Re: [sl4] Is there a model for RSI?

From: Bryan Bishop (
Date: Tue Jun 24 2008 - 18:36:01 MDT

On Wednesday 18 June 2008, Krekoski Ross wrote:
> Its also not altogether clear that the Turing test is even
> appropriate in relation to RSI. We might devise algorithms for RSI to
> do better at a Turing test, but that may not necessarily yield
> intelligence.  Its like the analogy of a Turing test for flight if
> the entity in question cannot be distinguished from a bird. But
> chickens dont fly.

I'm not sure that RSI can be tested with an external test of that sort
because of the nature of the SI aspect which suggests improvement of
the self, which perhaps includes certain aspects that others don't
understand or will not value as much. I'm not saying it's impossible,
but rather that the direction of the SI aspect is going to have to be
something that you (the reader) selects of the batches that turn up.
Separating signal from noise and such. But some of that noise can be
useful in other ways. (junk DNA)

- Bryan

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:03 MDT