[p2p-research] Drone hacking

[J. Andrew Rogers]

On Tue, Dec 22, 2009 at 10:43 AM, Andy Robinson <ldxar1 at gmail.com> wrote:
> So now we're getting somewhere. You've conceded the core point that people
> aren't really predictable at all, just that certain guesses of how people
> behave based on their past behaviour are right a certain proportion of the
> time - hence reducing unpredictability to a statistical fraction of cases.


I haven't conceded anything, it was my assertion from the very
beginning.  It is obvious at this point that many here are lacking the
context to see the implications. This is not statistics in the usual
sense, it is high-order algorithmic probability. Intuition will fail
you here.


> Which reminds me a lot of the models of risk management which have
> repeatedly caused economic crises in the last 20 years.


No it doesn't. This kind of math did not exist 20 years ago.


> So how exactly do you know that these packages buy a 10-30% sales increase,
> or indeed any sales increase at all?  Would you care to tell us where this
> information comes from - which websites are using these packages?  Is any of
> this information in the public domain?


I am have a lot of theoretical expertise related to this, so I am
often approached by implementors about specific issues. As a side
benefit I sometimes get insight into current performance metrics.
Always good to keep abreast of the state-of-the-art in implementation.


> Or are you working for the people who are marketing these packages, hence in
> on the secrets?  In which case I would seriously question your motives in
> trying to make them out to be as massively world-changing as you claim.


I have no involvement or vested interest in that business. Again, I
don't see the claims as remarkable since they follow from
well-understood math.

When you think about it, it is just marketing distilled to its purest essence.

-- 
J. Andrew Rogers
realityminer.blogspot.com