Background independence:

In many cases, we write classical part of a field as background and consider small fluctuations around it. String theory as formulated in the mid 80s involves this sort of separation. But with no classical background, it’s hard to interpret what the theory means! What does GR describe when g_{\mu\nu} = 0?

AdS/CFT sheds light on this– in AdS space, we need boundary data in order to get a well-posed initial value problem (for massless particles). Harvey says that a formalism that is independent of boundary data may or may not exist, but it’s fine if we only have a formalism that depends on boundary data, because finite energy excitations will never affect the asymptotic boundary.