Re: Artificial Intelligence will Kill our Grandchildren

From: justin corwin (outlawpoet@gmail.com)
Date: Sat Jun 14 2008 - 17:33:55 MDT


Ok, I'll bite.

On Sat, Jun 14, 2008 at 2:50 PM, john ellis <john_f_ellis@yahoo.co.uk> wrote:
> You'd better get a move on. Roadrunner's ability to program the entire visual cortex, which is 50% of the intelligent part of the human
>brain does, exists now.

What do you know about the PetaVision software they talk about? It
struck me as fundamentally similar to the Blue Brain project, and
other proported neural sims, interesting, but not implying any serious
analysis or new functional understanding, at least until they produce
some data which we can then compare to some model.

> Human intelligence is at most 2-3 years away in machines by this route, and post human intelligence by IJ Good's dictum, the day after that.

Do you have any factors you believe add up to that number? Does that
mean simply running a simulated human?

> There is but one way to guarantee safety now, which is to build containable AGI first and neutralise all other attempts as humanely as
>possible, which luckily for me I'm working on and expect to complete before Roadrunner does.

Are you affiliated with a AI project I'm not familiar with, or is this
a solo/stealth effort?

-- 
Justin Corwin
outlawpoet@hell.com
http://outlawpoet.blogspot.com
http://www.adaptiveai.com


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