Re: A conversation on Friendliness [HUMOUR]

From: Michael Roy Ames (
Date: Thu May 27 2004 - 11:42:40 MDT

Damien Broderick,

You wrote:
> As an exercise, examine the semantic content
> of Eliezer's interlocuter's contributions:
> [snip]
> How is this distinguishable from the responses
> of an Eliza-bot?

As a real question, the answer is: Because the interlocuter's contributions,
when considered as responses to Eliezer's contributions, where on-topic and
appropriate. Much of common conversation between two (or more) people when
split up into its separate contributions is mundane. This is one of the
reasons why people find bots fun to talk to **at all**.

As a joke, your Elizer-bot question made me smile. Here's a reformulation
of the content (with appologies to James)...

> <Jester> Any luck with that new chocolate cake recipe?
> <James> Nothing terribly communicable. I am wondering if a
> correct implementation of initial state is even generally
> decidable.
> <James> I'm having a hard time seeing a way that one can
> make an implementation that is provably safe.
> <James> At least generally.
> <Jester> Yeah, that is always a problem.
> <Jester> If it tastes good then your tempted to eat it all,
> which if continued leads inevitably to obesity, congestive
> heart failure and death.
> <Jester> If it tastes bad then no-one wants to eat it and
> you have to toss the damn thing, which if continued leads
> inevitably to a rotting pile of cake, attracting infestation
> and disease, then death.
> <James> There might be a shortcut for narrow cases.
> <Jester> Could be. There might be a class of cake that is
> tasty, but not too tasty, so that you only eat a moderate
> amount - and give the rest away to your friends.
> <James> right
> <James> but how do you prove that the invariant constrains
> expression correctly in all cases?
> <James> sorry. stays friendly
> <Jester> The correct balance of cocoa vs. sugar *is*
> difficult to get right, and even to measure correctly!
> <James> at runtime
> <Jester> Oh, that is an additional problem. Running while
> eating can often produce nausea.
> <James> Right
> <Jester> Leading one to spew half-digested chocolate cake
> all over one's friends - thus largely defeating the point.
> <James> precisely.
> <James> that's the problem
> <Jester> Its hard to convince even the friendliest person
> to eat cake after that. :/
> <James> heh
> <James> I've been studying it, from a kind of theoretical implementation
> standpoint. Very ugly problem
> <James> Yeah. Any thoughts on the how to remove chocolate
> puke stains?
> <James> No thoughts yet.
> <Jester> Well, I suppose we could always ask Eli - I think
> he knows the answer.
> <James> Yep. Always have to be aware of that

and so on...

Michael Roy Ames

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT