A view on cryonics (was Re: [extropy-chat] Bad Forecasts!)

[Eliezer Yudkowsky]

Slawomir Paliwoda wrote:
>>
>>This definition is trivial to deflate.  Just suppose that we interlace two
>>sets of neurons and synapses, neither interconnected with the other, but
>>both occupying the same volume of space.  Better yet, suppose that we run
>>two minds on the same neurons, the neurons having time-sharing registers
>>that swap between the two identities twenty times per second.  How does
>>your spatial algorithm distinguish between these two minds?
> 
> The algorithm distinguishes between two minds by tracking their unique
> trajectories not only in space, but also in time. Your argument *seems* to
> rest on the assumption that just because the space-time parameters of two
> minds register values that are very close, the trajectories will be
> indistinguishable. Even though the values are going to be very close, they
> will never be the same.
> 
> Let's use your example where minds A and B share the same computational
> medium whose proverbial "CPU" swaps two identities every 1/20s.
> 
> The position of an electron that flows as part of mind A is recorded every
> femtosecond. At time t1, the electron is at x, y, z, 1, and we record that
> position in the log. Then, at time t2, we make a new entry, x+1,y+1,z+1, 2.
> The entries pile up until CPU switches to mind B after 1/20s.
> 
> Now, suppose that an electron flowing as part of B makes the journey from x,
> y, z, to x+1, y+1, z+1, just like the previous electron. However, this
> particular electron's time parameter is not going to be 1 or 2 anymore. A's
> record will then necessarily differ from B's reflecting different
> trajectories of these two minds in space-time.
> 
> Now, think of the overall process and all the positions mapped by all points
> of matter whose flow in space-time leads to the emergence of a mind. Unless
> you find a way to make A and B's matter flow exactly within the same
> location AND time, I will always be able to verify their identities by
> investigating A and B's recorded trajectories in space-time.

Aha!  But if you're tracking spacetime coordinates on such an extremely 
fine grain, what you're actually doing is tracking every element of the 
entire computation.  No longer can you point to two different people on two 
sides of a room, and know complacently that they shall always be distinct 
for they are separate lumps of matter.  In particular, to establish 
continuity, you will need to track chains of causality through that complex 
skein of computation - not just the fact that an electron was here and then 
moved there, but the fact that it bumped into another electron which bumped 
into another electron and so on.

Now what if an electron in this skein of computation bumps into *two* 
electrons?

In this fashion may a person be duplicated so that there is no copy, there 
are two originals.

-- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence