[ExI] Mono or Poly?

[Jason Resch]

On Tue, Mar 4, 2025 at 10:39 AM Adrian Tymes via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> On Mon, Mar 3, 2025 at 4:25 AM Rafal Smigrodzki via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
>
>> Nothing is definite but I bet you 100$ to 1$ that ten years from now
>> there will be a superhuman AI on this planet, most likely a vastly
>> superhuman AI, as much smarter than you or I as MuZero is better at Go than
>> Lee Sedol.
>>
>
> No bet.
>
> If you win, then by your hypothesis, money will basically be worthless.
> Someone with that attitude, I don't trust will have (or at least, be able
> to part with without complications) the money to pay up if I win.
>
> More importantly, how exactly would we define and measure this?
> Internationally recognized ladder rankings of Go, as might numerically
> compare MuZero to Lee Sedol, do not directly translate to measurements of
> overall intelligence.  The bet may prove impossible to adjudicate.
>
> Though, by certain measures there already exist superhuman AIs.  Some
> would credit the most well known publicly usable LLMs as such, and
> computers have been able to outcompute humans for decades.  If we used
> those measures, you'd win the bet already.
>
>
>> But I only say it will be the last fork in the road that we take on our
>> own, not the last developmental fork ever.
>>
>
> If superhuman AGI is achieved by upgrading human intelligence - possibly
> mind upload, more likely external augmentations resulting in an entity that
> is a merger of AI and human - would that still be "we" who drives further
> forks after that?
>
>
>> Lots of important choices will be made, but not by unenhanced humans who
>> will pretty much equal mice in their ability to affect the world.
>>
>
> That I will agree with, and go further: forms of this exist today.  What
> impact on the world does a third world subsistence farmer have, as compared
> to a typical worker in an AI company, considering the training and
> resources the latter has to increase their productivity?
>
>
However we might choose to measure it, in terms of patents or academic
papers published/year, transistors you can buy for a dollar, even economic
measures of GDP, it is hard to say that we're not in the midst of very
interesting times:
https://ourworldindata.org/grapher/global-gdp-over-the-long-run

An analysis (
https://alwaysasking.com/when-will-ai-take-over/#The_Doomsday_Equation )
done in the 1960s, found that the doubling time for the world economy is
shrinking. Based on their data, it suggested that the doubling time would
reach 0 on Friday, the 13th of November 2026.

The population of human inventors largely ( https://arxiv.org/abs/1206.0496
) determines the rate of technological and economic growth. With the rise
of AI, the potential population of competent inventors will soon, and for
the first time, become decoupled from the human population. This could
easily occur sometime over the next few years, in line with this 65-year
old projection.

Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20250304/6a60e4f5/attachment.htm>