RE: Why FAI Theory is both Necessary and Hard (was Re: SIAI's flawed friendliness analysis)

From: Christian Rovner (crovner@lacasilla.com.ar)
Date: Sun May 11 2003 - 16:59:42 MDT


>===== Original Message From sl4@sl4.org =====
>Are we saying the intelligence required to create an FAI is greater than
>the intelligence required for the Manhatten project? Because we all
>know what that ended up with!
>
>It's not that I don't trust the people now doing the work, it's just
>that if it's publicized and demonstrated as doable, what's to keep
>parties with less honorable motives from coming in and taking over the
>project?
>
>The only way I see to protect the FAI from exploitation until it has
>achieved near omnipotence is to conceal it's progress and mask it's
>achievements by crediting them to human frontmen.
>
>
>
>-----Original Message-----
>From: owner-sl4@sl4.org [mailto:owner-sl4@sl4.org] On Behalf Of
>Christian Rovner
>Sent: Sunday, May 11, 2003 1:26 PM
>To: sl4@sl4.org
>Subject: RE: Why FAI Theory is both Necessary and Hard (was Re: SIAI's
>flawed friendliness analysis)
>
>
>
>Eliezer wrote:
>>
>> I find it hard to believe that any human truly capable of learning and
>
>> understanding that art would use it to do something so small and mean.
>
>Cliff wrote:
>>
>> You "find it hard to believe"? _I_ find it hard to believe you would
>> use such a phrase casually or unintentionally, without awareness of
>> the implications. *Especially* when such a statement is made about a
>> probability of the form P(small & mean | capable)...
>
>Benw wrote:
>>
>> I agree, Cliff.
>>
>> Human history is full of twisted atrocities that are "hard to believe"
>
>> -- except that they actually happened.
>
>Apparently, the kind of wisdom required to create Friendly AI is way
>greater
>(and includes) the kind of wisdom required to realize the smallness and
>meannes of seeking political power. If someone failed to achieve that
>sort of
>understanding of human morality, alas, they will hardly be able to
>design and
>raise a Friendly AI.
>
>-- Christian Rovner
>Singularity Institute for Artificial Intelligence, http://intelligence.org/

-- Christian Rovner
Singularity Institute for Artificial Intelligence, http://intelligence.org/



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:42 MDT