running thesimulationmay
ormay not resemble thephysics of theworldthatweobserve. While the world
weseeisinsomesense“real”,itisnotlocatedatthefundamentallevelofreality.
Itmaybepossi bleforsimulatedcivilizationstobecomeposthuman.They
maythenruntheirownancestorsimulationsonpowerfulcomputerstheybuild
in their simulated universe. Such computers would be “virtual machines”, a
familiar concept in computer science.(Javascript webapplets,for instance, run
on a virtual machine
a simulated computer inside your desktop.) Virtual
machinescan bestacked: it’s possibleto simulatea machine simulating another
12
machine,andsoon,inarbitrarilymanystepsofiteration.Ifwedogoontocreate
ourownancestorsimulations,thiswouldbestrongevidenceagainst(1)and(2),
andwewouldthereforehavetoconcludethatweliveinasimulation.Moreover,
we would have to suspect that the
posthumans running our simulation are
themselvessimulated beings; and their creators, in turn, may also be simulated
beings.
Reality may thus contain many levels. Even if it is necessary for the
hierarchy to bottom out at some stage the metaphysical status of this claim is
somewhat obscure there
may be room for a large number of levels of reality,
and the number could be increasing over time. (One consideration that counts
against the multilevel hypothesis is that the computational cost for the
basementlevel simulators would be very great. Simulating eve n a single
posthuman civilization might be
prohibitively expensive. If so, then we should
expect our simulation to be terminated when we are about to become
posthuman.)
Although all the elements of such a system can be naturalistic, even
physical,itispossibletodrawsomelooseanalogieswithreligiousconceptionsof
theworld.Insomeways,
theposthumansrunningasimulationare likegodsin
relation to the people inhabiting the simulation: the posthumans created the
world we see; they are of superior intelligence; they are “omnipotent” in the
sense that they can interfere in the workings of our world even in ways that
violate its physical
laws; and they are “omniscient” in the sense that they can
monitoreverythingthathappens.However,allthedemigodsexceptthoseatthe
fundamental level of reality are subject to sanctions by the morepowerful gods
livingatlowerlevels.
Furtherruminationonthesethemescouldclimaxinanaturalistic
theogony
thatwouldstudythestructureofthishierarchy,andtheconstraintsimposedon
its inhabitants by the possibilitythat their actions on their own level may affect
can be sure that they are at the basementlevel,
then everybody would have to
consider the possibility that their actions will be rewarded or punished, based
perhaps on moral criteria, by their simulators. An afterlife would be a real
possibility. Because of this fundamental uncertainty, even the basement
civilization may have a reason to behave ethically. The fact that
it has such a
reason for moral behavior would of course add to everybody else’s reason for
behaving morally, and so on, in truly virtuous circle. One might get a kind of
universal ethical imperative, which it would be in everybody’s selfinterest to
obey,asitwere“fromnowhere”.
In
additionto a ncestorsimulations,one may alsoconsiderthepossibility
of more selective simulations that include only a small group of humans or a
13
single individual. The rest of humanity would then be zombies or “shadow
people” humans simulated only at a level sufficient for the fully simulated
people not to notice anything suspicious. It is not clear how much cheaper
obviousthat
itispossibleforanentitytobehaveindistinguishablyfromarealhumanandyet
lack conscious experience. Even if there are such selective simulations, you
should not think that you are in one of them unless you think they are much
more numerous than complete simulations. There
would have to be about 100
billion times as many “mesimulations” (simulations of the life of only a single
mind)asthereareancestorsimulationsinorderformostsimulatedpersonstobe
inmesimulations.
There is also the possibility of simulators abridging certain parts of the
mental lives
of simulated beings and giving them false memories of the sort of
one can consider the following (farfetched) solution to theproblem of evil: that
thereisnosufferingintheworldandallmemoriesofsuffering
areillusions.Of
course,thishypothesiscanbeseriouslyentertainedonlyatthosetimeswhenyou
arenotcurrentlysuffering.
Supposing we live in a simulation, what are the implications for us
humans? The foregoing remarks notwithstanding, the implications are not all
that radical. Our best guide to how our
posthuman creators have chosen to set
up our world is the standard empirical study of the universe we see. The
revisionstomostpartsofourbeliefnetworkswouldberatherslightandsubtle
inproportionto our lack of confidence inourability to understand theways of
posthumans.
Properly understood, therefore, the truth of (3) should have no
andmakingplansandpredictionsfortomorrow.Thechiefempiricalimportance
of (3) at the current time seems to lie in its role in the
tripartite conclusion
established above.
15
We may hope that (3) is true since that woulddecrease the
probability of (1), although if computational constraints make it likely that
simulators would terminate a simulation before it reaches a posthuman level,
thenoutbesthopewouldbethat(2)istrue.
If we learnmoreabout posthuman
motivations and resource constraints,
maybeasa result of developing towards becoming posthumans ourselves, then
the hypothesis that we are simulated will come to have a much richer set of
empiricalimplications.
15
For some reflections by another author on the consequences of (3), which were sparked by a
privatelycirc ulated earlierversionofthis paper, see R. Hanson, “HowtoLiveina Simulation.”
JournalofEvolutionandTechnology,vol.7(2001).
14
VII.CONCLUSION
A technologically mature “posthuman civilization would have enormous
computing power. Based on this empirical fact,thesimulation argument shows
that at least one of the following propositions is true: (1) The fractionof human
level civilizations that reach a posthuman stage is very close to zero; (2) The
fraction of posthuman civilizations that are interested in running ancestor
simulations is very close to zero; (3) The fraction of all people with our kind of
experiencesthatarelivinginasimulationisveryclosetoone.
If (1) is true, then we will almost certainly go extinct before
reaching
posthumanity.If(2)istrue,thentheremustbea strongconvergenceamongthe
courses of advanced civilizations so that virtually none contains any relatively
wealthyindividualswhodesiretorunancestorsimulationsandarefreetodoso.
If (3) is true, then we almost certainly live in a
simulation. In the dark forest of
our current ignorance, it seems sensible to apportion one’s credence roughly
evenlybetween(1),(2),and(3).
Unless we are now living in a simulation, our descendants will almost
certainlyneverrunanancestorsimulation.
Acknowledgements
I’m grateful to many people for comments, and
especially to Amara Angelica,
Robert Bradbury, Milan Cirkovic, Robin Hanson, Hal Finney, Robert A. Freitas
Jr., John Leslie, Mitch Porter, Keith DeRose, Mike Treder, Mark Walker, Eliezer
Yudkowsky,andseveralanonymousreferees.
www.nickbostrom.com www.simulationargument.com