[ExI] How could you ever support an AGI?

Bryan Bishop kanzure at gmail.com
Mon Mar 3 12:24:20 UTC 2008


On Monday 03 March 2008, Stefano Vaj wrote:
> What is going to happen is that from machines that can pass the
> Turing test over 10 questions 1% of the times with 1% of the users
> (we are probably already there) we will get to machines that can do
> that over 100 questions 50% of the times with 50% of the users to
> those who score better than real humans on any finite number of
> interactions with any finite number of users.
>
> As for the obsolescence of Homo sapiens, well, biology is not static

While reading this, I began to think: why not automate the Turing tests? 
Let's set up a cloning operation so that we can characterize both the 
human and ai.

- Bryan
________________________________________
Bryan Bishop
http://heybryan.org/



More information about the extropy-chat mailing list