[ExI] How to survive forever

[Anders Sandberg]

Here is a small tech report I did with a colleague
http://www.aleph.se/papers/Survival2.pdf
(it will move to a proper home on the FHI server shortly)

The result is likely not new at all, but it is pretty cute: if you 
gradually (at a logarithmic rate) make more and more backup copies you 
can keep your survival probability finite to the end of time. This works 
even for time-varying risks and some correlated risks.

I just wonder if one can come up with a more secure defence against 
systemic risks than the hierarchical one we sketch near the end. Could 
one do something clever based on the Byzantine general problem?

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