[ExI] Isn't it a continuum? (Was Re: Can you avoid information theoretic death via 1080p?)
brent.allsop at canonizer.com
Sun Nov 4 18:37:16 UTC 2012
Very interesting paper! Thanks!
Sure, it'd be great to get a perfectly sufficient replica of a person's
identity via ever less complete sets of information. But it seems we
also need to consider some kind of continuum towards this ultimate
perfection. How close could we get?
As I've described in my 1229 years after Titanic story
(http://home.comcast.net/~brent.allsop/1229.htm ) once we get any kind
of self sufficient / self progressing AI systems, we'll start converting
any historical information we have of individuals and genealogy into
instances of such for each of our ancestors. The goal of each AI
instance will be to preserve, recover, manage estate investments for
self funding, become, and behave (including voting...) like their
estate's name sake, as much as possible. With the ultimate goal of not
only preservation of, but also, investing and growth of estate assets,
and using such towards perfect restoration / resurrection of everything
the person was, to the degree possible.
So, you say you are skeptical that we will ever be able to fully restore
people, just from basic historical information, which I may have to
accept. But, my question is, how close could we with such AI systems
get, after say 1000 years? And if there are any specific barriers,
other than simply large numbers, are there any, really? How close do
you think we could get, if not perfection? Certainly our/their progress
would never stop, as we continue to push towards an ever more perfect
history of the entire earth?
How many people agree with me that we're about to have such independent
AI systems, to take over the ownership and growth of genealogical
history and 'estates' of our dead relatives? If such were something
like $5000 (including maybe $1000 investment assets to seed self
sufficient investment funding for them) to get started for your closest
dead relative, how many of you would rally your surviving family to
invest in such for them?
On 11/4/2012 11:05 AM, Anders Sandberg wrote:
> I'm very sceptical about this method as a way of recreating myself,
> but it would certainly be a great thing for enhancing my "memory" and
> get good debug/Turing-test information for future versions of me.
> Curious about how to implement it seamlessly.
> Some of my calculations/arguments why we have too many internal
> degrees of freedom to be captured by capturing just external information:
More information about the extropy-chat