Re: [agi] What best evidence for fast AI?

From: Joshua Fox (joshua@joshuafox.com)
Date: Sat Nov 10 2007 - 09:12:23 MST


> So I am here to ask: where are the best analyses arguing the case for
> rapid (non-emulation) AI progress?

Some answers examine the progress of engineering and suggest that AGI will
probably come at a certain date -- soon -- just as the telephone and the
airplane seemed fated to emerge when they did: there were multiple inventors
racing for the finish line.

There is some justice in that, but another approach is this: All that
remains on the road to AGI is figuring out how intelligence works. Some
researcher(s) will do so. We don't know when, but it is fundamentally a
mathematical/philosophical advance rather than a nitty-gritty understanding
of the quirks of a biological systems. The implementation is just a Simple
Matter of Programming <http://www.catb.org/jargon/html/S/SMOP.html>.

The understanding could be achieved by one genius with a brain-flash, or by
a community making progress together. We can compare the breakthroughs of
Wiles and Perlman, which depended on a series of mathematical developments,
yet could not be predicted like the progress of some engineering project
(not that those are too predictable).

This is different from, e.g., reaching Alpha Centauri, which would require
solving more than one engineering problem (despite the good work of
pioneering researchers who have laid out advance plans) and enormous
economic resources. It is also different from time travel or teleportation
which would require re-defining laws of physics (despite some speculative
thought in these directions). Indeed, I'd say that it's different from life
extension, which requires understanding multiple biological systems and
engineering a well-balanced fix to certain problems.

Joshua



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:00 MDT