At 22:43 27/08/04 -0700, George Levy wrote:

Let's write Lob's formula as B2(B1p -> p) -> B1p
where B1, B2, and p are binary variables.
Note that B1 applies to p and B2 applies to the implication (B1p -> p). (Should I have done this differently?)

Why B1 and B2 ? Lob's formula is really B(Bp->p)->Bp, meaning
(with our naive stance toward machine) that if the machine
ever believes that: [if I (the machine) ever believe p then p] then I will ever believe p).
We will come back to it. This is true, and actually believable by any
Self-Referentially Correct machine (and even by larger classes of machines,
but for the physics extraction I interview only the SRC machine).


Reply via email to