[ExI] Intelligence inside a black hole

[Anders Sandberg]

On 17/06/2012 01:37, Anders Sandberg wrote:
> The real thing to try if you have a Kerr hole like this is to use 
> closed timelike curves to send information to the past, and hence use 
> acausal computation. That will likely beat most forms of computronium, 
> since you can do hypercomputation.

Incidentally, I came across Scott Aaronson's nice paper
http://www.scottaaronson.com/papers/ctc.pdf
http://www.scottaaronson.com/talks/ctc-dc.ppt
today. This is the one where he proves that if you can get CTCs 
classical and quantum computation become equally powerful (*very* 
powerful - PSPACE, which includes NP: these systems eat the travelling 
salesman problem for breakfast, and find roots to any equations in O(1) 
time).

Strictly speaking this is not hypercomputation since the PSPACE-computer 
is just solving intractable problems quickly: it is not computing 
uncomputable functions. However, http://arxiv.org/pdf/1105.0047.pdf show 
that CTCs can do hyperturing computation. They also mention that you can 
use a Kerr black hole to do hypercomputation of a kind, by jumping into 
it and having a computer on the outside calculating forever, sending the 
answer to you (since you get infinite blueshift as you approach the ring 
singularity it will reach you in time). But they argue that wormholes 
work better.

Whether this is a good argument against CTCs and wormholes, or whether 
this is an argument that we live in a universe where very poweful 
computation is possible, nobody knows.

-- 
Anders Sandberg,
Future of Humanity Institute
Philosophy Faculty of Oxford University