Re: Many-worlds (was Re: [sl4] Re: Uploads coming first would be good, right?)

From: Mike Dougherty (
Date: Sun Mar 08 2009 - 13:06:00 MDT

On Sun, Mar 8, 2009 at 1:00 PM, Matt Mahoney <> wrote:

> The human brain computes probability based on a model that assumes that
> memories accurately reflect the past and that the past predicts the future.
> Suppose you are given a choice of (A) being tortured, and then having the
> memory of the torture erased, or (B) not being tortured, but then having a
> false memory of torture programmed into your brain. Our model says that
> after many trials you will consistently choose (A) even if you prefer not to
> be tortured.

I assume A also involves ensuring evidence is erased along with memory, such
that scar and broken limbs have been healed? If not, I think I'd prefer B -
since I would probably divorce my identity from that tortured self in much
the same way I've divorced my current self from the punk I was at age 19. A
inexplicable scar or missing limb would be difficult to understand without
the associated memory.

I think this example only furthers the point Matt was making about memory,
identity and probability.

Consider also your poker strategy against 1) unknown players vs. 2)
well-known players. By observation of #1, you hope to observe tendencies
similar to those of #2 - so you may profitably predict your opponent's
reaction to your own betting and bluffing behaviors. A powerfully adaptive
intelligence would make a formidable Texas Hold'em player: both by reading
'tells' and presenting false 'tells' of its own.

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