[ExI] Upload/Download (was: Re: Fwd: Open Individualism)
Keith Henson
hkeithhenson at gmail.com
Thu Jan 18 19:12:55 UTC 2024
"but due to what we see at Tabby's Star, things might go differently."
All we can see at Tabby's Star is dimming. But we can tell something
about whatever is blocking the light. (I ran the math previously.)
If what is blocking the light is a data center, it is more than 400
times the area of the Earth, and even out at 7 AU, it is soaking up
about 1.4 million times the total human energy use.
The edge-to-edge speed of light delay is in the 2.5-second range.
That might tell us something about the clock rate they run at,
assuming they have a similar tolerance for communication delays.
Or perhaps not. I sincerely hope we are not looking at aliens because
we don't need the competition and it looks like they have spread to at
least 24 stars.
On the other hand, they got through their local singularity some
thousands of years ago, so perhaps we can. On the gripping hand, we
could be looking at AIs and even simulations of the original
biological race are extinct.
We live in strange times indeed.
Keith
On Thu, Jan 18, 2024 at 10:52 AM Keith Henson <hkeithhenson at gmail.com> wrote:
>
> Jason wrote
> > Also, consider that if we upload ourselves to computing substrates that run our minds a million times faster, then the "real world" becomes incredibly slow.
>
> I wrote about this 12 years ago, but due to what we see at Tabby's
> Star, things might go differently.
>
> Transhumanism and the Human Expansion into Space: a Conflict with Physics
> By: Keith Henson
> Published: April 12, 2012
>
> Orbital Space Settlements
>
> There are deep memetic roots between expansion into space and
> transhumanism, largely through Eric Drexler, creator of the modern
> concepts of nanotechnology and an early activist in the post-Apollo
> flowering of the space movement. I was deeply involved in the
> post-Apollo space movement (L5 Society) and moderately involved in
> early transhumanism, having been a reviewer of Engines of Creation
> when it was being written and a participant on the early Extropian
> mailing list where such concepts as Jupiter brains, M brains, and
> computronium were first discussed. No less an authority than RU Sirius
> has referred to me as “sort of an ur-transhumanist”.
>
> I have long had misgivings about large aggregations of computing nodes
> forming a mind because of speed-of-light delays. That will reduce
> “thinking speed,” since a mind cannot “be of one mind” if much it is
> not aware of the current situation due to speed-of-light delays.
>
> There is an analogous problem within a culture. We have a world
> culture today headed toward a monoculture because electronic
> communications on fiber optics has cut the delay to getting
> information from one side of the world to the other to sub-seconds. It
> hasn’t always been this way.
>
> Ferdinand Magellan’s expedition of 1519–1522 became the first to
> circumnavigate the Earth, a project in its day not unlike the moon
> landing. My point in mentioning this is not the privations they
> suffered (extreme) or the low survival rate (232 of 270 died before
> they got back, including Ferdinand) but the sheer length of time
> (three years) they were out of touch.
>
> How much communication delay can a society tolerate before it breaks
> up into smaller units? There are historical examples; the Roman Empire
> broke up partly because their communications failed, but I am not sure
> how they apply in a post-singularity world.
>
> At the speed of light, it takes 2/15 of a second to go all around the
> planet. (40,000 km/300,000km/s) What’s to cause communication delay?
>
> The problem is uploaded humans speeding up. Human brains are
> asynchronous, but, given reaction times, we can impute an equivalent
> clock rate of ~200 Hz. Which means a human brain (or brain equivalent)
> running in moderately fast hardware could run a million times faster.
> (200 MHz is not fast hardware.)
>
> You might ask, “Why would humans do such a silly thing?” Because
> intelligence is a large factor in sexual attraction. This was probably
> established in the Stone Age. Of course, intelligence is valuable
> outside of sexual attraction, being correlated with many other
> personality and life-history traits, and is especially valued by
> transhumanists. One aspect of being smart is thinking fast, or at
> least thinking faster than the person you are trying to impress. That
> leads at once to a runaway “Red Queen” situation where, when we can
> run our thinking faster, we would rapidly push the computational speed
> to the limit, whatever it is.
