[ExI] simulation as an improvement over reality

Eugen Leitl eugen at leitl.org
Thu Dec 30 14:31:33 UTC 2010


On Thu, Dec 30, 2010 at 03:15:48AM -0800, Samantha Atkins wrote:

> > But can teleoperation become so good that it reduces the instrumental interest in moving a mind to the moon?
> 
> Not as long as we are limited by the speed of light in signal 

Teleoperation is a crutch. It allows slow meat that can't travel
(because it's expensive, or dangerous) to operate remotely.

In case of fast minds (10^6 to 10^9 speedup to current
wall clock) the relativistic latency limits teleoperation
(or telepresence, as in avatars) to very short distances.

> propagation.  Locality of mind/intelligence will be important 
> as long as we are so limited.  You can't teleoperate well 

Importance of locality will increase as the systems get
more advanced.

> enough for many operations to do work in GEO from earth much 

It's good enough to precision kill people halfway around
the world.

> less on the moon.    The light speed barrier also means that 
> there will be many many AGIs when they are developed (and are 
> cheap enough to produce and replicate) rather than only one.

Precisely. In a relativistic universe with difference in local
state there cannot be a single acting agents, as each local
instances will bifurcate. Instead of synchronized monoclone 
you wind up with a population with a locally different
response, to locally diverging physical layer input. 
 
> More autonomous systems are increasingly being deployed.  
> Eventually it isn't teleoperation but more a matter of remote supervision.   

In order to walk you must first learn to crawl.
 
> It would be good for space robotics to not model their motor 
> abilities after humans.  We were not evolved for work in space 
> and thus do not exhibit an optimal design.

The main reason current aerospace teleoperation
platforms are anthropomorphic is because they have
to be operated by people, and share infrastructure
which was designed by people for people.
 
> > Given the transport problems over the last year (due to an overstretched infrastructure and too much institutional risk aversion) I have made the following "probable surprise": more transport system crashes will occur due to fairly normal fluctuations, and together with expanding IT infrastructure more and more people will be fine with at least social telepresence.
> 
> I have been expecting telework and telecommuting to be 
> bigger than they are for a very long time.  It amazes 

While I realize why we're not seeing more of that it
appears that few people realize the reasons, orelse the
problems would have been addressed already.

Instead, we're getting semi-random advances, where
random advances (e.g. simulated environments, avatars 
and motion capture trackers) result in sudden improvements.

> me that hundreds of billion in production are wasted 
> every year driving to jobs that can be done as well 
> remotely.   Not to mention the energy costs and risk 
> to life and limb.   But we chimps have to go to some 
> location where a more or less alpha chimp can keep 
> an eye on us and where we can send all the chimp 
> signals on the evolution habituated channels.  

Precisely. 
 
> > Whether that can be carried over to useful telework remains to be seen.
> 
> It certainly can.  A large part of IBMs workforce is now remote.   
> We have to learn to do this anyway as many we could benefit by 
> working with are not close enough to meet physically except a 
> few times a year at most. 

There are still many parts in the technology puzzle
missing before telepresence and teleoperation becomes 
the rule.

Some of it is infrastucture. You need reliable, low
latency broadband wireless everywhere, and rental platforms
(even docked quadcopters with remote video would
be a start) with good coverage. Apart from insurability,
recharging infrastructure, low-overhead payment, there
is also handicaps like navigability (Segways do
poorly on stairs), and the like.

The list is endless, and you need all of this
before the culture accepting it ever can take
hold.
 
> Those that haven't read it may enjoy "This is not a game" 
> by Walter Jon Williams. 

Right to my wishlist.

-- 
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