Why the Castles of Silicon Valley are Built out of Sand

Ambrogio_Lorenzetti Temperance with an hour glass Allegory of Good Government

If you get just old enough, one of the lessons living through history throws you is that dreams take a long time to die. Depending on how you date it, communism took anywhere from 74 to 143 years to pass into the dustbin of history, though some might say it is still kicking. The Ptolemaic model of the universe lasted from 100 AD into the 1600’s. Perhaps even more dreams than not simply refuse to die, they hang on like ghost, or ghouls, zombies or vampires, or whatever freakish version of the undead suits your fancy. Naming them would take up more room than I can post, and would no doubt start one too many arguments, all of our lists being different. Here, I just want to make an argument for the inclusion of one dream on our list of zombies knowing full well the dream I’ll declare dead will have its defenders.

The fact of the matter is, I am not even sure what to call the dream I’ll be talking about. Perhaps, digitopia is best. It was the dream that emerged sometime in the 1980’s and went mainstream in the heady 1990’s that this new thing we were creating called the “Internet” and the economic model it permitted was bound to lead to a better world of more sharing, more openness, more equity, if we just let its logic play itself out over a long enough period of time. Almost all the big-wigs in Silicon Valley, the Larry Pages and Mark Zuckerbergs, and Jeff Bezos(s), and Peter Diamandis(s) still believe this dream, and walk around like 21st century versions of Mary Magdalene claiming they can still see what more skeptical souls believe has passed.

By far, the best Doubting Thomas of digitopia we have out there is Jaron Lanier. In part his power in declaring the dream dead comes from the fact that he was there when the dream was born and was once a true believer. Like Kevin Bacon in Hollywood, take any intellectual heavy hitter of digital culture, say Marvin Minsky, and you’ll find Lanier having some connection. Lanier is no Luddite, so when he says there is something wrong with how we have deployed the technology he in part helped develop, it’s right and good to take the man seriously.

The argument Lanier makes in his most recent book Who Owns the Future? against the economic model we have built around digital technology in a nutshell is this: what we have created is a machine that destroys middle class jobs and concentrates information, wealth and power. Say what? Hasn’t the Internet and mobile technology democratized knowledge? Don’t average people have more power than ever before? The answer to both questions is no and the reason why is that the Internet has been swallowed by its own logic of “sharing”.

We need to remember that the Internet really got ramped up when it started to be used by scientists to exchange information between each other. It was built on the idea of openness and transparency not to mention a set of shared values. When the Internet leapt out into public consciousness no one had any idea of how to turn this sharing capacity and transparency into the basis for an economy. It took the aftermath of dot com bubble and bust for companies to come up with a model of how to monetize the Internet, and almost all of the major tech companies that dominate the Internet, at least in America- and there are only a handful- Google, FaceBook and Amazon, now follow some variant of this model.

The model is to aggregate all the sharing that the Internet seems to naturally produce and offer it, along with other “compliments” for “free” in exchange for one thing: the ability to monitor, measure and manipulate through advertising whoever uses their services. Like silicon itself, it is a model that is ultimately built out of sand.

When you use a free service like Instagram there are three ways its ultimately paid for. The first we all know about, the “data trail” we leave when using the site is sold to third party advertisers, which generates income for the parent company, in this case FaceBook. The second and third ways the service is paid for I’ll get to in a moment, but the first way itself opens up all sorts of observations and questions that need to be answered.

We had thought the information (and ownership) landscape of the Internet was going to be “flat”. Instead, its proven to be extremely “spiky”. What we forgot in thinking it would turn out flat was that someone would have to gather and make useful the mountains of data we were about to create. The big Internet and Telecom companies are these aggregators who are able to make this data actionable by being in possession of the most powerful computers on the planet that allow them to not only route and store, but mine for value in this data. Lanier has a great name for the biggest of these companies- he calls them Siren Servers.

One might think whatever particular Siren Servers are at the head of the pack is a matter of which is the most innovative. Not really. Rather, the largest Siren Servers have become so rich they simply swallow any innovative company that comes along. FaceBook gobbled up Instagram because it offered a novel and increasingly popular way to share photos.

The second way a free service like Instagram is paid for, and this is one of the primary concerns of Lanier in his book, is that it essentially cannibalizes to the point of destruction the industry that used to provide the service, which in the “old economy” meant it also supported lots of middle class jobs.

Lanier states the problem bluntly:

 Here’s a current example of the challenge we face. At the height of its power, the photography company Kodak employed more than 140,000 people and was worth $28 billion. They even invented the first digital camera. But today Kodak is bankrupt, and the new face of digital photography is Instagram. When Instagram was sold to FaceBook for a billion dollars in 2012, it employed only thirteen people.  (p.2)

Calling Thomas Piketty….

As Bill Davidow argued recently in The Atlantic the size of this virtual economy where people share and get free stuff in exchange for their private data is now so big that it is giving us a distorted picture of GDP. We can no longer be sure how fast our economy is growing. He writes:

 There are no accurate numbers for the aggregate value of those services but a proxy for them would be the money advertisers spend to invade our privacy and capture our attention. Sales of digital ads are projected to be $114 billion in 2014,about twice what Americans spend on pets.

The forecasted GDP growth in 2014 is 2.8 percent and the annual historical growth rate of middle quintile incomes has averaged around 0.4 percent for the past 40 years. So if the government counted our virtual salaries based on the sale of our privacy and attention, it would have a big effect on the numbers.

Fans of Joseph Schumpeter might see all this churn as as capitalism’s natural creative destruction, and be unfazed by the government’s inability to measure this “off the books” economy because what the government cannot see it cannot tax.

The problem is, unlike other times in our history, technological change doesn’t seem to be creating many new middle class jobs as fast as it destroys old ones. Lanier was particularly sensitive to this development because he always had his feet in two worlds- the world of digital technology and the world of music. Not the Katy Perry world of superstar music, but the kinds of people who made a living selling local albums, playing small gigs, and even more importantly, providing the services that made this mid-level musical world possible. Lanier had seen how the digital technology he loved and helped create had essentially destroyed the middle class world of musicians he also loved and had grown up in. His message for us all was that the Siren Servers are coming for you.

The continued advance of Moore’s Law, which, according to Charlie Stross, will play out for at least another decade or so, means not so much that we’ll achieve AGI, but that machines are just smart enough to automate some of the functions we had previously thought only human beings were capable of doing. I’ll give an example of my own. For decades now the GED test, which people pursue to obtain a high school equivalency diploma, has had an essay section. Thousands of people were necessary to score these essays by hand, the majority of whom were likely paid to do so. With the new, computerized GED test this essay scoring has now been completely automated, human readers made superfluous.

This brings me to the third way this new digital capabilities are paid for. They cannibalize work human beings have already done to profit a company who presents and sells their services as a form of artificial intelligence. As Lanier writes of Google Translate:

It’s magic that you can upload a phrase in Spanish into the cloud services of a company like Google or Microsoft, and a workable, if imperfect, translation to English is returned. It’s as if there’s a polyglot artificial intelligence residing up there in that great cloud of server farms.

But that is not how cloud services work. Instead, a multitude of examples of translations made by real human translators are gathered over the Internet. These are correlated with the example you send for translation. It will almost always turn out that multiple previous translations by real human translators had to contend with similar passages, so a collage of those previous translations will yield a usable result.

A giant act of statistics is made virtually free because of Moore’s Law, but at core the act of translation is based on real work of people.

Alas, the human translators are anonymous and off the books. (19-20)

The question all of us should be asking ourselves is not “could a machine be me?” with all of our complexity and skills, but “could a machine do my job?” the answer to which, in 9 cases out of 10, is almost certainly- “yes!”

Okay, so that’s the problem, what is Lanier’s solution? His solution is not that we pull a Ned Ludd and break the machines or even try to slow down Moore’s Law. Instead, what he wants us to do is to start treating our personal data like property. If someone wants to know my buying habits they have to pay a fee to me the owner of this information. If some company uses my behavior to refine their algorithm I need to be paid for this service, even if I was unaware I had helped in such a way. Lastly, anything I create and put on the Internet is my property. People are free to use it as they chose, but they need to pay me for it. In Lanier’s vision each of us would be the recipients of a constant stream of micropayments from Siren Servers who are using our data and our creations.

Such a model is very interesting to me, especially in light of other fights over data ownership, namely the rights of indigenous people against bio-piracy, something I was turned on to by Paolo Bacigalupi’s bio-punk novel The Windup Girl, and what promises to be an increasing fight between pharmaceutical/biotech firms and individuals over the use of what is becoming mountains of genetic data. Nevertheless, I have my doubts as to Lanier’s alternative system and will lay them out in what follows.

For one, such a system seems likely to exacerbate rather than relieve the problem of rising inequality. Assuming most of the data people will receive micropayments for will be banal and commercial in nature, people who are already big spenders are likely to get a much larger cut of the micropayments pie. If I could afford such things it’s no doubt worth a lot for some extra piece of information to tip the scales between me buying a Lexus or a Beemer, not so much if it’s a question of TIDE vs Whisk.

This issue would be solved if Lanier had adopted the model of a shared public pool of funds where micropayments would go rather than routing them to the actual individual involved, but he couldn’t do this out of commitment to the idea that personal data is a form of property. Don’t let his dreadlocks fool you, Lanier is at bottom a conservative thinker. Such a fee might balance out the glaring problem that Siren Servers effectively pay zero taxes

But by far the biggest hole in Lanier’s micropayment system is that it ignores the international dimension of the Internet. Silicon Valley companies may be barreling down on their model, as can be seen in Amazon’s recent foray into the smartphone market, which attempts to route everything through itself, but the model has crashed globally. Three events signal the crash, Google was essentially booted out of China, the Snowden revelations threw a pale of suspicion over the model in an already privacy sensitive Europe, and the EU itself handed the model a major loss with the “right to be forgotten” case in Spain.

Lanier’s system, which accepts mass surveillance as a fact, probably wouldn’t fly in a privacy conscious Europe, and how in the world would we force Chinese and other digital pirates to provide payments of any scale? And China and other authoritarian countries have their own plans for their Siren Servers, namely, their use as tools of the state.

The fact of the matter is their is probably no truly global solution to continued automation and algorithmization, or to mass surveillance. Yet, the much feared “splinter-net”, the shattering of the global Internet, may be better for freedom than many believe. This is because the Internet, and the Siren Servers that run it, once freed from its spectral existence in the global ether, becomes the responsibility of real territorially bound people to govern. Each country will ultimately have to decide for itself both how the Internet is governed and define its response to the coming wave of automation. There’s bound to be diversity because countries are diverse, some might even leap over Lanier’s conservativism and invent radically new, and more equitable ways of running an economy, an outcome many of the original digitopians who set this train a rollin might actually be proud of.

 

Advertisements

Maps:how the physical world conquered the virtual

World map 1600

If we look back to the early days when the Internet was first exploding into public consciousness, in the 1980’s, and even more so in the boom years of the 90’s, what we often find is a kind of utopian sentiment around this new form of “space”. It wasn’t only that a whole new plane of human interaction seemed to be unfolding into existence almost overnight, it was that “cyberspace” seemed poised to swallow the real world- a prospect which some viewed with hopeful anticipation and others with doom.

Things have not turned out that way.

The person who invented the term “cyberspace”, William Gibson, the science fiction author of the classic- Neuromancer- himself thinks that when people look back on the era when the Internet emerged what will strike them as odd is how we could have confused ourselves into thinking that the virtual world and our work-a-day one were somehow distinct. Gibson characterizes this as the conquest of the real by the virtual. Yet, one can see how what has happened is better thought of as the reverse by taking even a cursory glance at our early experience and understanding of cyberspace.

Think back, if you are old enough, and you can remember, when the online world was supposed to be one where a person could shed their necessarily limited real identity for a virtual one. There were plenty of anecdotes, not all of them insidious, of people faking their way through a contrived identity the unsuspecting thought was real: men coming across as women, women as men, the homely as the beautiful. Cyberspace seemed to level traditional categories and the limits of geography. A poor adolescent could hobnob with the rich and powerful. As long as one had an Internet connection, country of origin and geographical location seemed irrelevant.

It should not come as any surprise, then, that  an early digital reality advocate such as Nicole Stenger could end her 1991 essay Mind is a leaking rainbow with the utopian flourish:

According to Satre, the atomic bomb was what humanity had found to commit collective suicide. It seems, by contrast, that cyberspace, though born of a war technology, opens up a space for collective restoration, and for peace. As screens are dissolving, our future can only take on a luminous dimension! / Welcome to the New World! (58)

Ah, if only.

Even utopian rhetoric was sometimes tempered with dystopian fears. Here is Mark Pesce the inventor of VRML code in his 1997 essay Ignition:

The power over this realm has been given to you. You are weaving the fabric of perception in information perceptualized. You could – if you choose – turn our world into a final panopticon – a prison where all can been seen and heard and judged by a single jailer. Or you could aim for its inverse, an asylum run by the inmates. The esoteric promise of cyberspace is of a rule where you do as you will; this ontology – already present in the complex system know as Internet – stands a good chance of being passed along to its organ of perception.

The imagery of a “final panopticon” is doubtless too morbid for us at this current stage whatever the dark trends. What is clear though is that cyberspace is a dead metaphor for what the Internet has become- we need a new one. I think we could do worse than the metaphor of the map. For, what the online world has ended up being is less an alternative landscape than a series of cartographies by which we organize our relationship with the world outside of our computer screens, a development with both liberating and troubling consequences.

Maps have always been reflections of culture and power rather than reflections of reality. The fact that medieval maps in the West had Jerusalem in their centers wasn’t expressing a geologic but a spiritual truth although few understood the difference. During the Age of Exploration what we might think of as realistic maps were really navigational aids for maritime trading states, a latent fact present in what the mapmakers found important to display and explain.

The number and detail of maps along with the science of cartography rose in tandem with the territorial anchoring of the nation-state. As James C. Scott points out in his Seeing Like a State maps were one of the primary tools of the modern state whose ambition was to make what it aimed to control “legible” and thus open to understanding by bureaucrats in far off capitals and their administration.

What all of this has to do with the fate of cyberspace, the world where we live today, is that the Internet, rather than offering us an alternative version of physical space and an escape hatch from its problems has instead evolved into a tool of legibility. What is made legible in this case is us. Our own selves and the micro-world’s we inhabit have become legible to outsiders. Most of the time these outsiders are advertisers who target us based on our “profile”, but sometimes this quest to make individuals legible is by the state- not just in the form of standardized numbers and universal paperwork but in terms of the kinds of information a state could only once obtain by interrogation- the state’s first crack at making individuals legible.      

A recent book by Google CEO Eric Schmitt co-authored with foreign policy analyst Jared Cohen- The New Digital Age is chalk full of examples of corporate advertisers’ and states’ new powers of legibility. They write:

The key advance ahead is personalization. You’ll be able to customize your devices- indeed much of the technology around you- to fit your needs, so that the environment reflects your preferences.

At your fingertips will be an entire world’s worth of digital content, constantly updated, ranked and categorized to help you find the music, movies, shows, books, magazines, blogs and art you like. (23)

Or as journalist Farhad Manjoo quotes Amit Singhal of Google:

I can imagine a world where I don’t even need to search. I am just somewhere outside at noon, and my search engine immediately recommends to me the nearby restaurants that I’d like because they serve spicy food.

There is a very good reason why I did not use the world “individuals” in place of “corporate advertisers” above- a question of intent. Whose interest does the use of such algorithms to make the individual legible ultimately serve? If it my interest then search algorithms might tell me where I can get a free or even pirated copy of the music, video etc I will like so much. It might remind me of my debts, and how much I would save if I skip dinner at the local restaurant and cook my quesadillas at home. Google and all its great services, along with similar tech giants aiming to map the individual such as FaceBook aren’t really “free”. While using them I am renting myself to advertisers. All maps are ultimately political.

With the emergence mobile technology and augmented reality the physical world has wrestled the virtual one to the ground like Jacob did to the angel. Virtual reality is now repurposed to ensconce all of us in our own customized micro-world. Like history? Then maybe your smartphone or Google Glasses will bring everything historical around you out into relief. Same if you like cupcakes and pastry or strip clubs. These customized maps already existed in our own heads, but now we have the tools for our individualized cartography- the only price being constant advertisements.

There’s even a burgeoning movement among the avant garde, if there can still be said to be such a thing, against this kind of subjection of the individual to corporate dictated algorithms and logic. Inspired by mid-20 century leftists such as Guy Debord with his Society of the Spectacle practitioners of what is called psychogeography are creating and using apps such as Drift  that lead the individual on unplanned walks around their own neighborhoods, or Random GPS that have your car’s navigation system remind you of the joys of getting lost.

My hope is that we will see other versions of these algorithm inverters and breakers and not just when it comes to geography. How about similar things for book recommendations or music or even dating? We are creatures that sometimes like novelty and surprise, and part of the wonder of life is fortuna–  its serendipitous accidents.

Yet, I think these tools will most likely ramp up the social and conformist aspects of our nature. We shouldn’t think they will be limited to corporate persuaders. I can imagine “Catholic apps” that allow one to monitor one’s sins, and a whole host of funny and not so funny ways groups will use the new methods of making the individual legible to tie her even closer to the norms of the group.

