Re: AGI investment (Re: What are "AGI-first'ers" expecting AGI will teach us about FAI?)

From: Samantha Atkins (
Date: Sat Apr 26 2008 - 14:06:07 MDT

Whether there would be any trading at all seems to me to be a very different
question than what the worth of an AI is or, as implied by the original
question, what its worth would be from our current pre-AGI perspective.
Even AGIs will find some means of deciding what to devote limited time and
capacity to. That can be modeled as a price structure. The relative value
of an hour (objective) of AGI labor would certainly dwarf many thousands of
hours of skilled human labor in some problem domains. Some problems the
AGI can tackle are beyond the capabilities of all humans combined. From
that POV an AGI is priceless.

That said, the relative price offered will be a function of the agreed
valuable assets (money or goods) availabled to be traded, the value of that
which is to be produced to those who wish to receive it, and the cost to
produce it. The cost for many items we value will fall rapidly toward zero
given MNT. AGIs competing to trade for values held by humans or simply
benevolent to humans are unlikely to charge exorbitant prices for things
cheap and easy to produce.

On Sat, Apr 26, 2008 at 8:06 AM, Byrne Hobart <> wrote:

> On Sat, Apr 26, 2008 at 4:14 AM, Samantha Atkins <>
> wrote:
> > If AGI enables machines to do all the work that humans would otherwise
> > > be paid
> > > to do, then how much would it be worth?
> > >
> > >
> >
> > Nothing (or 42) as monetary valuation in its former terms would
> > disappear along with most existential scarcity.
> Do you expect the AIs to be precisely identical in terms of skills, goals,
> and circumstances? You would have to prove that assertion to demonstrate
> that they wouldn't want to trade. To prove that they would want to trade,
> but not to denominate it in currencies, you'd have to explain how that's any
> different from having a 'barter' system of measurement: "How far away is
> Cleveland, Ohio?" "It is the number of seconds since the last sunset,
> multiplied by the length of Freud's first cat on its eighth birthday." Why
> oh why oh why would we stop converting things to standard measurements to
> compare them more easily?

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT