A black hole collapses and emits Hawking radiation with average wavelength 1/M. The observer at I^+ sees it, so do the local observers (infalling or hovering). None of these cases should violate the equivalence principle. What is the size of the 'local' patch associated with the observer, the size beyond which he is allowed to see deviations from flat space because curvature effects become measureable?
I'm confused because on the one hand I'd have thought this size goes to infinity if the observer moves towards I^+, but on the other it seems to me it should stay at 1/M so that the (average) Hawking radiation has wavelengths the size of the locally flat patch. What am I getting wrong?
(Forget about firewalls, just thermal radiation.)