Re: Sleep, Memory, SIs & Evolution [was Re: curing sleep]

Anders Sandberg (asa@nada.kth.se)
01 Nov 1999 15:35:21 +0100

Fascinating. So instead of living in an universe filled with sleeping super-cats we might be living in an universe filled with super-cave fishes...

"Robert J. Bradbury" <bradbury@www.aeiveos.com> writes:

> Given the above, you look at the SI, its in an environment similar
> to the cave-fish. It wants to *live* in the moment. Since nothing
> around it is a threat (all threats being handled by preprogrammed
> sub-systems) it becomes pointless to remember anything! Just
> as the cave-fish doesn't need to exercise its memory pathways
> to have stuff available for future recall, neither does an SI!
> So all the resources get devoted to short-term memory and
> computation of the current virtual reality.

Makes the behavior of the Transmuters in Egan's _Diaspora_ seem quite rational. Maybe he was onto something?

> Extending my "We don't talk to nematodes and SI's don't talk to us"
> quote, "We don't talk to nematodes and SIs have forgotten us!"
> Both humans & SIs are totally "unaware" of the things they take
> for granted. Humans rarely think about or devote "long term memory"
> to breathing (exceptions being scuba diving, plane decompressions,
> astronauts and strangulations). Why should SIs devote thought or
> attention to the universe or the sub-entities in it? *Only* if
> surviving in the universe *indefinately* is a *possible*, but
> *very difficult* problem, will SIs pay attention to the universe.

If Dyson is right about the universe being maximally interesting/complex, then it seems likely that the SIs would be up and running with long-term memories. If the complexity of the universe is too small, then being memoryless or asleep makes sense (however, SIs have autoevolved from natural entities or been created by them, so they might have residual fears of removing aversions against Eternal Return and similar stuff). If the universe is too complex to be dealt with (like living in a type III cellular automaton with gaslike chaos) intelligence beyond a certain point doesn't help you any further, and there would be an upper bound of necessary intelligence - resources would better be spent on safety, backups and "entertainment".

> If its impossible or easy (which seem more probable) they may
> abandon the costs of long term memory for
> "Live for today, for tomorrow you *will* be dead".
> or
> "Live for today, for tomorrow the game is reinitialized".
>
> So, now we are left with one final "extropic" question --
> "Do you want to evolve slowly or quickly to that point?"
>
> If the energy available to do work is finite, and if we can
> use it slowly or quickly, but the work (thought quantity) that
> it can accomplish is the same, which do you prefer?

To me it doesn't matter, since I prefer to measure time in thought. Not much time passes when I'm asleep, but a lot when I'm meeting fellow transhumanists. So from my perspective the available time above is constant, and the distribution of it in physical time is a matter of aesthetics (and some practical problems). Most likely I would like to do it slowly so I could see what happens.

"What do you want then, Presence?"
"I want what I already have! Eternal wonder, eternally fulfilled... Not the eternal, even, just the Indefinite, that is where all beauty is... I'll wait out the heat-death of the Universe to see what happens next! And in the meantime, isn't it something, all of it?" - Bruce Sterling, Schismatrix

-- 
-----------------------------------------------------------------------
Anders Sandberg                                      Towards Ascension!
asa@nada.kth.se                            http://www.nada.kth.se/~asa/
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y