<!DOCTYPE html>
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
On 10/11/2024 21:54, Keith Henson wrote:<br>
<blockquote type="cite"
cite="mid:CAPiwVB4FLFAwKXNPw5sHuKYfKUOLE0rcKJA=j3GmfKJ7YCfzig@mail.gmail.com">
<pre>Uploaded biological creatures would need data centers. The centers
need energy from their star and radiation heat sinks, both favoring
large areas. This conflicts with the presumed need to communicate
with minimal delay due to the speed of light.
I don't think free-floating computation nodes are a reasonable
engineering solution.</pre>
</blockquote>
<br>
Maybe I used the wrong word. I didn't mean individual nodes in a
larger computing system, like neurons in a brain or something
similar. I meant the same as you when you say 'data centres'. Each
one would represent something in the range from a single individual
to an entire civilisation, but not smaller than a single brain.<br>
<br>
So what does 'large areas' mean, when it comes to energy collection
and heat dissipation?<br>
And what is an acceptable minimal delay in communications?<br>
<br>
A balance between these two would determine the size and shape of
the individual data centres.<br>
<br>
It would seem to indicate a structure like a small spider in the
middle of a large web, or a marble embedded in the centre of a large
circular membrane. The marble would be the data centre, maybe on the
metre scale, and the membrane the energy collection/dissipation
part, on the km scale (10s, 100s, or 1000s of km).<br>
<br>
Processing speed, or 'thinking speed' would determine the range of
sizes needed, and I think at some time you mentioned a one-metre
sphere sunk in the deep ocean as being suitable for a data centre
for minds running at 1 million times biological thinking speeds. I'd
assume that something similar in free space would be like I describe
above.<br>
<br>
Slower speeds would mean larger structures.<br>
<br>
So how large do you think one data centre (including its energy
collection and heat sink) would be likely to be?<br>
<br>
I'm guessing something between 10km and 1000km in diameter, and very
thin. So trillions of these could be spread out round a star, and
probably be fairly invisible from a distance.<br>
<br>
Or do you think the data centres would be humungous?<br>
<br>
As you say, this conflicts with speed-of-light delay, and I'm
wondering what would be the point of connecting individual
collections of computing substrate with solid mass rather than empty
space? Would this be necessary for the energy collection etc? So
you'd have a huge plate of solar collectors and heat sinks, with
individual processing centres dotted about in it? Or something else?<br>
<br>
I'm not a numbers guy, and I know that you (and others here) are,
and I know better than to argue with the numbers, so what do the
numbers say? Lots of small (relatively small, on a solar-system
scale. Say no bigger than 5000km) units spread out around a star, or
one (or a few) ginormously huge structures*, hundreds of times
bigger in area than the surface of earth, easily detectable from a
distance?<br>
<br>
Or are there other options that make more sense?<br>
<br>
And how does what we see at Tabby's star etc., fit in with this?<br>
<br>
--
<pre class="moz-signature" cols="72">Ben
* Interesting question: How big could a huge flat (or curved) continuous structure be? There must be factors that limit the size
(gravitational forces, tensile strength, perhaps other things)?
Ben
</pre>
</body>
</html>