Re: My attempt at a general technical definition of 'Friendliness'

From: Marc Geddes (
Date: Fri Jan 21 2005 - 19:52:22 MST

 --- Eliezer Yudkowsky <> wrote:

> Complete trash. Either Geddes didn't read "A
> Technical Explanation of
> Technical Explanation", or he missed the point so
> completely that he might
> as well not have.

No so fast there!

Firstly did you read the comments on the
philosophy/science debate? O.K, I admit I shouldn't
have used the word 'technical' for much of what I
said. But I am attempting to begin with the most
general (philosophy) and gradually move towards the
precise (science). I know you are have a certain
animosity towards philosophy but remember...all
science started with philosophy. Should one have
trashed the alchemists when they were doing alchemy?
Then chemistry would never have gotten all the ground.

> > If you gave it to him using a term other than
> > "friendliness", would he even recognize it as
> describing the same thing
> > that he frequently talks about?
> No.

At this point *no one* can claim to fully understand
'Friendliness' (or else they would have built an AGI).

Are you are sure you read what I actually said?

Here was the technical part: I defined friendly
sentients as functions (computations) which take in
particular kinds other functions as input and modify
them in a certain way. I said that the 'particular
kind' of functions operated on were approxmiations to
certain uncomputable functions. The 'certain way'
they are modified is that a poor approximation
function is taken in as input, and then a better
approximation function is given as output.

Admittedly very general, but why is it 'trash'?
Excuse me, but the paragrah I have given above makes
precise technical (mathematical) sense.


Find local movie times and trailers on Yahoo! Movies.

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:50 MDT