Re: FAI prioritization

From: Stuart Armstrong (
Date: Tue Apr 08 2008 - 07:17:59 MDT

> The fact that currently science lacks understanding of how to
> construct it, doesn't mean that it's going to be difficult from the
> engineering standpoint -- so 'nowhere near' is unhelpful.

When we have no idea of what we want to do, I'd say that "nowhere
near" is a good educated guess. When you stumble along in the dark,
the odds are much greater that you're "nowhere near" your goal, rather
than just next to it.

And yes, penicillin is a counter-example :-) But those discoveries are
quite rare, even if they're very vivid.

But the probability of AI happening soon is not what's relevant to me;
I feel that probability is low, but that the mean expected changes due
to AI are very high, so this is an important field.


This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT