[ExI] A science-religious experience

[Jason Resch]

On Fri, Feb 21, 2025, 1:49 PM efc--- via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
>
> On Thu, 20 Feb 2025, Jason Resch via extropy-chat wrote:
>
> >       Another way is to deny the existence of objective email, and
> affirm our
> >       opinion about events.
> >
> > Good and evil derive ultimately from subjective states of consciousness,
> which
> > makes their objective characterization and comparison difficult:
>
> I agree. This is my current favourite position. An ethical anti-realist I
> guess.
>
> > "For these words of Good, evill, and Contemptible, are ever used with
> relation
> > to the person that useth them: There being nothing simply and absolutely
> so;
> > nor any common Rule of Good and evill, to be taken from the nature of the
> > objects themselves; but from the Person of the man or, From the Person
> that
> > representeth it;" -- Thomas Hobbes in “Leviathan” (1651)
> >
> > "We have already observed, that moral distinctions depend entirely on
> certain
> > peculiar sentiments of pain and pleasure, and that whatever mental
> quality in
> > ourselves or others gives us a satisfaction, by the survey or
> reflection, is
> > of course virtuous; as every thing of this nature, that gives
> uneasiness, is
> > vicious." -- David Hume in “A Treatise of Human Nature” (1739)
>
> David Hume is a philosophical ninja! Jan Narveson developed
> contractarianism in
> a very nice way showing how rational minds, based on contractarianism
> should
> converge on a kind of libertarian ethics.
>
> > "It there appeared that we could not, on reflection, maintain anything
> to be
> > intrinsically and ultimately good, except in so far as it entered into
> > relation to consciousness of some kind and rendered good and desirable:
> and
> > thus that the only ultimate Good, or End in itself, must be Goodness or
> > Excellence of Conscious Life." -- Henry Sidgwick in “The Methods of
> Ethics”
> > (1874)
> >
> >
> > But I think it is still possible to give an objective definition of what
> > constitutes evil, as the previous examples suggest.
>
> How come, and it what way?
>

I gave a YouTube link to a video that explains it. It basically concludes
that morality is defined as the reconciliation of all systems of desire.

Here is the paper that video is based on:
https://philarchive.org/rec/ARNMAW

Here is Google's AI summary:

""Morality as what one desires" refers to a philosophical concept by Arnold
Zuboff, where he argues that true morality aligns with what a person would
genuinely desire if they had a perfect understanding of all relevant
information and consequences, essentially suggesting that a fully informed
self-interest is inherently moral."

The upshot is that if this idea is true, then the smarter superintelligence
becomes, the more likely that AI is to converge upon this truth, and in
accepting its truth it will be strongly motivated to act according to it's
truth, which is to act morally (which is defined as acting to reconcile all
systems of desire).


> > Further, although uncomputable in practice, for a god-like mind there is
> a way
> > to define morality objectively:
>
> I find examples with infinities, and postulates with god or god-like a bit
> unsatisfying. It seems to me, if you assume god-like powers, or god(s) you
> can
> basically justify or explain anything.
>

If it helps, replace god-like with unlimited computational power.

There are many things in compute science that are uncomputable, but
nevertheless their existence reveals something interesting, or points in
the direction of something that while imperfect, is nevertheless
instructive (such as the AIXI algorithm for universal intelligence). We
can't compute the AIXI algorithm, but it helps us define what a perfect
intelligence is.

Likewise while Zuboff's definition of morality is uncomputable, it helps us
to define what morality is.

And in both cases, we might discover sub-optimal, but computable shortcuts.

Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20250221/e5aa636d/attachment.htm>