RE: SIAI's flawed friendliness analysis

From: Rafal Smigrodzki (
Date: Thu May 29 2003 - 17:46:34 MDT

Eliezer wrote:
> Rafal Smigrodzki wrote:
>> This uncertainty is due to our general lack of knowledge about how to
>> build an FAI - you cannot rationally regulate a non-existent field of
>> research, because all of the "experts'" guesses are as good as mine
>> or yours at this point.
> Not really. There is no way for the government to tell the
> difference; that is not the same as saying no difference exists.

### Could experts today (whether governmental or not) provide anything but
the most general guesses? I agree that the government, due to the very small
number of FAI-interested persons and lack of experience in finding such
persons, cannot reliably select and retain the people needed to tell the
difference. But isn't the situation even more premature? Are there any
persons who could honestly and without self-deception say they can reliably
tell good FAI projects from bad ones? (aside from excluding obviously flawed
ideas, like evolutionary programming, or happiness as reinforcement). This
is not to belittle anybody's ideas, but the field is still a pretty early
stage, especially as Friendliness content goes.


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