Tuesday, November 6, 2007

interface and automation

Paraphrase from The Simpsons wherein Bart and Lisa go to military school: The wars of the future will be fought by machines. Your job as soldiers will be the cleaning and maintenance of those machines.

A recent 60 Minutes story defines the present state of human-machine interaction in warfare. In Afghanistan, the U.S. military dropped a bomb on a house where a suspected Taliban supporter lived. It did so in response to an RPG that had been fired at a U.S. base (it missed, badly). The air strike didn’t kill the suspected Taliban supporter but rather 8 members of his family, mostly old men, women and children.

The story described the basic method being used to order an air strike. Soldiers on the ground did initial investigations of the house the day before and requested the strike. Someone at a U.S. Air Force control center in the Middle East had to make the final decision based on the report from the ground, grid coordinates, time, tonnage, etc. The Air Force allowed 60 Minutes to film the control center, which was filled with large screens and banks of computer operators (imagine the big board from Dr. Strangelove perched above telemarketing cubicles).

Even if war machines have predatory capabilities and sophisticated AI, would the result have been any different? Probably not, because it seems as if the decision was largely deductive. The U.S. should eliminate its enemies; an enemy lives in that house; therefore, a bomb should be dropped on that house. A machine could make that terrible, simplistic decision.

De Landa’s argument is that, in order for machines to be effective without human interface, information processing would have to at least demonstrate that it could be used to automate self-organizing processes using a complex singularity. The key to this would be the development of inductive and heuristic reasoning, or learning from experience and making good judgments, and the inability to generate this in a machine at a sophisticated level is the reason the U.S. military must utilize a human interface. There is a certain merit to this argument between automation and interface. Interface, at least for the military, is a temporary fix to a problem with automation not easily resolved. Once, resolved, at least humans can still interface with machines with a cloth and some polish.

"Machinic Phylum" as explanation and description

This text is pleasantly bewildering. I will attempt to pinpoint a peculiar confusion as it pertains to the concept of the machinic phylum; this concept seems of utmost importance for the mode of insight that this piece promises.

On the one hand, "machinic phylum" can be explanatory. Harkening back to scientists (along with their theoretical perspectives and vocabulary), this term explains phenomenon which occur... but I hesitate to say "naturally" since it is the logic of the non-organic. By "explaining" we could say that it provides an analysis of "causality;" Things happen because they happen for a simple, mathematically derivative reason. This, however, is also tricky since we would like to say causality (and the knowledge of what mathematics describes) is natural and unidirectional. The point of the causal analysis as regards the machinic phylum though, suggests that there is a bizarre emergent property, a "cooperative" formation arising arbitrarily to be a whole greater than the sum of its parts. Something could thus gain an intelligence, like a hive of bees or collection of amoebas or flock of birds nesting, and this intelligence can be "traced."

On the other hand, "machinic phylum" appears to be a concept which operates as an analogy (a term he conveniently uses on page 23 and elsewhere without considering too many of the ramifications of this view). In this respect, the phylum is not necessarily a reflection of what is actually happening but a way to see history "from the perspective of the robot." Hence, DeLanda seems to use this conceptual insight to describe much more complex interactive, human, social phenomena. We can see the mass "flow" of history like a fire or jet stream turbulence - the way in which order emerges out of chaos.

It would seem that these two aspects of the phylum would link up to the straightforward bifurcation of what Delanda gives us on page 20 - for he wishes to look at two notions, order of out chaos and the specific formations that result. In one of his examples, he looks at the singularity of forging iron at the non-organic, machinic level and the use of the iron at the social, phylum level. However, if there is an overlap as I see it, if DeLanda offers us a machinic description of history as non-organic to some extent - the machine of war - then he is attempting to elide the distinction of explaining machinic social formations with describing phylum "singularities." Two holistic problems arise which are greater than the parts of the argument that DeLanda can control: 1. Singularities are by definition that which we cannot explain or account for - the very idea of emergence as beyond the logic of the governing machine of mathematical rules of operation. 2. There is something odd about the "looping" back of intelligent interface going on here. I hope to have this formulated by the time we meet in class, but my suspicion is that intelligence is both inactively part of the assembly and greater than the sum of the war machine whole to reflexively interface with computers...

What I am having trouble understanding is the singularity of conceptual understanding itself once it arises out of the matrix of the machinic phylum of ordering - once intelligence gains the perspective to look back upon the conditions of its own formation. The very emergent property looking at itself now is more than that which produced it. It seems here we could go one of two ways - Hegelian consciouness and the concept of aufhebung or the materialistic soft determinism of mathematical and lower level operations. DeLanda seems to perform both of these positions somehow.

Monday, November 5, 2007

I quit.

I’m not going to post anymore. In fact, I’m converting to Neo-Luddism. It’s just so depressing to read this without feeling like we missed the boat in the early 1990s: Rheingold called for virtual communities relying on an anarchic gift economy with alternative media outlets and De Landa called for augmenting man’s faculties rather than sloughing off humans. What happened?

The answer, which De Landa seems to outright ignore, seems to be capitalism. I tried to lay out his periodization but I had trouble because it seems like he fails to account for the role of capital (or if you’re Dyer-Witherford, labor). Is there such a thing as a militaristic-determined narrative? De Landa claims, “Interactivity, the passing of the machinic phylum between humans and computers, was developed both as an intellectual goal by visionary scientists and conquered in battle by the hackers” (226). It’s not about connecting or informating, it’s about interfacing better with machines? It seems he forgets the struggle between labor and capital when he claims, “Logistic considerations regarding the procurement of manpower are behind the drive to get humans out of the loop” (229). Even his move from the mercenary of the clockwork age to the patriot of the motor age seems to elide some pretty important changes in subject-formation and geo-political organization which can’t be so easily chalked up to militaristic ambition.

Clockwork Age

Motor Age

Distributed Network Age

Geared Mechanisms

Sequential Procedures

Parallel Procedure (Non-Sequential)

Aristotelian Syllogism

Boolean Calculus

Pandemonium

War of Attrition

War of Annihilation

War of Machines?

Mercenary

Loyalist / Patriot


Aristocratic Military

Meritocratic Military



Panopticon

Panspectron

Batch-processing

Time-Sharing

Hypertext

Gears

Thermodynamics / Steam

Integrated Circuit

De Landa stands at a fork in the road: “The future of interactivity would depend on the events that took place at that surface of contact: the computer screen could be turned into a new method of enslaving people or transformed into a means to ‘augment man’s intellect’” (220). Machine as prosthesis or substitute? This seems to be the same struggle between Marx’s Grundisse and The Fragment of Machines, which sees all variable capital eventually becoming fixed capital; and Virno’s General Intellect, which reinterprets Marx and notes the importance of humans. Hardt explains the latter’s position, “Interactive and cybernetic machines become a new prosthesis integrated into our bodies and minds and a lens through which to redefine our bodies and minds themselves” (“Affective Labor” 95)

I think De Landa’s most interesting point comes toward the end of the reading when he explains the importance of the Pandemonium, reminding us of the importance of adaptability and the emergent properties we read about in Terranova’s fourth chapter. “Pandemonium is such a machine concrete enough to allow the control of physical processes, but abstract enough to allow the spontaneous emergence of order out of chaos” (229). De Landa goes on to note the paradox at the heart of the military’s attempts to create autonomous weapons: “The software control structure with the least amount of central control, the Pandemonium, is the only one that works for the purpose of creating true Artificial Intelligence” (230).

Ho Chi Minh

Delanda's article takes me back to the not-so-distant past (September 4th) and our reading of Paul Edwards' work "“Why Build Computers? The Military Role in Computer Research." This reading seems to cover similar themes concerning automation/interactivity concerns....

Early on in the Edwards piece is an account about complex sensors on the Ho Chi Minh trail that were used to dispatch airplanes to attack spots where the sensors had detected some form of trouble. The trouble turned out to be something the military with all its machines and might couldn't easily detect - "The guerrillas had simply learned to confuse the American sensors with tape-recorded truck noises, bags of urine, and other decoys, provoking the release of countless tons of bombs onto empty jungle corridors which they then traversed at their leisure," (p. 4).

Delanda's oppositions and the construction of military machines reminds me of the "Boids" we discussed in class. Following very simple rules the system is able to simulate complex behaviors. Delanda seems to state that future development of military machines will leave the only window of hope to come from hackers and visionaries. Well, what about IED's?

Improvised Explosive Devices (IED's for short), have been readily used in the resistance to the current occupation in Iraq. According to the Washington Post, 63% of U.S. deaths in Iraq have been killed by IED's. "They [IED's] now account for nearly two-thirds of military deaths from hostile actions in Iraq, and slightly less than half in Afghanistan."

IED's, bags of urine, recordings of truck noises...How do these "guerrilla" tactics fit in with the picture Delanda paints?

Well, and maybe these examples don't exactly fit in with what Delanda discusses concerning the development of military intelligence. It strikes me that these simple machines (IED's and bags of urine) do so well to subvert a highly mechanized and even computerized army.

War Machines

The war machine is the result of royal science capturing the limbdinal energy of science, the hacker is fast and smart. Nomadic science is that which escapes the conflation of desire. Odepial psychology seems to demand that our AI be obsessed with killing its own father. The question of desire is what I am concerned with. This is all a fantasy of overcomming thermodynamics, we can only program into the machine desire for so long, we must develop a model of the real that would drive the machine forward. The pattern of mechanical consciousness would fit the model of consciousness that is prevalent in the military.

Here is the thought problem- can you socialize an AI? How does it learn? It if is programmed with the Odepial logic (which I am sure it would be) then it will have desire, there will be a real that would drive it forward and demand that it destroy. The machine would need to be programmed with really bad Freud, the death instinct and some odepial action and we are good to go.

