[ExI] Oxford scientists edge toward quantum PC with 10b qubits.

[spike]

-----Original Message-----
From: extropy-chat-bounces at lists.extropy.org
[mailto:extropy-chat-bounces at lists.extropy.org] On Behalf Of Adrian Tymes
Sent: Thursday, February 03, 2011 3:24 PM
To: ExI chat list
Subject: Re: [ExI] Oxford scientists edge toward quantum PC with 10b qubits.

On Thu, Feb 3, 2011 at 2:04 PM, spike <spike66 at att.net> wrote:

>> Tried that: started with an insect level AI, scaled it up.  Ended up with
a simulation of a huge pile of bugs.

>So simulate yourself fixing bugs.  ;)

Tried that: didn't work either.  The huge pile of simulated bugs were
smarter than me.  They first devoured my avatar.  Then they devoured each
other.

>From that failed exercise, I figured out the way to go however.  Instead of
starting with insect level AI and scale it up, I would start out with a
me-level AI and scale that up.  Reason: I am not all that great a coder.  I
can do it, but I suck.  In debugging code, I am not all that far above
insect level AI.

It's a challenge.  I am really good at writing bugs, but haven't yet figured
out how to write a software simulation of my intelligence.  If I ever do, I
will write an AI simulated spike, then have it rewrite itself better, then
have that new simulated spike do all the work.  While it is at that, I will
have it write a new sim-spike to have fun watching the other sim-spike work.

I did learn something else interesting.  If I attempt to write a really
simple minded routine, such as a prime number generator, I can write that
code without any bugs in it.  But if I write something complicated, such as
my latest digital guidance and control scheme, that routine is full of bugs.
So now my strategy is this: instead of writing a simple insect-level AI, I
will write a really complicated sophisticated transpike or spike+ algorithm,
even if it has lotsa bugs.  Then I will make it debug itself.  When it is
finished debugging itself, I will make it scale itself up.

spike