Re: [sl4] FAI development within academia.

From: Roko Mijic (rmijic@googlemail.com)
Date: Wed Feb 25 2009 - 07:28:58 MST


2009/2/25 William Pearson <wil.pearson@gmail.com>:
> 2009/2/25 Roko Mijic <rmijic@googlemail.com>:
>> Since this thread seems to have been somewhat distracted, I might just
>> bring us back to the original more immediately salient question.
>> Perhaps we could Matt's quantitative theory to another thread?
>>
>> "I want to help out with FAI research, what do I do?"
>
> There used to be a FAQ. "So you want to be a seed AI programmer," I
> suspect it is deprecated.
>
> My guess is you would be out of luck trying to do FAI/AGI in academia.
> You might try attaching yourself to something like SOAR for a PhD and
> then striking out on your own from there. Aaron Sloman at Birmingham
> Uni (UK) is somewhat more open minded about AI approaches IIRC.

Indeed, though he told me that he doesn't think that there is such a
thing as general intelligence....

>
> The other way to be helpful is probably advancing the state of the art
> of decision theory so that it is more realistic/more powerful or
> otherwise better suited for reasoning about FAIs with. Just my 2 penny
> worth, I am not in any way affiliated with FAI work, I'm one of those
> crazies that thinks AGIs won't go foom.

Well, this isn't a such a crazy position to take, but it is enough
that there is a significant probability that "general" AI will behave
in a volatile way to make it worth doing research on FAI/AGI. Think of
FAI researchers as insurance against the possibility that AGI will "go
foom".

>
> Will
>

-- 
Roko Mijic
MSc by Research
University of Edinburgh


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:04 MDT