Note to Ray Chiao
Because the light cone tips over so much I don't think a stable circular orbit inside the event horizon is possible - at least classically for a non-rotating black hole.


light cones at a black hole

 

I do agree that there will be a real electron-positron plasma stuck just outside at the event horizon where the tensor acceleration of virtual quanta off the timelike geodesic is
g(r) = (c^2rs/2r^2)(1 - rs/r)^-1/2
where r = rs + H
H << rs
1 - rs/(rs + H) = 1 - 1/(1 + H/rs) ~ 1 - 1 + H/rs = H/rs
g(r) ~ (c^2rs/2rs^2)(rs/H)^1/2 = (c^2/rs)(rs/H)^1/2
The Unruh energy is
kBT = hg(r)/c  ~ (hc/rs)(rs/H)^2
to get a real electron-positron plasma requires
2mc^2 ~ (hc/rs)(rs/H)^2 = hc/(rsH)^1/2
note geometric mean of rs with H.
2(rsH)^1/2 ~ h/mc