Re: How to make a slave

From: Harry Chesley (chesley@acm.org)
Date: Fri Nov 23 2007 - 16:22:58 MST


John K Clark wrote:
> On Fri, 23 Nov 2007 "Harry Chesley" <chesley@acm.org>
>
>> Goals will necessarily come from one of two places: pre-programmed
>> or random.
>
> Or goals evolved from ancestral goals in ways that are deterministic
> but imposable to predict. As I said before I can write a very very
> short program that will do things nobody can predict, that is why I
> get impatient when somebody says we can just tell a Jupiter Brain
> "always be a slave to human beings" and we can be certain he will be
> exactly that slave from now to the end of time.
>
>> So, are you arguing that it's impossible to eliminate the presence
>> of random top-level goals? Or that it's impossible to predict the
>> actual result of pre-programmed top-level goals? Or that new
>> non-random top-level goals will get introduced from thin air?
>
> My top-level goal will not allow me to respond to the above because
> it exceeds your quota of rhetorical questions. Please rephrase and
> resubmit.

Not rhetorical at all. I believe your answer from the previous paragraph
maps to "it's impossible to predict the actual result of pre-programmed
top-level goals." The issue here seems to be the fine line between true
randomness and chaotic systems. This is quite a valid concern. But you
seem to be going so far as to say that all GAIs are necessarily chaotic.
I wouldn't go that far myself.

>> BTW, the condescending tone of many of your posts does not further
>> your arguments.
>
> That could be but it can't be helped. The truth is I have no talent
> for pretending an idea is clever when in fact I don't believe it is.

I would not ask you to. Only to be polite.

Personally, I try to be very tolerant of fools because I myself am one
so often.



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:00 MDT