pharos at gmail.com
Mon Sep 29 19:51:06 UTC 2014
On Mon, Sep 29, 2014 at 7:06 PM, Adrian Tymes wrote:
> Indeed, knowing what the program is to do is often the problem.
> We want it, ultimately, to be at least human equivalent and probably
> superior. Figuring out what that means, in terms of stuff that can be
> directly coded, is a large part of the challenge.
That's why some people speculate that the internet might eventually
appear to be 'intelligent'.
Not by itself, but by linking together other AI machines, like Watson,
Wikipedia, Wolfram, Google, etc.
Just as humans actually 'know' less and less, but appear to know more
than ever because they can look it up in seconds on the internet.
More information about the extropy-chat