Seth Goldstein

There is usually so many things I want to write about that I end up overdigesting my ideas and write very little… and when I do write it tends to be long and expansive and hard to process in one gulp.

I had drinks last week with Josh Kopelman and Rob Hayes of First Round Capital. I have the unique distinction of having Josh as an investor in companies I founded (Majestic Research, Root Markets), of being an investor in Josh’s fund as an LP, and of investing directly alongside of Josh (del.icio.us and Aggregate Knowledge). Suffice it to say that Josh knows my many sides and so I appreciated his perspective on my blogging: “Seth, all you ever seem to write are long, esoteric posts about… Attention. You have such great, funny perspective to share.” He wanted to know not what I thought of metadata ownership, but rather my off-the-cuff reflections as a recent New York transplant in Silicon Valley.

I told Josh and Rob how enthusiastic I was about the Facebook ecosystem that seemed to be bubbling up like thick layer of foam over a double shot of Google. It is as if in a matter of months, both the high end of LinkedIn and the high-end of MySpace had been absorbed into the Facebook social graph. LinkedIn is suddenly no longer the social network of choice for us chic geeks. Yes, we learned how to tell our professional stories these past few years in the LinkedIn profile fields, but- as in summer of our 8th grade- we are now ready to lose the awkward friends we had accumulated , and start from scratch in a new environment.

Meanwhile, the kids who treated their MySpace profile, and concomitant friend requests, with the same reckless abandon that we have done with our LinkedIn profiles, have now de-camped for Facebook. While I don’t have fresh data on hand to support this hunch, the well-sourced rumor I heard last week ab out MySpace scrambling feverishly to open their API’s reinforces what is becoming obvious: MySpace’s Kremlin-esque behavior towards 3rd party widget developers -”we buy them or we crush them!”- is on a crash course with the debauched dirty-dancing going on amidst the MySpace spring-breakers. As these kids move from junior high to high school, from high school to college, and from college to the work force, they are increasingly choosing the meritocratic social logic of Zuckerberg over MySpace’s “hot or not?” popularity contest

There can be little doubt now that Facebook is a platform for social media, as opposed to simply a web site community. Time will tell whether it can continue to scale through opening up its audience to 3rd party developers like Microsoft did in the 80’s. This weekend I watched the Gates/Jobs conversation from D on my iPod. The elephant in the room that nobody really discussed was the fact that competition stopped at a certain part in their relationship, Microsoft became a monopoly, and Gates became the richest person in the world.

This was not an accident, but was in fact a direct result of the platform strategy that Microsoft so successfully pursued. Back in March, 2005 in a blog post, I recounted a meeting that taught me more about platforms than anything since:

In 1999 I sat down with Brad Silverberg of Ignition VC who Microsoft recruited out of Borland in the early 90’s to become the lead developer and project manager of Windows 95. Never has there been a more valuable platform. He described 3 things that platforms needed to have:

• wide distribution
• application developers making money
• good tools

Let’s test those three axioms against the preeminent platform play of our time, Google:

• Wide distribution? YES
• Application developers making money? YES (if you count all the adsense publishers)
• Good tools? YES (all the adwords and adsense self-service goodness)

Now let’s test these axioms against Facebook:

• Wide distribution? YES
• Application developers making money? NO (at least not yet, I will comment on 3rd party Facebook developers such as Slide, Rockyou, and AttentionSoft)
• Good tools? YES

So, the question for establishing Facebook’s value as a platform is no longer whether Facebook itself can make money but whether its developers can do so. And in a world where retail software sales are no longer a legitimate business model for developers, the default assumption is that these developers will make money through advertising. Which begs the question as to whether there is a pragmatic alternative to Google adsense (no) and therefore suggests that Facebook will need to create this for its developers.

Later in that post from March 2005 I related platform strategy to API structure:

Nobody controls the web as a platform the way that Microsoft controlled the desktop. But certain parties do control enormous pools of user data and direct their behavior…API’s are fountains of data, mostly consumer meta data, that are the byproduct of some other functionality… The value of a web service API is tied to its ability to convert granular feeds of individual data into useful social media contexts.

I wrote this before I had ever used Facebook but the implication is clear now. Google does not offer this Social Media API. Facebook does. Here is an example of its API syntax:

facebook.friends.areFriends

Returns whether or not each pair of specified users is friends with each other. The first array specifies one half of each pair, the second array the other half; therefore, they must be of equal size.

What could be more useful as a social media context for a software application than being able to ask whether two users are friends with eachother?

I always loved you

You always had a lot of style

I’d hate to see you on the pile
Of ‘nearly-made-it’ s

You’ve got the essence, dear

If I could have a second skin

I’d probably dress up in you

Belle & Sebastian, The Life Pursuit

Welcome to the readerverse. Just as pages and sites have their community in Ted Nelson’s concept of the docuverse, so users and visitors have theirs in the readerverse. It is a place where responses are generated as the primary activity. This occurs when we are reading, browsing, searching, scanning, tuning into, subscribing and, generally, using the Internet passively, automatically. The readerverse shadows the more explicit actions of writing, commenting, rating, taging and coding.

Via my various widget logs, I have been trying to illuminate my own readerverse.  I write things and then listen for the barely audible click steps that you make when you visit; the slight pinging sound you make when your reader checks my RSS feed.  With some of the emerging blog statistics and Attention tracking services that are emerging, the web is increasingly rendering as visible what we have come to think of as invisible.

One way of thinking about this in the real-world would be imagine what it would be like if your gaze left a mark? What if when you looked at somebody, instead of that being your private experience, that the person  immediately felt that she was being watched by you? How would that change the way we behave?

We feel free to watch certain things, listen to certain conversations, tune in to certain channels, without worrying about these Attention choices being exposed to others.

This is a fundamental media right: the preference we enjoy in knowing that our media choices (ie our decisions about what we choose to pay Attention to) are not only under our control but are private to us.

Bishop Berkeley asked whether the tree really falls if nobody is there to observe it.  This applies to the physics of Attention. If my gaze is imperceptible to those I am paying Attention to, then I remain the  sole source of information on my media consumption habits.

However, if my gaze has material properties that impact others, then there are by definition other sources of authority on my Attention data. 

Almost seventy years ago, Alan Turing,   the brilliant British computer scientist and war-time cryptographer, suggested that:

The behaviour of the computer at any moment is determined by the symbols which he is observing, and his "state of mind" at that moment.
Turing, On Computable Numbers, 1938

Now isn’t that interesting? A conception of computing, from 1938 no less, in which the computer’s behavior is driven by the Attention it is paying.  Putting this in the context of Goldhaber’s theory of the physics of Attention will lead us to important laws on how influence is created:

There is only so much attention (available from other humans), and many or most of us want more than we have.

In order to get attention one needs to express or do something — let us say, perform in some way. (This can be putting forth information, but that is not particularly what, e.g., a trapeze artist does.)

The more attention we get in comparison with the attention we pay in putting together our total performance, the greater our attention productivity.

The more attention we have, period, the more influential we are.

The more attention you get now, or have gotten in the past, the more attention you can get in the future. (Attention wealth is stored in the minds of the attention payers.)

Having others’ attention means you can rely on some attentiveness from them as well. Attentiveness is a willingness to satisfy your desires whatever they may be — as long as these desires do not go too much against what the attention payers (audients) would otherwise want.

Though all this has always been true, new attention technologies, and particularly the Internet, make all this work much more directly. They make it easy for more of us to seek attention, and if and when we get it, to have other desires satisfied as well.

Michael Goldhaber, February 2007

API stands for Application Programming Interface.  In the context of Media Futures, an API routes the output from one’s own unique Attention-processing Algorithm  into an Alchemical reaction triggered by the convergence of other human-driven APIs.

I wrote my first post about APIs in the Spring of 05, at a moment when APIs such as those of Flickr and del.icio.us were just starting to become becoming popular targets of developers.  Since then, the subject of APIs has become commonplace in any discussion of the future of media.  In fact AOL- that stalwart of old new media- is now obsessed with open APIs.  Tina calls it the “the liberation of egosystems.”  Open data transport has suddenly become de riguer among the even the most traditional media companies.  In less than two weeks, legions of their corporate development executives will descend upon SF to walk down the red carpet of the O’Reilly ceremony, ready to sign the top Web 2.0 talent to long-term studio deals.

But while we all fall over ourselves to proclaim our “openness,” we introduce a far heavier burden of trust into the mix.  Is one company’s “open” the same as another’s?  While I may be able to avoid data lock-in in that silo, how do i know for sure the next “open data” silo will be equally amenable to the mobility of my data?  These questions beg a deeper investigation into the history of APIs and their evolution both physically and electronically.

History

In a memorandum dated July 15, 1949, Warren Weaver, who held the position of director of the division of natural sciences at the Rockefeller Foundation from 1932 – 1955, wrote about the possibility of language translation by an electronic computer.  It was the first suggestion most had seen that such a thing might be possible, and as he draws the memorandum to a close, his words preview the emergence of the API:

Think, by analogy, of individuals living in a series of tall closed towers, all erected over a common foundation.  When they try to communicate with one another, they shout back and forth, each from his own closed tower.  It is difficult to make the sound penetrate even the nearest towers, and communication proceeds very poorly indeed.  But, when an individual goes down his tower, he finds himself in a great open basement, common to all the towers.  Here he establishes easy and useful communication with the persons who have also descended from their towers.

Thus may it be true that the way to translate from Chinese to Arabic, or from Russian to Portuguese, is not to attempt the direct route, shouting from tower to tower.  Perhaps the way is to descend, from each language, down to the common base of human communication – the real but as yet undiscovered universal language – and then re-emerge by whatever particular route is convenient.  Such a program involves a presumably tremendous amount of work in the logical structure of languages before one would be ready for any mechanization.

The key to examining the evolution of the role of the API in context of Media Futures lies, in fact, in the multiple resonances of its last term, Interface:  as a surface lying between two portions of matter or space, thus forming their common boundary; as a means or location of interaction between two systems or organizations; as an apparatus designed to connect two scientific instruments so that they can be operated jointly, the abstract concept of an interface contains in it the possibility of a very literal connection between two beings, two faces.  As a physical interface connects two pieces of hardware, a user interface connects a human and a computer and a software interface connects separate software components so that they may communicate with one another.  To interface is to come into interaction with a thing or being, to communicate, in manners both figurative and literal.

Mainframes

As a platform that allows a computer system, library or application to open itself to use by other computer programs, or to allow for the exchange of data between them, the APIs of yesterday were IBM mainframes and Microsoft SDKs, arcane languages of translation between hardware and software.

From the late 1950s through the 1970s, a number of American, German and British manufacturers (Burroughs, Control Data Corporation, General Electric, Honeywell, NCR, RCA and UNIVAC; Siemens and Telefunken; and ICL, respectively), produced such mainframes, computers used in large part by companies and government institutions for the purposes of bulk data processing in the context of, for example, the census or financial transaction processing.  IBM secured itself a position of power in the industry with the development of its 700/7000 series, based on vacuum tubes and transistors, and with its 360 series mainframe.  Unveiled in 1964, the 360 series was to be an all-around computer system, a series of compatible models for purposes both scientific and commercial – a series which, moreover, brought together features which were once only available in scientific or commercial computers, such as floating point arithmetic in the former and decimal arithmetic and byte addressing in the latter.  The 360 series also included supervisor and application mode programs and instructions and built-in memory protection facilities, making it one of the first computers manufactured with provisions specific to the use of an operating system.

Console of an IBM 360/67 mainframe

PCs

As the demand for the older mainframe systems fell off, new installations were seen mainly in the realms of finance and the government.  Personal computer networks came to challenge the mainframe.  It was during the rise of personal computing networks, though, that the APIs with which we are most familiar came into being and, in the case of Windows, achieved dominance. 

In 1975, the Altair 8800 was introduced in Popular Electronics, a personal computer that was affordable, user-friendly, and, some argue, the spark that set Apple Computer and Microsoft ablaze in their development of personal computers. The Apple II, though less capable and versatile than some of the larger computers of the day, gave computer enthusiasts an environment in which to develop their own programming skills and to operate simple office and productivity applications. 

The IBM PC released in 1981 took the personal computer into the realm of business, giving individual users word processing programs, spreadsheet programs and database programs which would change the way businesses stored, sorted and used their data. Four years later in 1985, in order to compete with the graphical user interfaces made popular by Apple, Microsoft released an add-on to MS-DOS – an operating environment known as Windows.

Though that release of Windows was not an operating system in the full sense of the term, it had pushed beyond the characteristics of a typical desktop, adopting some functions of operating systems.  Windows achieved a leg up on competing systems due in large part to the fact that MS-DOS dominated the early landscape of personal computing.  But the dominance of Windows (up until Google that is) is the API.  The APIs which enabled professional programmers to develop desktop applications on top of platforms (perhaps most notably the Microsoft Windows API), have now given way to APIs which feed off of the platform of the Internet.  And while Microsoft and the desktop are controlled by physical bodies, the Internet, despite the fact that certain companies do, in fact, oversee enormous pools of user data and have the ability to direct traffic as they see fit, is not governed by a particular body or set of bodies.  If the power flow of yesterday’s APIs was a vertical one, headed at top by the executives of companies like Microsoft, which allowed programmers to work off of their platform to develop applications to be used by the users at the  bottom, we might see the power flow of today’s APIs as closer to a horizontal one.   

Next:  The Thrilling Poverty of Physical Gestures