do we live in a futurologists think it will happen
within the next 50 years. But even if it
takes10 million years, it makes no
computer simulation? difference to the argument.
Let me state what the conclusion of
the argument is. The conclusion is that
NICK BOSTROM at least one of the following three
propositions must be true:
1 Almost all civilisations at our level
SCIENCE has revealed much about the infrastructure. Our habitat is now without mathematics. It starts with of development become extinct
world and our position within it. largely one of human making. The the assumption that future civilisa- before becoming technologically
Generally, the findings have been fact of technological progress is also in tions will have enough computing mature.
humbling. The Earth is not the centre a sense humbling. It suggests that the power and programming skills to be 2 The fraction of technologically
of the universe. Our species most advanced technology we have able to create what I call “ancestor mature civilisations that are interested
descended from brutes. We are made today is extremely limited and simulations”. These would be detailed in creating ancestor simulations is
of the same stuff as mud. We are primitive compared with what our simulations of the simulators’ almost zero.
moved by neurophysiological signals descendants will have. predecessors – detailed enough for 3 You are almost certainly living in a
and subject to a variety of biological, If we extrapolate these expected the simulated minds to be conscious computer simulation.
psychological and sociological technological advances, and think and have the same kinds of
influences over which we have limited through some of their logical experiences we have. Think of an How do we reach this conclusion?
control and little understanding. implications, we arrive at another ancestor simulation as a very realistic Suppose first that the first proposition
One of our remaining sources of humbling conclusion: the “simulation virtual reality environment, but one is false. Then a significant fraction of
pride is technological progress. Like argument”, which has caused some where the brains inhabiting the world civilisations at our level of develop-
the polyps that over time create coral stir since I published it three years ago. are themselves part of the simulation. ment eventually become technologi-
reefs, the many generations of The formal version of the argument The simulation argument makes no cally mature. Suppose, too, that the
humans that have come before us requires some probability theory, but assumption about how long it will second proposition is false. Then a
have built up a vast technological the underlying idea can be grasped take to develop this capacity. Some significant fraction of these civilisa-
| NewScientist | 00 Month 2006 www.newscientist.com
tions run ancestor simulations. It should be emphasised that the If this were true, it would be an predicting what will happen next are
Therefore, if both one and two are simulation argument does not show interesting constraint on the future still the familiar ones – extrapolation
false, there will be simulated minds that you are living in a simulation. The evolution of intelligent life. of past trends, scientific modelling and
like ours. conclusion is simply that at least one The third possibility is philosophi- common sense. To a first approxima-
If we work out the numbers, we of the three propositions is true. It cally the most intriguing. If it is tion, if you thought you were in a
find that there would be vastly many does not tell us which one. correct, you are almost certainly living simulation, you should get on with
more simulated minds than non- In reality, we don’t have much in a computer simulation that was your life in much the same way as if
simulated minds. We assume that specific information to tell us which of created by some advanced civilisation. you were convinced that you were
technologically mature civilisations the three propositions might be true. What Copernicus and Darwin and leading a non-simulated life at the
would have access to enormous In this situation, it might be latter-day scientists have been “bottom” level of reality.
amounts of computing power. reasonable to distribute our credence discovering are the laws and workings If we are in a simulation, could ever
So enormous, in fact, that by roughly evenly between them. of the simulated reality. These laws know for certain? If the simulators
devoting even a tiny fraction to Let us consider the options in a little might or might not be identical to don’t want us to find out, we probably
ancestor simulations, they would be more detail. Proposition one is those operating at the more never will. But if they choose to reveal
able to implement billions of straightforward. For example, maybe fundamental level of reality where the themselves, they could certainly do
simulations, each containing as many there is some technology that every computer that is running our so. Another event that would let us
people as have ever existed. In other advanced civilisation eventually simulation exists (which, of course, conclude with a high degree of
words, almost all minds like yours develops and which then destroys may itself be a simulation). In a way, confidence that we are in a simulation
would be simulated. Therefore, by a them. Let us hope this is not the case. our place in the world would be even is if we ever reach a point when we
very weak principle of indifference, Proposition two requires that there is humbler than we thought. are about to switch on our own
you would have to assume that you a strong convergence among all What kind of implications would ancestor simulations. That would be
are probably one of these simulated advanced civilisations, such that this have? How should it change the very strong evidence against the first
minds rather than one of the ones almost none of them are interested in way you live your life? two propositions, leaving us only with
that are not simulated. running ancestor simulations. One can Your first reaction might think that the third.
Hence, if you think that propositions imagine various reasons that may lead if three is true, then all bets are off
one and two are both false, you civilisations to make this choice. Yet and you would go crazy. To reason Nick Bostrom is the director of the
should accept the third. It is not for proposition two to be true, virtually thus would be an error. Even if we are Future of Humanity Institute at the
coherent to reject all three. all civilisations would have to refrain. in a simulation, the best methods of University of Oxford
www.newscientist.com 00 Month 2006 | NewScientist |