Re: Riddles (was: Lateral Thinking)

From: Dale Johnstone (
Date: Tue Jan 02 2001 - 09:49:27 MST

----- Original Message -----
From: "Damien Broderick" <>
To: <>
Sent: Tuesday, January 02, 2001 3:00 AM
Subject: Re: Riddles (was: Lateral Thinking)

> At 05:14 PM 1/01/01 -0800, Jason Joel Thompson wrote:
> >> >The riddle itself is a Hofstadterian self-answering riddle.
> >> How... disappointingly... inadequate. So much build up, so little pay
> >> I need to see this alleged lateral wonder unpacked.
> >I have to agree, and to be somewhat immodestly frank, I think I might
> >to claim that my response was the better answer.

I liked your answer too.

> Jason's response was wonderfully clever, but still left out some key
> elements and finessed others. I think I should elaborate a little on why I
> felt cheated by Eliezer's and Dale's meta-level `revelation'.
> Purported koans of this kind are rather too close for my tastes to that
> hoary deadener; `then she woke up, and it had all been a dream!' Well, no
> shit. I discussed some of the implications of apparently self-subverting


I can understand your disappointment. I would have been very disappointed if
it had turned out to be some sort of hypercube. Posing problems in >3
spacial dimensions doesn't *seem* fair.

It really stems from your expectations of what a puzzle should be. Had the
puzzle been posted in alt.maths.topology a hyperspacial solution would have
been more appropriate.

I see from your own stab at the puzzle you took a very literary angle. This
is hardly surprising considering you're an author.

I think the point of Eliezer's puzzle was to challenge our assumptions and
force us to think outside the box ('box' probably being another clue). The
box this time being the *form* of the riddle-answer pair.

You assumed it would have an answer like every other, probably of the
juxtaposition-of-letters form. Personally I never really enjoyed those
types, I find them disappointing and inadequate. Sound familiar?

Trying to design a machine that can think outside these solution patterns is
a central issue to real AI. Existing attempts at AI are notoriously brittle,
and suffer from inflexibility due to rigid classification schemes.

The only reason I solved it was because I've seen this meta-pattern before,
and it's no coincidence I also work on AI as Eli' does. Having half a dozen
goes helps too! :)

Have a go at the puzzle I posted. If it makes you feel any better (and at
the risk of spoiling it) I guarantee it has a spacial answer, although it
might still challenge your assumptions.

Dale Johnstone.
AI Researcher.

This archive was generated by hypermail 2b30 : Mon May 28 2001 - 09:56:16 MDT