25 May 2015.
25 May 2015
What should GCHQ do? Part 2
From: "t byfield" <tbyfield[at]panix.com>
Date: Sun, 24 May 2015 22:09:00 -0400
Subject: Re: <nettime> What should GCHQ do?
On 24 May 2015, at 7:09, William Waites wrote:
And so we have arrived at the economic problem. The business model of advertising
has the same basic requirements as mass surveillance. Thwarting one by
decentralisation and ensuring confidentiality of communications means thwarting
the other. Improving safety and security by encouraging pervasive encryption
means finding a new economic model for the Internet that does not depend
on surveillance, that transcends the Web2.0 model of capturing users in silos.
Surely this too can be a fruitful direction for research.
The ethos of our time is something like *I can therefore I will.* If something
technically possible, it becomes imperative to do it. It's not an individual,
philosophical imperative, something that you or I 'must' do -- that wouldn't
really matter. It's a systemic, probabilistic imperative, something that
someone else will end up doing sooner or later -- which matters very much.
If something is possible, it seems inevitable. 'Technology' is the field
where this dialectic plays out: whatever gets drawn into those dialectics
becomes 'technology.' I think this helps to explain why our time is dominated
by engineering and law (note: not 'business' or 'finance'), two disparate
professions that are united by one strange feature: both are organized around
limit cases. They operate through ultraist logic, extrapolating everything
to its logical extreme and proceeding on that basis.
Normally I don't go in for oracular bluster like that, but when it comes
to cryptography I've learned to make an exception. The alternative is to
trust the mathematicians. That's no exaggeration: one of the rallying cries
of the crypto crowd is 'trust the math.' I don't, because math doesn't exist
in the abstract. Its relationship to engineering is obvious: engineers implement
math, they make it real, make it happen. Its relationship to law is less
obvious. I don't mean ITAR, Wassenaar, or any other mechanism by which states
would standardize or regulate cryptography. Instead, I mean the kinds of
individual and collective sovereignty that cryptography enables through various
implementations. The Cypherpunks understood this potential in their own way
('crypto anarchy'), and the Bitcoin/altcoin advocates understand it in other
ways -- hence all the experimentation and excitement about things like side
Hard crypto everywhere all the time has become one of those internet pietisms
that's hard to challenge. First of all, anyone who does so ends up with some
really troubling bedfellows (e.g., the NSA). But even if we ignore that kind
of implication (i.e., ultraist extrapolation), we quickly come to basic,
practical questions: If you want anything less than absolute crypto, where
and how would you draw the lines? For example, the lines between what's permitted
and what's forbidden, or what's practically possible or impossible, or for
how long (e.g., key length vs 'Moore's Law' and misc innovations).
I'm skeptical about crypto absolutism because one of its first effects would
be, in effect, to *privatize* everything. 'Public' would be reduced to whatever
was cracked or leaked, as if Wikileaks and Snowden were the norm rather than
the harrowing exception. And that would apply not just to social or communicative
records but also -- as anyone who's lost a key or a password knows -- to
one's own records. And isolated cases, which now seem almost like thought
experiments -- questions about whether the US's Fifth Amendment, against
being compelled to provide witness against oneself includes passwords, for
example -- would become near-daily considerations.
Discussions of cryptography *should* involved questions like this, but they
don't because no one has an incentive to discuss them. Opponents of crypto
are happy mongering the ~four horsemen (terrorists, organized criminals,
money launderers, and c-pornographers, more or less), and too many crypto
advocates are absorbed in exotic last-mile opsec projects. Recently, I read
a casual remark that Baudrillard is cited more often in 'critical accounting'
than he was in the humanities. I doubt that's true, and the field of critical
accounting is completely new to me, but it sounds like it could be promising
-- as a part of a broader challenge to the naive positivism of [text missing]
I don't know the solution to this all, and I don't know where bright lines
should be drawn. But I do think that the growing 'moral' push toward secure
communications is troubling, and that preserving 'insecure' communications
channels as a legitimate choice is vital.
Date: Sun, 24 May 2015 19:39:49 -0700
Subject: Re: <nettime> What should GCHQ do?
There is a fine point here which is almost always missed, but from which
most of these conclusions come from.
It is about the concept that 'crypto' is created by some small set of Illuminati,
it needs to be standardized, and the rest of the world must trust them. These
'crypto wars' are then waged between the mentioned Illuminati and various
evil agencies that would like take away the tools, bestowed by Illuminati
upon the unwashed.
The concept works great both for the Illuminati and evil agencies - both
do everything they can to maintain it.
Illuminati get livelihood: denigrating terms like "home brew crypto" are
deeply entrenched and help maintain the guild exclusivity.
Evil agencies get their job made easy - it is trivial to subvert several
standards or rubberhose few dozen experts into submission. Mass surveillance
is only possible when there is a small number of crypto technologies.
This is all total bs.
While crypto is not the simplest technology in the world, it is far from
being rocket science in practical terms. If everyone that did some scripting
in any language would construct their own custom terribly weak cipher (ROT-14,
ROT-15, etc), and use it only between themselves and their personal
correspondents, totally incompatible with ways that "standard" web sites
and VPNs do crypto, it would become too expensive, for any evil entity, to
break millions of terribly weak ciphers. There is nothing "standard" about
your circle of correspondents. There is no need that everyone in the world
can participate in your crypto technology.
Back to the point: you don't need absolute crypto. You don't need to trust
anyone. Scramble your communications in some custom way that will take evil
agency's analyst 10 minutes to break: they can't afford it. And if they target
you, you are f*cked anyway, no matter what you use.