[simonh]

>Does someone really need to build a computer that carries out the simulation for a universe to be “real”? If there is a set of rules defining a universe, one can say that the universe already exists without having to simulate it.

Marvin Minsky made this argument and it's a compelling one. But rather than meaning all universes are simulated, it means it doesn't matter whether they are or not because for all possible universes they will exist as root universes and as simulations within more complex universes and there's no meaningful distinction between those. It's not the argument I'm making though.