[p2p-research] Drone hacking

[J. Andrew Rogers]

On Tue, Dec 22, 2009 at 2:39 AM, Michel Bauwens <michelsub2004 at gmail.com> wrote:
> to prove mathematically that humans have no free will, come on, that must be
> the worst possible category error ...


The category error is in thinking that "free will" is a binary state
or non-contextual.  There are limits to how grossly we can simplify
what the mathematics actually says here.  If a computer can predict
every single decision you make throughout the day with 85% (random
number) accuracy, you are not predictable in a deterministic sense --
no one ever will be -- but you are predictable for all practical
purposes.

One can trivially derive a working concept of "free will" from the
Invariance Theorem in information theory: no machine can perfectly
predict the behavior of a machine of equal or greater size in *any*
case, including itself. In more restricted cases, there are algorithms
that are opaque to induction by vastly larger machines -- encryption
is based on this. In other words, the mathematics requires that we see
ourselves as unpredictable even if there is a context in which we
*are* predictable.

Behavioral prediction is not a binary thing. The prediction error,
assuming the Solomonoff ideal, is a function of relative machine size
and certain machine characteristics. Any prediction error that is less
than chance says something about the complexity of the machinery being
predicted and in many cases is exploitable.


-- 
J. Andrew Rogers
realityminer.blogspot.com