Re: DiscoveryCh - AI

From: Anders Sandberg (asa@nada.kth.se)
Date: Thu Jan 25 2001 - 16:54:31 MST


Eugene.Leitl@lrz.uni-muenchen.de writes:

> Robin Hanson wrote:
>
> > I think you're missing the point. Some AIs can tax other AIs,
> > just as some humans now tax other humans. The relative
> > abilities of humans and AIs are irrelevant.
>
> You seem to see economy as a natural law. I don't think that
> taxation does exist for more than 10 kYrs of Earth's entire
> biological history, which counts several GYrs.

Taxation is a way entities can pay for public goods (or be forced to
pay for what some entity considers good, at least). There may very
well be alternatives that better will suit the economy of
hyperintelligences.
 
> My mental model of the future assumes species radiation and
> matter, energy and information flows in an ecology. I presume
> economics is a subset of that model, so I consider emphasizing
> postbiological beings in monetary relationships to other
> postbiological beings taxing said postbiological beings rather
> pointless. YMWV.

The important thing is that postbiological beings will likely be
involved not just in economic relationships but also form various
societies. This will make the development of institutions like
taxation to solve certain problems likely, even if the institutions
and problems are very different from current human one. Economics
isn't just about monetary relationships, but also about general
interactions involving utility (didn't Ludwig von Mieses suggest a
fancy Greek word meaning "getting enemies to work together" or
somthing similar rather than the zerosumish 'economics' as the name
for the field? A friend mentioned it today, but the exact term slipped
my mind).

When discussing AIs/SIs the economic ideas employed are often very
primitive. Often the issue is simply ignored or a kind of zero level
economy of getting what is needed through superior force or just
picking it up is assumed. This is generally linked with the idea of a
spiking SI, so that there is really a single agent with near total
autonomy. A more nuanced look at AI/SI economics would take into
account what has been learned in game theory and economics, the
economic context these entities come into existence inside and then
co-evolve with, the higher-order interactions that would emerge
between different kinds of agents and so on.

I have had a paper on the game theory and economics of human-posthuman
interactions sitting in one of my virtual desk drawers for several
years now. The reason that it is going slowly is that the more I think
about the issues and learn about economics the more inadequate I
consider my deductions - it is a paper that I am actually making
negative progress on! :-) At the same time the issue is important, and
I think that it *can* be tackled. Posthumans might not calculate taxes
like we do, but they will be economic entities. As Minsky pointed out
about aliens, math and economics might be the only thing we have in
common with them.

-- 
-----------------------------------------------------------------------
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


This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:05:21 MST