Re: What is Intelligence?

Robin Hanson (
Mon, 14 Oct 96 11:03:55 PDT

Dan Clemmensen writes:
>> We have had half a century of experience with computers, and no
>> similar takeoff has occured, or anything close. We have had feedback,
>> lots of it, but growth rates have been relatively steady. I think you
>> need a good story about why your scenario hasn't yet happened to make
>> it at all plausible.
> Recently, the computing power accessible to my nominal
>researcher/hacker has been increasing a lot faster than that as
>machines have joined the internet. ... My scenario has apparently not
>yet happened (as of last week) because the critical mass has not yet
>been reached ... Computing power is not necessarily the only critical
>element. It's possible that the final missing link will be a
>particular piece of software such as an information-visualization
>package, or a decision-support package, or a knowledge database. The
>point is that the internet may then enter a super-critical condition
>in which a single seed may precipitate a phase change.

Machines have been joining the internet for a long time, without any
visible takeoff resulting. So you must have in mind some critical
number beyond which your imagined effect happens. And visualization,
decision-support, KB, etc software has been around for awhile, so you
must be imagining some special effect of some as yet unseen aspect of
such programs. Both of these seem like wishful thinking to me.

Robin D. Hanson