From: Arona Ndiaye (andiaye@chello.nl)
Date: Wed Jan 23 2002 - 20:46:02 MST
>>If there is no easier way to build an AI (easier
>>defined by the amount of computer hardware required
>>to implement it), then I expect a sufficiently large
>>neural network will eventually work. This is, in
>>fact, how I personally estimate the time to the
>>singularity.
I do not agree. Danny Hillis with his 'connection machines', offered computer
hardware several orders of magnitude faster than anything else existing at
that time. Great: no one knew how to code for the damned things. The
breakthrough will be at our level. A very fast, large (whatever) machine, is
just that. Someone still needs to come up with code to run on it.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT