Re: SI Comparative Advantages [was Re: Free Will] (fwd)

Robert J. Bradbury (
Thu, 19 Aug 1999 08:51:01 -0700 (PDT)

> Anders wrote:

> I wrote:
> > The question comes down to *survival* in your ecological niche.

This becomes even *more* interesting if you realize that you not only have niches in space, but niches in time. With the capacities of an SI, they could predict the niche developments. So, SIs that use all their thought resources to master a current niche may be ill-equipped to master an emerging niche. For example, I, being a short term kind of guy might devote my resources to dominating the period when stars still burn (the "Bradbury era") Anders on the other hand, being a deeper thinker, could go off to some dark and remote corner of the universe and use resources that I'm unaware of or unconcerned about and plan for the day when the stars are all dead and only the black holes remain as power sources. This will eventually become known as the "Sandberg era".

The question becomes, as these eras proceed, they presumably become more "complex". Does the increase in complexity *guarantee* an increasing number of niches (for lesser SIs, life forms, etc.) or at some point (when you have the power to manipulate everything) do you simplify everything into a "perfect structure". [You could view the perfect structure at many levels -- atomic, molecular planetary, solar system, galaxy, etc.] This question would seem to revolve around whether you can structure (by design) completely controlled energy flows to do useful work. Niches exist today because the evolution of the universe creates opportunities to harvest those flows. But in the current situation (as we can view it), the evolution of life, intelligence, etc. is always playing "catch up". Do we get to a point where we have the capacity to design and build the entire structure from scratch?(!) It won't work if the structure has infinite complexity, but if much of the design is mostly simple copies of subunits, this seems feasible. So is the game really to restructure the universe into a perfect structure?

> > If SIs know all the laws of physics, how the universe works,
> > are able to watch & simulate any potential dangers, know that
> > the "magical fantasy worlds" are nothing but senseless diversions
> > for the uploads, etc. then *why not sleep*?

If you attain the "perfect structure", then there isn't much else to do.

> If we assume that SIs are hardcore realists (they all are decendants
> from Carter-Zimmerman polis) and avoid such fantasies,

You have lost me here, got a pointer?

> then the question is whether there are an endless supply of questions
> about the physical universe to explore.

If the universe is finite then it should be "knowable" within the limits of your storage technologies. If the laws it operates on are limited, then these should become known as well.

> This seems to relate to how complex it can be. It seems that
> the most complex systems around in this scenario are the SIs
> themselves, and their activities and artefacts.

The SIs can certainly make themselves complex by allowing niches to exist and allowing/promoting the evolution of energy harvesters into them. There would seem however to be a limit on how little energy you can use to do something useful with. That would seem to limit the complexity.

> I can't see any reason why not they could get involved in endless
> coevolutionary races towards increasing complexity, that are both
> relevant for their survival and interesting

I think the key term is "endless". Of course you can always setup the game so you end up with move, counter-move (i.e. an endless game of chess). But if complexity has to be based in reality (i.e. energy & matter), then I think there are physical limits to how much it can increase.

The thing which is unclear to me is whether I can create a "perfect structure" in the Bradbury era and that prevents the Sandberg era from ever arriving?