[ExI] [tt] Identity thread again
anders at aleph.se
Mon Apr 6 09:02:23 UTC 2015
Rafal Smigrodzki <rafal.smigrodzki at gmail.com> , 6/4/2015 8:19 AM:
On Thu, Apr 2, 2015 at 3:25 PM, John Clark <johnkclark at gmail.com> wrote:
On Thu, Mar 26, 2015 at Rafal Smigrodzki <rafal.smigrodzki at gmail.com> wrote:
> there will be a trade off between spawning independently goal-oriented processes (copies) and running non-sentient optimization and search processes.
You may create a copy of yourself to accomplish some specific sub-goal and after that has been accomplished you might want him to self destruct. However given the fact that he's identical to you as soon as he's created he would think it's a much better idea if you not him accomplish the sub-goal and then self destruct. In fact he might insist and sincerely believe that he is the original and you are the copy and it might be exceedingly difficult to convince him that is not the case. Actually it doesn't matter which is the original and which is the copy because they are identical.
### Well, I don't know about your copies, but my copy, after looking up its location and determining it is a copy created to accomplish a goal, would self-destruct in accordance with the decision that I, a mind identical to it, made before being copying.
Meanwhile the Anders-clan, believing 'Anders' to be the equivalence class of all sufficiently Anders-like processes is fine with having a lot of brief twigs on the main branches. It will have more resources running very fast Anderses, and it will also counteract the tendency of getting side tracked (since Anderses derived from my current instance likely have a high tendency for that).
The interesting case is how much I/we would consent to neural upgrading experiments. On one hand I am a bit more cautious there, but on the other hand I am also rather in favor of self-experimentation. Might be fun to see how it plays out.
My point is that some people may be very different in the styles they run their clans. Some do not accept at all being used as tools for somebody like them. Others think it is entirely OK.
Anders Sandberg, Future of Humanity Institute Philosophy Faculty of Oxford University
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat