CarlShulman comments on Friendly AI and the limits of computational epistemology

[CarlShulman]

Edited to reduce snark.