In an article on the SIAI blog, I found two papers by Max Tegmark on the theory that there is no difference between physical and mathematical existence. In short, his argument is that provided that our world can be fully explained by a mathematical model—and generations of physicists have made steady progress in that direction—then it becomes meaningless to talk about the “real world” and the model as two separate things. And if one particular mathematical model “exists”, Occam’s razor suggests that they all do—because otherwise one would have to postulate the existence of an “existence” property. (But of course, it’s possible that such a property is required for theory to match observations.) As a programmer, I find this idea immensely appealing. If a deterministic simulation can contain a conscious mind—and I have a hard time seeing how that could fail to be the case—the rest seems obvious. Because if the simulation is deterministic, the mind inside will experience the exact same thing every time the simulation is run. From the inside, it’s not possible even in principle to detect when the simulation is being run, if it’s being run multiple times, stopped in the middle and resumed a billion years later, etc. Postulating that the mind inside the simulation “exists” if and only if someone somewhere has run this exact simulation at least once seems pointless, since the “existence” property makes no observable difference to anyone involved—neither to the zero or more beings running the simulation, nor to the being inside.