Interface, Interaction, Autonomy

Delanda makes some rather sweeping pronouncements about periodization in addition to setting up some key oppositions (clockwork, motor, e.g.). I'm interested in your take on these examples of periodization and the terms of analysis. To his various taxonomies (of stages, components, and oppostions) "work" to make sense out of the changes he describes? Which ones did you find most productive or least and why? One of the key tensions that emerges in his account is that between automation and interactivity -- this is a recurring theme in the class and I'm wondering if you found Delanda's contributions useful in considering the relationship between the two. As Delanda puts it, "one and the same program may be used to take human beings out of the decision-making loop or, on the contrary, interfaced with them so as to create a synergistic whole. It is the design of the interface which will decide whether the machinic phylum will cross between men and machines, whetehr humans and computers will enter into a symbiotic relationship, or whether humans will be replaced by machines" (176). Much hangs on the interface, according to this account. How convincing is this account, and what does Delanda provide in the way of thinking about the interface design for symbiotic purposes?

Thursday, November 1, 2007

Ideology redux

I think Niko's question about the position of the reflexive academic is central to Dean's argument -- not necessarily because of she is one herself, but because I think the point of her argument is that it is directed squarely at a position that is quite familiar to contemporary academics, particularly those who "see through" not just ideology, but the ideology of ideology itself, if I can put it that way. That is to say, it is directed not just toward those who would engage in the process of unmasking (myself, naively thinking that I could cause the scales to drop from the eyes of my undergrads -- students who had long ago been descaled), but those who engage in the unmasking of ideology critique itself. Think, for example, of the familiar position that goes something like this: "Critiquing ideology always implicitly or explicitly invokes a truth claim (presumably guaranteed by some big Other) -- you can't tell someone they are "in" ideology unless you locate yourself in a position that is able to distinguish between ideology and something that is not ideology. But, thanks to our postmodern reflexivity, we understand that it is precisely this position (that of the debunker, not the debunkee) that is ideological: there are no grounds for positing the position from which this distinction can be made (demise of symbolic efficiency, decline of the "big Other"). I think a key point here is that such a formulation reproduces the cynicism described by Dean: it's not just that the student/member of the public has become savvy enough that unmasking doesn't work, it's also that the position of debunker comes to coincide with debunkee. That is the academics bent on "enlightening" the public come to adopt a cynical reflexive attitude toward their own position (we know that there is no position from which to adjudicate between ideology and the extra-ideological). There is no point in telling us that all attempts to seek a ground (or closure, suture, whatever) are merely ruses of power...we already know. This is the final debunking: debunking is just one more form of bunk(ery?). The second point that Dean, drawing on Zizek makes, is that this cynicism (savvy reflexivity) nevertheless remains ideological. The notion that we can stand outside ideology is ideological in both of its manifestations: 1) that we can adopt the position of science or truth from which to tell others they are caught in ideology AND 2) the notion that we escape ideology by moving beyond its problematic (ideology critique is quaint and outdated -- a relic of a naive era of critical theory). For what is this second position but an attempt to stand outside or detach ourself from ideology (we see through all truth claims as one more ruse of power, in order not to be seen to be dupes)? This second position might describe the conspiratorial subject position of the academic (always looking for the power interests to which any truth claim can be reduced).

So here is Dean's second point: this cynical distancing is the ideology of contemporary capital. And here is the point about what might be described as "postmodern criticism, or, the academic logic of late capitalism": it instantiates, at the level of theory, the reflexive cynicism characteristic of what Dean calls "contemporary technoculture" (and communicative capitalism). It incorporates a (wide?) stance of "ironic distance." Thus Dean asks, "how does the valorization of fragmentation and contingency benefit global capital by preventing it from being understood as a totalizing modality of power?" (and here, right away, Dean invites the reproach that she is not cynical enough -- she defaults to the position of the dupe from which the savvy subject seeks to distance her/himself).

OK, how? At the level of the savvy consumer/citizen, the subject knows the entire culture industry is a commercial fabrication but continues to participate (and this is the point I was trying to make, rather clumsily, with the example of the Halloween costumes: they all know the culture industry is a for-profit system of image manipulation, and yet, they go ahead and buy the 2Gether albums and dress up according to the images that have been manufactured for them. (Isn't Paris Hilton THE example of the ironic celebrity -- the celebrity who is famously reviled for being famous solely for being famous, and yet -- and yet, she continues to sell products and accumulate "mindshare" -- we mock her fame, and in so doing magnify and reproduce it: we are not "fooled" by it, but this doesn't make it go away. On the contrary.).

How does this ideology operate at the level of the savvy academic. I'll try this out to see if I can generate some controversy/response. One possible example. The author is one Alan McKee, poobah at the Creative Industries campus under the tutelage of John Hartley (well known in audience studies/cultural studies). Here is a quick breakdown of how I think this type of position operates: McKee's version of "distinction":
step 1) culture is a social construction that reproduces class distinction. step 2) (Bourdieu Anglicized and reduced) Culture is purely reducible to the role it plays in creating distinctions. Step 3) Culture itself does not matter -- what matters is what gets done with it. Step 4) there is no symbolic standard for adjudicating between cultural forms (decline of big Other). Step 5: all cultural distinctions are purely a matter of taste -- those who pretend otherwise are seeking to impose a debunked big Other (are exercising power). Step 6: The market is naturalized: our tastes are our own, do not take them away from us, and do not invalidate them by judging them.

The paradox of this position is that it naturalizes taste even as it acknowledges its contingency (and constructed character).