In the discussions (rabid rants) in favor of mind uploading, the product
of the upload is described as a personality, that is asserted to be
yours, by virtue of functioning just like yours. This is assenine but
the idea of a human-analog piece of software serves to further clarify
what is meant by referring to an AI.
So you have AI in general, (as apposed to general AI).
Then within that, general AI systems that could include mainframe type
systems similar to IBM's Watson (in general shape and form).
And finally, within that category we have systems that specifically
mimic the human mind and operate more or less like a human does.
I find this last type most compelling for a number of reasons, even if
it were networked with and served as an avatar for a more alien type of
AGI, or a human user via a neural interface.
Furthermore, the abandonment of trying to emulate the "identity" of a
specific human being is tremendously liberating because it permits a
much higher level of abstraction in the code, yielding up to a 1,000x
performance boost as well as eliminate all nececessity to emulate
specific brain quirks either common to humans or specific to the
individual being emulated and thus achieve even more performance.
There is an organization in Japan that appears to be actively pursuing
this line of research:
https://wba-initiative.org/en/about/
If you know of others, please share the links.
--
Please report bounces from this address to a...@numentics.com
Powers are not rights.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink:
https://agi.topicbox.com/groups/agi/Tbc1510191f9871a5-M2de6c9f814f17ab7f1840c7d
Delivery options: https://agi.topicbox.com/groups/agi/subscription