A world where I am surrounded by a swirl of constant spam, or helpful and not so helpful suggestions, the minute I am connected, indeed, a barrage that never ends except when I am sleeping because I am always connected, may be annoying, but it isn’t all that scary. It’s when we put these legibility tools in the hands of the state that I get a little nervous.

As Schmitt and Cohen point out one of the most advanced forms of such efforts at mapping the individual is an entity called Platforma Mexico which is essentially a huge database that is able to identify any individual and tie them to their criminal record.

Housed in an underground bunker in the Secretariat of Public Security compound in Mexico City, this large database integrates intelligence, crime reports and real time data from surveillance cameras and other inputs from across the country. Specialized algorithms can extract patterns, project social graphs and monitor restive areas for violence and crime as well as for natural disasters and other emergencies.  (174)

The problem I have here is the blurring of the line between the methods used for domestic crime and those used for more existential threats, namely- war. Given that crime in the form of the drug war is an existential threat for Mexico this might make sense, but the same types of tools are being perfected by authoritarian states such as China, which is faced not with an existential threat but with growing pressures for reform, and also in what are supposed to be free societies like the United States where a non-existential threat in the form of terrorism- however already and potentially horrific- is met with similar efforts by the state to map individuals.

Schmitt and Cohen point out how there is a burgeoning trade between autocratic countries and their companies which are busy perfecting the world’s best spyware. An Egyptian firm Orascom owns a 25 percent share of the panopticonic sole Internet provider in North Korea. (96) Western companies are in the game as well with the British Gamma Group’s sale of spyware technology to Mubarak’s Egypt being just one recent example.

Yet, if corporations and the state are busy making us legible there has also been a democratization of the capacity for such mapmaking, which is perhaps the one of the reasons why states are finding governance so difficult. Real communities have become almost as easy to create as virtual ones because all such communities are merely a matter of making and sustaining human relationships and understanding their maps.

Schmitt and Cohen imagine virtual governments in exile waiting in the wings to strike at the precipitous moment. Political movements can be created off the shelf supported by their own ready made media entities and the authors picture socially conscious celebrities and wealthy individuals running with this model in response to crises. Every side in a conflict can now have its own media wing whose primary goal is to shape and own the narrative. Even whole bureaucracies could be preserved from destruction by keeping its map and functions in the cloud.

Sometimes virtual worlds remain limited to the way they affect the lives of individuals but are politically silent. A popular mass multiplayer game such as World of Warcraft may have as much influence on an individual’s life as other invisible kingdoms such as those of religion. An imagined online world becomes real the moment its map is taken as a prescription for the physical world.  Are things like the Hizb ut-Tahrir which aims at the establishment of a pan-Islamic caliphate or The League of the South which promotes a second secession of American states “real” political organizations or fictional worlds masking themselves as political movements? I suppose only time will tell.

Whatever the case, society seems torn between the mapmakers of the state who want to use the tools of the virtual world to impose order on the physical and an almost chaotic proliferation using the same tools by groups of all kinds creating communities seemingly out of thin air.

All this puts me in mind of nothing so much as China Mieville’s classic of New Weird fiction City and the City. It’s a crime novel with the twist that it takes place in two cities- Beszel  and Ul Qoma that exist in the same physical space and are superimposed on top of one another. No doubt Mieville was interested in telling a good story, and getting us thinking about the questions of borders and norms, but it’s a pretty good example of the mapping I’ve been talking about- even if it is an imagined one.

In City and the City an inhabitant of Beszel  isn’t allowed to see or interact with what’s going on in Ul Qoma and vice versa otherwise they commit a crime called “breach” and there’s a whole secretive bi-city agency called Breach that monitors and prosecutes those infractions. There’s even an imaginary (we are led to believe) third city “Orciny” that exist on-top of Beszel and Ul Qoma and secretly controls the other two.

This idea of multiple identities- consumer, political- overlaying the same geographical space seems a perfect description of our current condition. What is missing here, though, is the sharp borders imposed by Breach. Such borders might appear quicker and in different countries than one might have supposed thanks to the recent revelations that the United States has been treating the Internet and its major American companies like satraps. Only now has Silicon Valley woken up to the fact that its close relationship with the American security state threatens its “transparency” based business- model with suicide. The re-imposition of state sovereignty over the Internet would mean a territorialization of the virtual world- a development that would truly constitute its conquest by the physical. To those possibilities I will turn next time…