[ExI] Mind uploading via Gmail

Wed Sep 29 05:08:29 UTC 2010

 Again another recent TV plot. Namely 'Caprica' the gallactica spin off.
I have noticed that the time for such ideas to hit the mainstream is getting shorter. Perhaps a good indicator of technical progression toward singularity in action, the steepening elbow at least. Greater and faster world wide comms get ideas out there much quicker, thus the researchers (like the scif fi writers) are reacting faster to theory and making it reality.

I think that given a person's genome, a large sample of their writing
and (even for this crowd) a mind boggling amount of computation,
really advanced technology could generate an entity that thought it
was the original person and would have pretty much the same
personality and much of the original's memory.  It's possible--if I
become "weakly godlike"--I might try it with some of the authors I
admire but was unable to talk into cryonics.

-----Original Message-----
From: Keith Henson <hkeithhenson at gmail.com>
To: ExI chat list <extropy-chat at lists.extropy.org>
Sent: Wed, 29 Sep 2010 1:33
Subject: Re: [ExI] Mind uploading via Gmail

Far as I know, the idea of reconstruction from writings originated
with Hans Moravec as a rejection of cryonics.  A "we will be
reanimated anyway" kind of argument.  I threw cold water on the idea
at the time on the basis of it being a colossal expenditure of
computer resources to iterate the process till you got a simulated
Hans to write a duplicate of "Mind Children."

If you do have computer resources to waste, the idea is barely
plausible.  Charles Stross used it as a plot element in Accelerando
set in an era where much of the solar system had been converted to
computronium.  (Great book, should be required reading for anyone on
this list.  Load it from here
http://www.jus.uio.no/sisu/accelerando.charles_stross/plain.txt and
search for FAQ or resimulated.

Of course, you could also animate fictional characters, but for some
reason this was considered illegal (copyright?).

I think that given a person's genome, a large sample of their writing
and (even for this crowd) a mind boggling amount of computation,
really advanced technology could generate an entity that thought it
was the original person and would have pretty much the same
personality and much of the original's memory.  It's possible--if I
become "weakly godlike"--I might try it with some of the authors I
admire but was unable to talk into cryonics.

But it's not anything close to as good as cryonics or (better) just
living till people quit dying.

Keith

On Tue, Sep 28, 2010 at 10:18 AM, Giulio Prisco <giulio at gmail.com> wrote:
> As you say, "GMail, or services extended from GMail including
> BCI-extracted emotional responses" is more precise. I am thinking
> about overall feasibility in principle, not of a detailed engineering
> plan. Gmail is just an example, and of course we can use whatever
> works.
>
> G.
>
> 2010/9/28 Adrian Tymes <atymes at gmail.com>:
>> 1) Why would BCI-extracted information be a part of GMail?  Adding such
>> things makes it
>> fundamentally different than an email client - different enough that it
>> would likely have a
>> different name.  Thus, saying "GMail" specifically makes this consideration
>> moot.  (Now, if
>> you said "GMail, or services extended from GMail including BCI-extracted
>> emotional
>> responses", that starts lending some plausibility.  It may seem like a mere
>> semantic
>> difference, but it isn't.)
>>
>> 2) When they start extracting said information via BCI, they'll have already
>> started making
>> an upload of you, whether or not you've given permission.  They will likely
>> insist on
>> resolving any permission issues at that time, and therefore, any previous
>> permissions -
>> even ones given by subsequently dead folk - will almost certainly be
>> disregarded as legally
>> impossible to have given for want of context, regardless of the letter of
>> what was said.
>> (Judging based on existing precedent.  Even in the Singularity, some things
>> can be
>> predicted.)
>>
>> On Tue, Sep 28, 2010 at 9:04 AM, Giulio Prisco <giulio at gmail.com> wrote:
>>>
>>> I agree that the information isn't there... today.
>>>
>>> But suppose I keep using Gmail and similar services for 30 years,
>>> adding geolocation, video and BCI-extracted emotional responses as new
>>> options become available. Suppose I start adding childhood visual and
>>> sensory memories via BCI...
>>>
>>> Suppose all my friends do the same.
>>>
>>> Don't you think my own Gmail account and, in a lesser measure, the
>>> Gmail accounts of my friends, would have a lot of information about
>>> me, orders of magnitude more than today?
>>>
>>> My post will be interpreted by many as a joke, but to me in is (at
>>> least half-) serious.
>>>
>>> 2010/9/28 Adrian Tymes <atymes at gmail.com>:
>>> > "2) AI-based mindware technology able to reconstruct individual
>>> > mindfiles by
>>> > analyzing
>>> > the information in their aggregate Gmail accounts and other available
>>> > information, with
>>> > sufficient accuracy for mind uploading via detailed personality
>>> > reconstruction, is available."
>>> >
>>> > There is much reason to believe this is beyond the realm of
>>> > plausibility.
>>> > Not as in, "we
>>> > don't have the AI for it today", but as in, "the information to do this
>>> > simply isn't there."
>>> >
>>> > Does this record your day to day lives?  Not really.  (If you sincerely
>>> > disagree, chronicle
>>> > how you spend each hour of a given day, and then honestly look at your
>>> > outbox and see
>>> > how much of that is even mentioned.  Deliberately rigging the system,
>>> > like
>>> > doing a diary via
>>> > email for just that one day, doesn't count: this needs to be a
>>> > representative sample of most
>>> > of your days.)
>>> >
>>> > Does this record your experiences in offline conversations, or most of
>>> > your
>>> > pre-email life?
>>> > Again, no.  Certain details might come up; most of it won't.  Further,
>>> > it'll
>>> > be a subset of
>>> > what you remember, because you have to have remembered it at some point
>>> > to
>>> > have
>>> > mentioned it in email.  (While you may forget something that email will
>>> > remember, even if
>>> > you spend the majority of your life with GMail, the amount that you know
>>> > and
>>> > it has never
>>> > seen will remain substantial.)
>>> >
>>> > This is the kind of thing a critic of the Extropians can latch onto, to
>>> > dismiss the entire
>>> > movement as "obviously woowoo pseudoscience bunk".
>>> >
>>> > On Tue, Sep 28, 2010 at 2:47 AM, Giulio Prisco <giulio at gmail.com> wrote:
>>> >>
>>> >> http://giulioprisco.blogspot.com/2010/09/mind-uploading-via-gmail.html
>>> >>
>>> >> You may wish to give your permission too (you never know). Use the
>>> >> comment field or any other _public_ means.
>>> >> _______________________________________________
>>> >> extropy-chat mailing list
>>> >> extropy-chat at lists.extropy.org
>>> >> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>> >
>>> >
>>> > _______________________________________________
>>> > extropy-chat mailing list
>>> > extropy-chat at lists.extropy.org
>>> > http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>> >
>>> >
>>>
>>> _______________________________________________
>>> extropy-chat mailing list
>>> extropy-chat at lists.extropy.org
>>> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>
>>
>> _______________________________________________
>> extropy-chat mailing list
>> extropy-chat at lists.extropy.org
>> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>>
>>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>

_______________________________________________
extropy-chat mailing list
extropy-chat at lists.extropy.org
http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20100929/ed4e9474/attachment.html>