[ExI] thought experiment part 1

[Adrian Tymes]

On Wed, Dec 10, 2025 at 8:25 AM John Clark <johnkclark at gmail.com> wrote:
> On Wed, Dec 10, 2025 at 7:23 AM Adrian Tymes via extropy-chat <extropy-chat at lists.extropy.org> wrote:
>>> >> technology is rapidly approaching the point where machines alone can do what somebody wants to get done, completely bypassing human labor, and thus also bypassing the abstract concept of "money".
>>
>> > Only in certain fields. [...]  Not politics, nor most things dealing with agreements that fundamentally involve other people.
>
> Politics is about choosing a small set of people that get to make a much larger set of people do what they tell them to do, and the foundation of that persuasion has always rested on money, until now.

Not entirely.  It's played a major role, sure, but there are plenty of
examples from modern and olden times of the better financed campaign
in a race losing.  Money usually does not directly buy votes.  It is
possible that AI will help level the playing field and make
non-financial factors more important - but it is not the case that AI
can get multiple people simultaneously elected (in the generally
publicly recognized version of "who was elected", niche cases that
amount to competing antipopes aside) to the same term of the same
office for which only one person can serve at a time.

I might want to be President of the United States.  It is a safe bet
that machines alone, without significant input from me or any service
I might buy with money, will not accomplish that goal this side of
2050, and probably not this side of 2100.  It is definitely the case
that machines alone will not be able to simultaneously make both you
and I President of the United States (in the singular, real life
version, leaving out all simulated or delusional instances that might
be merely labeled "President of the United States") at the same time
in that time.

>> > There is some doubt as to how creative the current sorts of AIs are turning out to be, although even if they always require a human to spark any truly new idea, they can fan the flames of millions (perhaps some day billions) of creators to where it can seem like the AIs themselves are coming up with the ideas.
>
> I can find no evidence to support the claim that humans have some sort of mysterious "spark", a secret sauce, that an AI could never duplicate.

I dud not say "all AIs that could ever theoretically exist".  I said
the current sorts (the ones that might cause a Singularity in the next
several years, which is the time frame you appear to be talking
about), for which you need to turn to tests of the creativity of the
AIs that actually exist today.