Re: MNT-enabled tyrannies

From: Aleksei Riikonen (
Date: Tue Jun 21 2005 - 05:35:41 MDT

Thomas Buckner wrote:
> --- Phillip Huggan <> wrote:
>>under present realities, an
>>AI of less than assured friendliness might be
>>less risky than facing the run-up to MNT;
>>turning on a "buggy" GAI system might be the
>>only defence against an emergent MNT enabled
> Although I somewhat agree with this, I suspect
> Eliezer would not; his position is that imperfect
> GAI is far, far more dangerous than almost anyone
> realizes.

An emergent MNT-enabled tyranny would in most cases relatively quickly
stumble upon a GAI of some sort (MNT enables that too). In many cases,
it would therefore be a justified last-resort defense to activate an
imperfect GAI system, provided that it would be significantly less buggy
than the one that the tyrannical regime was about to stumble upon.

Aleksei Riikonen -

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT