TRANSPARENT BUNDLES by Seth Goldstein

MEDIA FUTURES 2006: 2/5 ALGORITHM: The Transition from Automata to Algorithm

In the beginning there is the Automata.  It is the prime mover: an intention that drives human action.

Over time, the record of these actions, both individually and across people,  establish a unique pattern of behavior that is known as an Algorithm.  In the context of Media Futures, an algorithm is a computing engine designed to process behavioral data and convert it into content that engages ones Attention.

You can imagine an Algorithm like a strange Rube Goldberg machine with a complex set of routers, pulleys and chutes that turn a certain input into an equally certain output. 

Recall our reinterpretation of Hamlet:

We encourage others to participate so that we may consume them
and we make ourselves interesting for the blogosphere.  Your Internet CEO and your Joe Blogger are just different algorithms- two APIs, but to one network.

Each decision that I make as to what to pay Attention to, and the physical gesture that I use to effect this choice (search, click, form, sign in, etc) establishes a little personal algorithm that gets joined with all of my other personal algorithms.  Together, this bundle of personal information algorithms establishes and maintains my persistent, stable electronic identity.  This is a deeper, more authentic version of me than simply a numeric ID that establishes my offline physical presence.  The me that makes me me online is one that I actively create and reinforce every moment based on a series of interlocking gears (which I control based on data I produce).

 

And now if you pull up from the tree of me as an individual to the forest of all of us in society, then you see a much broader fabric.  The fabric represents Social Media, each of its infinite threads representing one individual's momentary micro algorithmic gesture. 

One would assume that each of these mini decisions was distributed to the edges, and that the control over it was determined by its owner.  But this would be to ignore the gravity of the Attention economy, which is Influence.  On the Internet, Influence is measured by the amount of Attention one gets relative to the amount of information one gives  The most influential online individual is able to syndicate a limited but steady stream of what makes me me-ness through his personal API and nevertheless generate a high pagerank that lands him above the Google fold.

The area that I am most interested in exploring in this current chapter on Algorithm is the rub between what you are searching for and where you emerge from other people's searches.  This is located between (1) the record of your Attention (for example as expressed recently by AOL through their disclosure of "anonymous" search histories) and (2) the position you occupy within the pagerank universe based on what keywords produce results that point to you, above the fold.

For me,  these keywords might include: "Seth Goldstein"  "Media Arbitrage"  "Algorithm Futures" "Transparent Soft Dollars"

Coming next, a brief history of Algorithm

Media Futures 2006: 1/5 Automata: The Human Computer

While the recent inventions of Web 2.0 and User Generated Content (UGC) seem to be radical departures from the computing culture we grew up in, their organic social metaphors are in fact rooted in the beginning of computer science.  In the 1940’s and 50’s work of Alan Turing, John Von Neumann and Norbert Weiner, most discussions of the future of computing evolve into a study of the brain.  The natural automata of human thought, the way in which our ideas express our independence, this is the machine intelligence that technologists tried to design into early computers.   

Alan Turing was fascinated by Automata and its relationship to natural human thought.  In his 1950 “Computing Machinery and Intelligence,” Turing outlined an experiment that was able to determine whether a computing machine could be defined as having the capacity to think.  The Turing test functions as follows: Human “X” and respondent “Y” take part in a teletype conversation, but X cannot know whether Y is human or a machine.   If, after a specified amount of time, X believes that Y has responded like a human, and Y is a machine, then Y can be defined as having that human capacity of thought.

In his biography of Turing, William Aspray writes that this:

“was among the earliest investigations of the use of electronic computers for artificial-intelligence research...He attempted to break down the distinctions between human and machine intelligence and to provide a single standard of intelligence, in terms of mental behavior, upon which both machines and biological organisms could be judged.   In providing his standards, he considered only the information that entered and exited the automata…Turing was moving toward a unified theory of information and information processing applicable to both the machine and the biological worlds.”

The fusion of machine and biology is promoted as a core computer architectural principle in the Interim Progress Report on the Physical Realization of an Electronic Computing Instrument:  Julian H. Bigelow, James H. Pomerene, Ralph J. Slutz and Willis H. Ware; Princeton: The Institute for Advanced Study; 1 January 1947.  This report was prepared for John Von Neumann, and the rest of the IAS authorities, on the development progress of a machine based entirely on mathematical equations.

Left to right: James Pomerence, Julian Bigelow, von Neumann and Herman Goldstine

Von Neumann had joined Princeton’s Institute for Advanced Study as a Mathematician in 1933.  About 10 years later he started concentrating on something less theoretical and more practical (which alienated many of his colleagues): building an electronic computing machine.  This project was a deep meditation on the act of creation.  Some of the greatest minds, across a variety of disciplines (math, biology, engineering, physics) converged in Princeton to help Von Neumann “physically realize” his ideas. 

IAS Report, 1947

According to the report, Organs are:  “portions or sub-assemblies of the machine which constitute the means of accomplishing some inclusive operation or function; as “arithmetic organ.”  Note how the processor in this case is able to extend its influence onto others in an “inclusive operation.”  The organ of social media was anticipated already then, in 1947, even without an Internet to enable it at scale.   

Von Neumann continued to extend his computer research towards an understanding of the human brain.  He described this specifically in his introduction to his 1958 work The Computer and the Brain:

In 1948, Norbert Weiner, the leader of cybernetics wrote Control and Communication in the Animal and the Machine.  His use of the word animal is different than Turing’s logic or Von Neumann’s brain, but he is similarly concerned with the organs of information and their ability to relay information between systems:

“It is a noteworthy fact that the human and animal nervous systems, which are known to be capable of the work of a computation system, contain elements which are ideally suited to act as relays.  These elements are the so-called neurons or nerve cells... The mechanical brain does not secrete thought <as the liver does bile>, as the earlier materialists claimed, nor does it put out in the form of energy, as the muscle puts out its activity.  Information is information, not matter or energy.”

Weiner, Control and Communication in the Animal and the Machine, 1947

In late 2004, the creator of del.icio.us Joshua Schachter described to me that tags were simply crystallized attention.  Both terms interested me: while attention has become my chief investigation, the transparent materialism expressed by “crystallized” has also been a key focus.  When you put these together, you get, in Weiner’s words, a “secretion” of passive behavioral data.

Seth Goldstein, April 2006

Just because a tag is a form of  information doesn’t mean that it lacks physicality  Without being matter or energy, can a tag be made of something else, something that comes closer in nature to mirror neurons?  Attentrons.  Remember that mirror neurons are a form of biological material.  These mirror neurons fire when the subject performs an action, but also when it observes somebody else performing an action.  In this latter case, the successful firing of a mirror neuron is based entirely on its ability to passively mimic the behavior of somebody else.  In this quiet absence of a human impulse, attention is full.

Media Futures 2006: 1/5 Automata: Industrial Automata: From Performance to Prosthesus

Automata evolved from acting like us to acting on behalf of us.  What if it was possible for an automaton to do the work of a human?   Would the stuff of Aristotle's ruminations come to pass, eliminating the need for servants and slaves?  To realize that fantasy – as Albertus Magnus supposedly did in the 13 th century, with the construction of a life-size human domestic servant automaton – is to eliminate the need for humans to pay Attention to certain aspects of work in the home.  These automata promised to give owners a surplus of energy and attention, but at a cost.

Scientists have long recognized, Sir Geoffrey Jefferson argued in his 1960 "The Mind of Mechanical Man", that aside from the mind, “both animal and human bodies were nothing more than a collection of pumps, reservoirs, bellows, fires, cooling and heating systems, tubes, conduits, kitchens, girders, levers, pulleys and ropes.”   (Clearly, as John Stewart points out, little has changed since)

In this model, automata do not have minds, hearts, or souls.  This was all the better, since those human aspects might have forced the automata into unnecessary error.  These perfect machines began to gain power over the very humans who operated them, a power which became even more threatening when humans wrestled with the possibility that their creations might actually come to life.  What, then, would these exploited classes do?   Embedded in their very name are the seeds of revolutionary threat: robot (as human-shaped automata are widely referred to) comes from the Czech robota, meaning forced labor, and it is a term that was first used in Karel Čapek's 1921 play Rossum's Universal Robots, or R.U.R. 

 

 

The questions raised by this term in Čapek's play are central questions in the discourse of modernity – questions of the nature of man and machine and of the boundaries between the two.   Human-like machines threaten to come to life and overpower the control of their former masters; machine-like humans threaten to destroy life, overpowering any sense of humanity and the body of humanity itself.   

 

(Picture: Georg Grosz. Republican Automatons.)

From the faceless figures in Georg Grosz's Republican Automata (above) with hooks for hands and gears for souls, to the orphaned machine of Francis Picabia's The Child Carburetor (born of the work of man but devoid of his agency) the machine-like human and the human-like machine confuse our sense of where us stops and the machine we created to stand for us begins.

 

(Picture: Francis Picabia. The Child Carburetor.)

In 1950, Norbert Wiener wrote in his 1950 work on cybernetics "The Human Use
of Human Beings":

When human atoms are knit into an organization in which they are used, not in their full right as responsible human beings, but as cogs and levers and rods, it matters little that their raw material is flesh and blood.  What is used as an element in a machine, is an element in the machine.  Whether we entrust our decisions to machines of metal, or to those machines of flesh and blood which are bureaus and vast laboratories and armies and corporations, we shall never receive the right answers to our questions unless we asks the right questions.

Let us bear this careful warning in mind, as we evaluate the visions being articulated now by our most noble leaders of the Internet.  Take for example the recent speech by Ray Ozzie, Chief Software Architect at Microsoft to an audience of financial analysts:

But beyond infrastructure services, what's most unique and valuable about a very large-scale services platform is what I'll refer to as optimization. By optimization I mean the monitoring and utilization of both collective end-user behavior and individual behavior to rank content for the user. That ranked content might be the order of advertisements in a search or e-mail window, or the order of relevant news items or playlists or video clips or items in a marketplace that are presented to the user...Optimization always respectful of a user's privacy will be increasingly key to delivering great user experiences, and it's already a key factor in the area of profitability, because the larger the number of users that are connected to any services platform, the more behavioral the data that can be generated. The larger the number of PCs and other devices that are connected to that platform, the more behavioral data that's available; the larger the number of applications connected to the platform, both Web apps and desktop apps, the better our optimizations will be and the more profitable it will be for us and for our partners.

It is remarkable the extent to which Ozzie seems to ignore the fundamental Web 2.0 premise that users are in control, and that just because behavioral data may be generated automatically, that does not mean that the companies enabling such data (ie Microsoft) have necessary dibs on it.

Over history, automata were at once objects calling for our attention – objects meant to enrapture us in spectacle – as well as objects that offered to do our work without requiring our Attention.   This represented the threat of triumph over human mastery: that these automata might take on lives of their own, rendering humans obsolete and placing us at the mercy of the machine which always acts without emotion, error or thought.  The power of choice manifests itself in one's ability to ask, in Weiner's words, "the right questions."  Right questions might be those queries specifically which elude their engine's best attempts at matching them to willing advertisers.