I keep thinking about cloud computing: I can't say that I really understand what it is, and this may be in part because it names something that I have believed more or less all along: that the internet "is nowhere." Even when I had to connect my computer to a phone jack in order to get online, I never sensed that the internet "occupied" this or that server. Now wireless, the internet (or "connectivity") is dispersed, an atmosphere. Which is all just to say that the internet has been a "cloud" for some time. When Ted Stevens called it "a series of tubes," he caught hell for his "analog" image, but the idea is the same, I think: a passage, betweenness.
I've been spending a lot of time with Luke Howard's wonderful On the Modifications of Clouds (available to read or download here), the 1802 lecture to the Askesian Society that first named the clouds (cirrus, cumulus, stratus, etc.). Howard's classification is in some respects perfectly conventional (the sort of Enlightenment project Michel Foucault describes in The Order of Things), in other ways unprecedented. Most remarkable, I think, is Howard's choice of "modification" rather than "type" or "category." It is a term singularly suited to his object, denoting "the bringing or coming of a thing into a particular mode of existence" or "differentiation into a variety of forms or modes."
(This is one of John Constable's cloud studies)
Such is the "manner" of clouds. They have no fixed "form or magnitude." (And as such lack two of the four "variables" of classification identified by Foucault: form, quantity, distribution, and magnitude.) Howard adapts classification to an object that would seem to dispute its very foundations. Cirrus, cumulus, and stratus, in turn, do not behave like conventional categories. Howard confidently asserts that "the principle modifications are commonly as distinguishable from each other as a tree from a hill, or the latter from a lake." But, he continues, "clouds in the same modification, considered with respect to each other, have often only the common resemblances which exist among trees, hills, or lakes, taken generally." No problem: we can accept that a palm and a beech are both trees. But things get weirder. In the case of clouds, "the same aggregate which has been formed in one modification ... may pass into another." Trees do not "pass into" hills, nor hills into lakes.
Cloud as "modification" is "bringing or coming into" existence as well as "passing on" from one "mode" of existence to another, and passing out of existence, dissipating. In this respect it is something like the "process" Alfred North Whitehead describes in Process and Reality: a "concrescence," first, and secondly a "transition from particular existent to particular existent." The "passing on" (his phrase) of process is, he writes, "'creativity,' in the dictionary sense of creare, 'to bring forth, beget, produce.'" "Process" as "production" describes, in turn, the "flow" of capital in described by Deleuze and Guattari. As "process," Howard's cloud anticipates the radical contingency of weather’s "sensitive dependence," or chaos theory, introduced in the work of the meteorologist Edward Lorenz and later adopted by economists to describe the vagaries of international capital. A cloud is a "deterritorialization," or "line of flight" (hence the inadequacy of "arborescent" classification). It is a "body without organs" in this sense: "that which serves as organs … is distributed according to crowd phenomena, in Brownian motion, in the form of molecular multiplicities." (See Lucretius.)
When we name Google a "cloud" we refer to the virtualized labor and resources of "cloud computing"--to the "cloud" of (deterritorialized) capital. Capital "is nowhere." Weather was this sort of "cloud" long before the internet, at least since the mapping of the trade winds (in the seventeenth century), or the advent of weather "networks." I suspect that its flow has always been entwined with capital's.