[ExI] The digital nature of brains

jameschoate at austin.rr.com jameschoate at austin.rr.com
Thu Feb 4 16:43:16 UTC 2010

This is a perfect example of my 'understanding inversion' claim...

First, we're not talking about different things. The Turing Test was suggested, not 'designed' as it's not a algorithm or mechanism. At best it's a heuristic. If you read Turing's papers and the period documentation the fundamental question is 'can the person tell the difference?'. If the answer is 'yes' the -pre-assumptive claim- is that some level of 'intelligence' has been reached in AI technology. Exactly what that level is, is never defined specifically by the original authors. The second and follow on generations of AI researchers have interpreted it to mean that AI has intelligence in the human sense. I would suggest, strongly, that this is a cultural 'taboo' that differentiates main stream from perceived cranks.

They way you flip the meaning of 'can the person tell the difference' to 'machine to convince' are specious and moot. The important point is the human not being able to tell the difference. You say it is not meant to test the ability of humans, but it is the humans who -must be convinced-.

I would say you're trying to massage the test to fit a preconceived cultural desire and not a real technical benchmark. It's about validating human emotion and not mechanical performance.

---- Ben Zaiboc <bbenzai at yahoo.com> wrote: 

> Well, we're talking about different things.  I said "it was designed to..", and you replied "no it does not".  Both of these can be true.
> The test was intended to test the abilities of a machine to convince a human, not to test the abilities of the human.  Of course that may well be one of its side effects! (apparently a disturbingly high proportion of people - mostly teenagers I think - are convinced by some chatbots)

 -- -- -- --
Venimus, Vidimus, Dolavimus

jameschoate at austin.rr.com
james.choate at g.austincc.edu
james.choate at twcable.com
h: 512-657-1279
w: 512-845-8989

Adapt, Adopt, Improvise
 -- -- -- --

More information about the extropy-chat mailing list