June 1, 2003

June 2003

Book Review - Thomas Georges's Digital Soul: Intelligent Machines and Human Values, 2003
Artificial Intelligence for dummies.

The False Promise of Pseudoscience
Real science offers hope. Mysticism and belief in the paranormal are just plain dangerous.

Read this Dharma inspired interpretation of The Matrix: Reloaded.

In light of Matrix: Reloaded, I've been thinking about the simulation argument again. I'm wondering if we can apply Occam's Razor to a number of overly complicated theories that seek to reconcile the Fermi Paradox, and in their place conclude that, yes, the absence of ET life is further evidence to the idea that we may in fact reside in a simulation.

This can lead to some curious and provocative questions. For example, should we assume that we're a simulation that is being run by our posthuman descendants? If so, why are they running a simulation? Do they also experience the Fermi Paradox, and if so, do they also believe that they are living in a simulation?

Perhaps these posthumans find themselves alone in the universe and are running a large number of simulations of their history to assess the likelihood (or unlikelihood) of taking the step from biological to postbiological life. By conducting this experiment, they may be trying to a) gauge the probability of postbiological evolution for intelligent life, and b) to determine if they themselves are in a simulation.

Thus, there are far more simulations than realities. And if this is indeed the case, then we can conclude – from a statistical/probabilistic point of view – that we are most likely in a simulation.

Along these lines, here's a perspective from my friend, William Korvine:

Are We Living in a Simulation?
World simulations are all the rage, from the hit movie The Matrix, to the popular computer game The Sims. The question is, are we, ourselves, in a simulation?

One opinion is that we probably are. The evidence to support this is statistical. The assumption is that a highly advanced civilization would develop computers with sufficient power to fully simulate a world, right down to the quantum effects. Once this capacity was achieved, any number of simulations could be run which would be indistinguishable from reality. Since there is only one reality, and possibly millions of simulations, the odds are that any given universe, such as our own, is a simulation.

First, let me define some terms that I will use. I divide simulations that include life into two types, “perfect” and “imperfect”. A perfect simulation is one in which the inhabitants cannot prove that they are in a simulation. An imperfect simulation, obviously, is one in which the inhabitants can know that they are simulated.

Another way of dividing simulations is into “deterministic” and “free”. In a deterministic simulation, the inhabitants are simply programs that provide pre-determined responses to simulated situations. A free simulation permits the inhabitants an unrestricted set of thoughts and reactions.

From these definitions, it is clear that any deterministic simulation must be perfect. The Sims will never discover that they are in a simulation because it is not within their program to even ask the question. By corollary, an imperfect simulation must be free. The Matrix is imperfect, since some inhabitants know they are simulated. Therefore, those inhabitants must be free in their thinking.

If we exist in a simulation, it is perfect (so far). There is no physical evidence that we are in a simulation, only statistical and philosophical theory. Further, we are definitely in a free simulation (if we are in one), because we can ask the question of whether we are in one or not.

Why would any society simulate the world we live in? It seems a lot of trouble. Since we are not experiencing frequent mischievous “acts of God”, I doubt we are in an entertainment simulation like The Sims. Thus, we are probably in a program to determine the end result given a set of initial conditions, and possibly minute adjustments. Why would someone run such a program? More importantly, how many would they run? This is important because an estimate of the number of simulated universes gives us the likelihood of our universe being simulated.

To answer this question, we must go back to the fundamental question: how can we prove we are in a simulation? By definition, if the simulation is perfect, we cannot. But if we had the computing power, we could produce a statistical proof. How? We create simulated universes, give them the initial conditions and physical laws of our universe, allow them the randomness required by those physical laws (the Hiesenberg Uncertainty Principal), and let them run. We run millions of them. However many of them evolve into a universe that looks like ours, that’s the likelihood that our universe is real. If very few, or none, develop a human-like society, we can conclude beyond doubt that we are in a simulation that has been “jigged” to simulate us.

So, are we in a simulation? Almost certainly. We are one of millions of free, perfect universes that an advanced civilization is running to determine the answer to the very question we are asking. Are they in a simulation.

Of course, only perfect simulations would count. Any simulation in which the inhabitants determine that they are in a simulation would be eliminated from the statistics they are gathering.Thus we can expect our simulation to be terminated very shortly. Good-bye.

-- William Korvine