- From: Harry Halpin <hhalpin@ibiblio.org>
- Date: Thu, 4 Aug 2011 11:15:09 -0400
- To: jeff putnam <jayfoo@gmail.com>
- Cc: Melvin Carvalho <melvincarvalho@gmail.com>, Semantic Web <semantic-web@w3.org>
On Wed, Aug 3, 2011 at 11:22 AM, jeff putnam <jayfoo@gmail.com> wrote: > Isn't it possible that the internet is already sentient (it certainly > has lots of power, interconnections and the like). > > The interesting question is more likely to be "How would we know?" > It is not likely to exhibit human style intelligence. Do we have > any way to measure sentience (or intelligence) in things that are > wildly non-human? > No, because we lack a good definition of intelligence. The Turing Test just measures "human-like" intelligence. There is a field of philosophy trying to figure this out, but results look dim so far: http://en.wikipedia.org/wiki/Artificial_consciousness In essence, as regards the precise argument put forward on that video, neural networks are in general a bad model of the human brain (essentially throw enough nodes on and one can approximate any non-linear function, nothing particularly human or even interesting about that), and thus, there is no reason to believe that anything as simple as adding "weights+backpropagation" to - god help us - the Semantic Web would in anyway produce consciousness. Also, Dennett's view of consciousness is that anything with multiple competing models qualifies as consciousness, but as Chalmers correctly points out, that's kinda crazy as it doesn't take into account the emergence of qualia. I'd look at work by Thomas Metzinger for a better story about consciousness and subjectivity and for upcoming work from Andy Clark on a better model of human neural activity than neural networks from the 80s... cheers, harry > > On Tue, Aug 2, 2011 at 1:59 PM, Melvin Carvalho > <melvincarvalho@gmail.com> wrote: >> According to this TED Talk >> >> http://www.youtube.com/watch?v=RGF9NbPFfRo&t=10m34s >> >> Enjoy! :) >> >> > > >
Received on Thursday, 4 August 2011 15:15:46 UTC