[ExI] Conspiracy epistemology (Was: cut off worried)

[Anders Sandberg]

On 2016-02-16 04:38, spike wrote:
> Camping in Death valley no Internet no radio email intermittent. 
>  Hearing wild stories from other campers pls refute oh pls do.

The best one I have heard is that it was Leonard Nimoy, ruler of the 
Illuminati (he faked his death, you know) who ordered it. Because of 
evil ObamaCare reasons. That story was made up by a spoof right-wing 
site, but then a lot of people started to run with it (either to prove 
the other side are nuts, or actually supporting a variant of the theory).

More seriously, after reading Tetlock's work, I have reduced my estimate 
of the likelihood of clever political conspiracies. I have no doubt some 
people *try*, but given that policy experts are lousy at predicting 
consequences of events any plan involving steps like "And then X will 
appoint Y, who will implement plan Z, with no trace back to me. 
Excellent!" is very likely to fail. When even superforecasters selected 
for their objective accuracy cannot do reliable predictions very far in 
the future it seems unlikely conspirators are any better. Especially in 
multi-agent situations like the current US election, or situations where 
random events can blow up (like in many parts of the middle east).

A philosopher colleague pointed out that it is a strange world where 
some of the best work in epistemology is being done by the US 
intelligence community (especially IARPA). The really crazy part is that 
they started only a few years ago - before that people generally thought 
they knew what they were doing prediction-wise. I think that 
overconfidence in how good we are at predicting consequences is driving 
both conspiracy theorists (who take comfort in that somebody is running 
things, even though they are evil) and would-be political manipulators 
(who churn the mess with their attempts).

So if you want to run a "conspiracy", focus on actions that have 
predictable effects - physical effects, things going according to the 
normal legal, bureaucratic or traditional routine. Don't rely on outside 
people acting as you want, don't act in domains where outside events can 
overwhelm your influence. Shanteau's theory of expertise gives great 
input to what skills you need to affect things. Information leakage and 
side agendas grows nonlinearly with group size. And so on. It is much 
easier to do a non-conspiracy where you get people with shared interests 
to coordinate, pool resources, and push for changes they want to see 
without all the cloak-and-dagger stuff. Non-conspiracies are also more 
agile in areas where predictability is lesser, but will of course have 
to deal more with opposition. Conspiracies in theory avoid opposition by 
being secret, but in practice they have to deal with it anyway: the 
opposing interests will be acting in unpredictable ways in any case, 
potentially wrecking plans with no idea what they were about.


(I always smile inwardly when I cash checks from Steve Jackson Games at 
my local bank, since the checks have the company's eye-in-the-pyramid 
logo prominently displayed. I am getting paid - tiny amounts - by the 
Illuminati! Fnord.)

-- 
Dr Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20160217/69057b83/attachment.html>