>
> Speeding up will mandate uploading into faster hardware. Humans could
> think a little faster in existing bodies, but not a lot. The problem
> is that the faster you run your brain, the more the world around you
> seems to slow down. With only a modest speedup, movement would seem
> like wading through molasses. If you desire serious speedup, it
> probably has to be in a simulation of the environment to match your
> faster perception.
>
> Speeding up rapidly leads to a situation where distance causes severe
> communication delays. We hardly notice telephone communication delays
> unless they are going through satellites. Speed us up a million-fold
> and the communication round-trip delay gets to be serious. A
> million-to-one speed-up would impose a subjective round-trip delay of
> three days from one side of the earth to the other. The subjective
> round trip delay to the moon would be two months. The delay from
> hearing back from a computronium node on the far side of Earth’s orbit
> would be 2,169 years, a long time even by Ferdinand ‘s standards. The
> subjective delay talking to an interstellar spacecraft 10 light-years
> out would be 20 million years. I don’t know what the maximum is, but
> two thousand years seems a long response time for a single polity.
> Twenty million subjective years to get an answer back is just
> ridiculous.
>
> The speed-up limit may be 100 times as high.
>
> Due to this line of thinking, I no longer think it’s practical to
> surround a star with computronium. Instead, I suspect population
> centers will shrink to sizes in the few hundred-meter range and sunk
> in the deep oceans for cooling.
>
> Taking two subjective seconds as the upper limit round-trip delay for
> telephone-like communication, the distance could be up to 300,000 km
> or 300 million meters.
>
> For a million-to-one speedup, that means that all the communicating
> nodes can be no more than 300 meters apart, i.e., configured as a
> sphere 300 meters in diameter with a hole to pump water in or out (for
> cooling). The area of the sphere is ~283,000 square meters.
>
> We now need a number for how big a human-class computer might be. Eric
> Drexler gave considerable thought to this problem and came up with the
> volume of a coffee cup. For simple calculations let us take that as a
> 10-cm cube, or 100 minds to the square meter arranged one layer deep.
>
> Taking the area times Eric’s number, the population of fast-uploaded
> humans per communicating fast culture could be as high as 28 million.
> If each drew 20 kW (1,000 times the 20 W our biological brains use),
> the total power draw would be 540 GW. The other factor of a thousand
> in the million-to-one speed-up comes from the hardware being more
> efficient than biological brains.
>
> With a fair amount of pressure difference and many fine passages, a
> water flow of one liter per second through each person (10 square cm)
> would carry off 20 kW with a temperature rise of about 5° C. The deep
> ocean is definitely the place for this community.
>
> If you wanted to talk to someone without a delay of up to a second,
> one of you could move close to the physical location of the other
> party via a “core swap.” The community might have ten percent or so of
> empty brains to facilitate moving around.
>
> With each order of magnitude speed up, the maximum community size
> falls by a factor of 100. That is, a community running at only 100,000
> to one could have 2.8 billion residents and one running at 10 million
> to one would only hold 280,000. The hundred-million-to-one speedup
> habitat would only hold 2,800. Talk about a physics-enforced
> oligarchy!
>
> Would you want to move from a slum where there was only a
> 100,000-to-one speed-up to one of these “elite” places with a
> million-to-one speed-up? Do you yearn for the 100-million-to-one
> existence?
>
> This might sound like total nonsense, except we can already see the
> beginnings of serious economic concerns with the speed of light. The
> avatars that run programmed trading must be close physically to the
> computers that run the stock exchanges.
>
> “In the U.S., high-frequency firms represent only 2 percent of the
> 20,000 or so trading firms operating today. But they now account for
> nearly three-quarters of all trades.
>
> “And the average time a stock investment is held these days is 22
> seconds. If time is money, microseconds are now millions. In a recent
> so-called TED talk on cutting-edge technology, tech whiz Kevin Slavin
> wowed the audience by describing buildings now being hollowed out in
> Lower Manhattan. Why? So that high-frequency trading firms can move in
> and get as close as possible to New York’s point of entry for the
> Internet at a so-called carrier hotel in Tribeca.
