Re: Sysop Scenario FAQ

From: Gordon Worley (
Date: Mon Feb 04 2002 - 17:36:54 MST

On Monday, February 4, 2002, at 05:49 PM, Hooky Sun wrote:

>> From: Gordon Worley <>
>> Eep, I guess I've learned too much about writing from statistics (or
>> maybe mathematics in general). I try very hard to *not* reargue points
>> that have already been argued (successfully) elsewhere.
> I'm feeling you, but why write the FAQ at all then? To take
> Yudkowsky's point further present FAQ, answered with links, unless of
> course you can really shed new light.

Well, it's a matter of what `shedding new light' is. To me, if Friendly
SI has already been argued, and a Sysop is an SI, then the Sysop will be
Friendly because the SI decided that that was best. Or, maybe it will
be better than Friendly. At any rate, Eliezer is suggesting that I need
to go back and argue that the Sysop will be friendly, at least briefly.
This is reasonable since, to some if not most readers, the Sysop is a
new context in which to argue Friendliness and so I must at least
briefly argue it, even if I just reapply `obvious' facts of Friendly AI
to the Sysop, to convince them that the Sysop is Friendly.

As to why write an FAQ. Well, there really are frequently asked
questions that need to be prevented from being rehashed and an FAQ is an
effective way of doing this.

Gordon Worley                     `When I use a word,' Humpty Dumpty            said, `it means just what I choose                it to mean--neither more nor less.'
PGP:  0xBBD3B003                                  --Lewis Carroll

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT