[ExI] How to survive forever

[Anders Sandberg]

On 15/06/2012 07:40, BillK wrote:
> I know it is a common assumption that if you wait long enough, then 
> *anything* might happen. But 'long enough' is probably longer than the 
> life of our universe, so we can ignore the risk.

Well, the universe is open, so "long enough" looks like it is going to 
happen.

Freeman Dyson calculated the black hole implosion time for human-sized 
objects in
http://www.aleph.se/Trans/Global/Omega/dyson.txt
to 10^10^26 years, assuming quantum black holes.

Using the Gamow formula in section F for the head teleportation case, we 
can assume the mass = 1 kg, d = 0.01 m, U=0.1 J (lifting a head one cm) 
and we get the action sqrt(8MUd^2/hbar^2) = 8.48e31, and hence the 
characteristic timescale will be around 10^(3.6*10^31) [This is in the 
big range where it doesn't matter much whether we measure time in Planck 
time or years, hence T_0 also disappears].

So, we should be more worried about imploding than decapitation.


> So for the sake of the theory you might be setting your risk limit 
> higher than is necessary in practice.

It is still applicable if you are dealing with high risks, like storing 
your files. Assuming you want to have a decent chance of them all 
surviving, you should make an increasing number of separate copies.

Sure, you can make loads of them right now - spam the cloud! But that is 
expensive. So instead you make a reasonable number. Next year you add an 
extra backup. You wait two years, then add another. Wait four years, add 
another. This is not too cumbersome.

The big problem is common mode faults: end of civilisation, EMP erasing 
everything, you going nuts and wrecking the system, and so on. Once the 
risk  of all backups breaking is lower than this set of risks there is 
not much point in expanding.



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