>
> “. . . . this is really where the wires come right up into the city.
> And the further away you are from that, you’re a few microseconds
> behind every time. These guys down on Wall Street, they’re eight
> microseconds behind all these guys going into the empty buildings
> being hollowed out up around the carrier hotel.
> “Just to give you a sense of what microseconds are, it takes you
> 500,000 microseconds just to click a mouse. But if you’re a Wall
> Street algorithm and you’re five microseconds behind, you’re a loser.”
>
> – from Kevin Slavin on algorithms
>
> One consequence that Eric Drexler discussed in Engines of Creation
> (end of Chapter 5) was a million years of science and engineering
> being done in one year. He didn’t discuss the subjective effect of a
> whole society uploading and subjectively experiencing a million years
> per calendar year.
>
> If uploads happen by mid-century, then by the end of the century human
> culture could experience 50 million subjective years (or more).
>
> As a conclusion, if humanity takes the speed-up route, then I don’t
> see a future for M brains, S brains or even Luna-sized brains, and the
> maximum size of a communicating civilization becomes a good deal
> smaller than the Earth.
>
> Unless, of course, we can find a way around the speed of light.
>
>
> Keith
>
> On Thu, Jan 18, 2024 at 8:18 AM Jason Resch via extropy-chat
> <extropy-chat at lists.extropy.org> wrote:
> >
> >
> >
> > On Thu, Jan 18, 2024, 6:51 AM Ben Zaiboc via extropy-chat <extropy-chat at lists.extropy.org> wrote:
> >>
> >>
> >> Keith Henson said:
> >> >> I suspect that people will prefer the uploaded state, but I also
> >> suspect that people will save bodies.
> >>
> >> Jason Resch said:
> >> > Yes, I think there will still be many reasons for interfacing with
> >> the external world.
> >>
> >> That remark seems to imply you think that uploads would not interface
> >> with the external world. I don't see why they wouldn't, and I'd
> >> certainly expect they'd be able to if they wanted.
> >
> >
> > Perhaps, but then again, that might be considered too dangerous. Future computing substrates might be too fragile and too subtle for any human mind to manage in any way. And if they break something they might kill a billion people running on that node.
> >
> >>
> >> If nothing else, uploads should be able to maintain their own hardware.
> >
> >
> > In the near future we will be able to build hardware they maintains itself. Or we could envision nanobots that maintain things.
> >
> >>
> >> If you were an upload, would you want to entrust the care and
> >> maintenance of your hardware to biological beings? That will probably be
> >> the case at first, but in the long run, we don't want to depend on
> >> beings that can be disabled or destroyed so easily. It would be ironic
> >> if civilisation was wiped out by a disease, say, in spite of a good part
> >> of it being resistant to the disease!
> >
> >
> >
> > My views are inspired, among others, by this vision of the future:
> >
> > http://frombob.to/you/aconvers.html
> >
> > Also, consider that if we upload ourselves to computing substrates that run our minds a million times faster, then the "real world" becomes incredibly slow. It would take subjective weeks for each click tick of each second in the real world. They really does isolate the worlds. Even the speed of light is slowed to something like a jetliners, so sending an text message to the other side of the world is something that takes hours to arrive.
> >
> > To me this suggests the future will be not only vastly miniaturized but also localized. We will create many copies (for redundancy) of large collections of individuals (perhaps all of humanity) in localized clusters for speed of access and interface.
> >
> > There would be no way any human could maintain such machines, as they would operate on atomic or subatomic scales, and their million+ fold time speed difference makes the physical world a static alien thing. You could run around in the rain at normal speed in the real world with a robot body for a few minutes, but it will cost you years of subjective time, all your family and friends will go years without seeing or hearing from you while you dance in the rain. But you could have had the exact same experience in virtual reality running at the same speed as everyone else in the uploaded state.
> >
> > Jason
> > _______________________________________________
> > extropy-chat mailing list
> > extropy-chat at lists.extropy.org
> > http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
More information about the extropy-chat
mailing list