Re: QUICKIE PREDICTION NEEDED

Eliezer S. Yudkowsky (sentience@pobox.com)
Sun, 25 Oct 1998 13:13:02 -0500

E. Shaun Russell wrote:
>
> Eliezer wrote:
> >
> >By the year 2050 at the absolute latest (assuming no civilization-destroying
> >wars) I expect that there will be nothing recognizably mortal, much less
> >human, in our Solar System. It's a matter of positive feedback.
>
> Out of curiosity, what parameters are you basing this hypothesis on?
> True, technological progress is being manifested at a rapidly quickening
> pace, but the moral mindset of this planet's human habitat is generally
> stubborn, pessimistic and myopic. Given the opportunity of "right here,
> right now" transcendence into a higher form, it is doubtful that every
> single mortal would want to upload.

Anyone who thinks there is going to be some kind of choice involved is being naive and optimistic to the point of lunacy. Do we give children a choice as to whether or not they have to grow up? Most imaginings on this subject basically resemble a bunch of children who have never seen or heard of adults discussing whether they ought to grow up and what kind of wonderful society they'll make if they do.

> There is far to much human stigma.

First-graders attach a large stigma to education. It hasn't made a noticeable difference. They are not major factors in politics.

> Furthermore, the assumption that technology will have progressed *that* far
> in the next fifty years is not an extrapolation that I would care to gamble
> on. Though I would honestly love for this claim to be true, my optimism
> wanes in the face of my realism. Show me good, hard, *tangible* evidence
> for your hypothesis and I will probably believe it as well.

Why does it matter (except to us) whether this happens in five years or five hundred? But anyway, we've *got* two-way computer-telepathy implants, collaborative filtering, enough 'Net-connected computing power to match the human brain, quantum computing, atomic-level manipulation, and neurological intelligence enhancement. Sure, most of these technologies are at the vacuum-tube stage - but so what? At this point, everything is just a matter of predictable miniaturization and known paths of research. No unexpected brilliancies are needed, although I'm sure there will be many - who'd heard of quantum computing a few years ago? Where was the Web when _Mind Children_ was written? Had I.J. Good heard of nanotechnology?

All these lines, projected into the future, intersect in and about 2025. If just one breakthrough is needed (quite plausible), we could get it by 2005. In 2025 we'll have a lot of converging, synergetic Singularity tech, making it the most arguable point.

By 2050 we'll be supersaturated. A thousand times human processing power on every desktop, home-neurosurgery kits, Parallel Scanning Tunneling Probe atomic desktop printers, thousand-qubit CPUs, a Net with a *quadrillion* times human power, million-neuron two-way interfaces... Even uploading would be possible at this point, given a Manhattan Project. This scenario is ludicrous. The Singularity occurs long, long before, or grey goo eats the world, or some paranoid blows up the planet.

-- 
        sentience@pobox.com         Eliezer S. Yudkowsky
         http://pobox.com/~sentience/AI_design.temp.html
          http://pobox.com/~sentience/sing_analysis.html
Disclaimer:  Unless otherwise specified, I'm not telling you
everything I think I know.