[p2p-research] Drone hacking

[J. Andrew Rogers]

On Thu, Dec 24, 2009 at 2:30 AM, Tere Vadén <tere.vaden at uta.fi> wrote:
> J. Andrew Rogers wrote:
>>
>> There is no strong theoretical reason to believe they are genuinely
>> random, so Occam's Razor would favor the "deterministic but not
>> measurably so in this universe" hypothesis.
>
> No? What about Bohr contra EPR? At least Einstein found that argument sound,
> even though he tried to find holes in it for years. In fact, some people
> have claimed that Bohr's argument for complementarity is one of the most
> beautiful & precise arguments in the history of natural science. Some of
> those people are mathematicians.


Quantum mechanics was largely invented in the 1920s. The basic
mathematics of predictability was largely invented in the 1960s.
Obviously they used the tools available at the time, and in the
absence of meaningful distinction continued using the assumption that
had been in use for two generations when an alternative materialized.


> Oh but it does. The whole point of Bohr's argument is that quantum
> randomness is of a different kind than classical ("deterministic but not
> measurably so in this universe") randomness. Quantum randomness is not an
> effect of dependence of initial conditions, but a physical "brute" fact.


Non sequitur. They are indistinguishable cases. The only difference is
that strong unpredictability allows (in the mathematics) that a
context can exist in theory in which the algorithm is measurably
deterministic.  In the case where no such context can be available to
us in theory, there is no difference in practice. A physicist should
not care one way or the other.


> Therefore whatever your interpretation of quantum mechanics (the mathematics
> of it), you are bound to have something non-classical in nature
> (complementarity, incompleteness, uncertainty, many-worlds, or what have
> you).


I think you might be conflating the specific case of simple *local*
determinism in physics with generalized mathematical determinism. We
know the former case is not a valid model, but I've never heard a
physicist assert in discussions specifically about this that the
latter model is invalid. The generalized model of determinism has far
weirder manifestations than anything we can easily conceptualize.


> What non-local deterministic model do you have in mind? Bohm's?


I have no particular model in mind (from the standpoint of physics it
doesn't matter), but I do understand the advantages of the
deterministic model from the mathematical standpoint.

The case of strong unpredictability is a trivial extension of existing
axioms and assumptions already used in physics and is mathematically
equivalent to "random" in all cases -- all upside, no downside. You
just have to accept that it is a determinism that sounds weird.
Asserting true mathematical randomness creates a new construct that
has very different (and greatly complexifying) properties relative to
the existing constructs in physics. It is a very ugly hack. ("random"
is generally a very murky concept in math, a portmanteau for anything
that is unpredictable in some context.)

That is where Occam's Razor comes in. If we can trivially extend the
existing assumptions without any loss of utility, why assert an
equivalent alternative that throws a monkey wrench into the clean
mathematics of the model?


> So the claim
> about full deterministic predictability is exactly the kind of unnecessary
> complication that Occam is talking about. We can do science, more simple
> science, without the claim.


Determinism does not imply predictability in an arbitrary context. The
relationship is not that simple.

We have three real-world facts to work from:

1.) Behavior is robustly predictable in a way that mathematically
implies significant determinism.
2.) The non-predictable residue can be either deterministic or not,
since there is no meaningful distinction.
3.) Future machines may further shrink the size of the non-predictable
residue, but we won't know until we try.


-- 
J. Andrew Rogers
realityminer.blogspot.com