[ExI] destructive uploading

Eugen Leitl eugen at leitl.org
Tue Sep 6 14:14:26 UTC 2011


On Sat, Sep 03, 2011 at 12:50:31PM -0700, Ben Zaiboc wrote:

> Well, as much as uploading can be regarded as resurrection, I suppose it is.

Be sure to specify the type of reinstantiation desired in your
cryopreservation contract, should you be squeamish in
the sense of "Do you think that's air you're breathing now?".

Notice that this could cut your options down to none, so
make sure to leave some escape clause (do not resuscitate
unless XY). 
 
> Of course, you'd be able to run, fly, teleport, jump, displace, simply be somewhere... whatever you like (and whatever the software supports).

Notice that relativistic latency would require physical relocation
of the image in case of superrealtime (up to 10^6 to 10^9 the
wall clock rate) rate of execution instead of just casting your 
sensorium around (which is a bit like today's virtual
environments, be it centralist SL or distributed determinism
like OpenCroquet).

I presume transparent migration would be less crippling than
halting, copying and resuming elsewhere, as the image size would
be very large by today's notions and result in a Rip van Winkle
effect (minus the wrinkles).
 
> Er.. you don't need to breathe!  You're an upload, remember?  You'd have to build in a requirement to 'breathe' in the virtuality before swimming with mermaids presented a breathing problem.  Your world consists of software, so there'd be no problem having a set of controls to specify which biological (and other) parameters to emulate.  You could specify that you have to 'breathe', have a pulse, etc., etc. You could specify that you have a sense of breathing, regardless of your environment, or whatever.

Minimizing suprises when instantatiated it's probably
a good practice.
 
> > Did George Carlin ever get wind of this scheme?  
> Not that I'm aware of, although I'm sure he would have had great fun with the concept.
> 
> > Is there somewhere we
> > can sign up for uploading?  
> 
> Not yet.
> 
> > Is George already there?  
> 
> No. Sadly, George is nowhere.

Gone into the great bit bucket into the sky.
 
> > What kind of
> > operating systems will support us?  
> 
> Who knows?

For efficiency reasons alone there will be no software
in the emulation engine itself (other than you,
which is not at all like software, think of
it as a 3d integer gas lattice). For security
(pwnage) reasons there should be as little 
software as possible elsewhere, and of course
there's the issue of execution on untrusted
systems when you do a live relocation to a remote
location not owned and built by you.

If you thought today's security situation is a nightmare,
I have bad news for you...
 
> > I like JAVA myself so I can
> > virtually exist anywhere that has a runtime, plus I can use it to stay
> > alert.
> 
> LOL.
> Just remember that Java is owned.  

Pwned, even.

> I'd probably prefer Ruby, or more likely, there will need to be some as-yet-uninvented OS for the kind of hardware and architecture needed to implement uploading.

The basic capabilities are simple enough. Once you start
allowing more features (like live migration, griefing
mitigation and such) it will get hairy. It can't get too
hairy, as it will be then impossible to get sufficiently
watertight, and also too slow (gate delays are just as
scarce as iterations in a loop, if you insist in common
metaphors you unroll everything directly into virtual
hardware, running on a minimal physical hardware aka
computronium).

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE



More information about the extropy-chat mailing list