by Jaron Lanier
When Are Siren Servers Monopolies?
As explained in the sections on network effects, when users put effort, money, or important data into a particular service, like a social network, then network effects tend to create a single Sirenic presence, a monopoly for that particular kind of data or pattern of use.
Many Siren Servers of this kind are subject to something like a Pauli exclusion principle, or if you prefer, they tend to discover and uniquely occupy pseudo-monopolistic roles. There can’t be both a Friendster and a Facebook. One of them must win.
When a Siren Server is more a mediator than an accumulator of primary data sources, then it can have company. There can be multiple travel sites, as explained earlier, because they don’t own the primary reservation-related data that they mediate. There can be multiple Sirenic financial services because none of them own Wall Street.
Similarly, there can be both a Bing and Google, since neither owns the Web. To be more precise, there can be two search engines,* but Google still tends to be monopoly-like in selling advertising based on search, which is a different matter. That is because, as an accumulator of advertiser relationships, Google does enjoy a monopoly-like network effect.
*This observation only applies to traditional personal computers. On mobile phones, Google generally enjoys a structural advantage because of preferred placement.
Another example is that Amazon and Barnes & Noble can coexist as booksellers, because they don’t own the books, but if they also become major publishers, then one would probably have to kill the other.
Sometimes potential Sirenic monopoly is blocked because of a structural or legal blockade that limits reach. For instance, a language barrier might limit a social network to certain regions of the world, or a mobile carrier might be able to capture users by contract instead of through pure data effects.
Even when there is only one Siren Server to a niche, there can be a lot of niches, however.
Free Rise
What’s the threshold for rewarding network effects to kick in? For consumer-facing sites, it is the point at which enough people are using a site to support each other’s expectations of dynamism. An additional threshold is that a critical number of people have to stick together long enough so that the site becomes a habit for them. Then the dynamism won’t decay.
It’s not as if there is no technical requirement at all for a site to catch a wave and become huge. The site generally has to be at least consistently available, though in its early years Twitter wasn’t.
Once you reach a critical point, you have a population or two locked in. You might very well grow to global proportions and exert influence like a messiah, tweaking the design of human experience at large.
If you’ve made it to the point that growth is accelerating, you’ve entered the honeymoon phase, or free rise (the opposite of free fall). Some entrepreneurs promote like crazy during this phase, while others are just consumed with keeping the thing running. If you want free rise to continue until your Siren Server becomes a monster, you’ll have to attend to a few things . . .
If the first phase goes well, you can experience an amazing lift, as you aggregate connections and data at an intense clip. During this period, all the usual rules of life and commerce are suspended. It’s free rise, and anything can happen.
During free rise, you can see patterns in data no one else can see, as if you were an oracle. You will suddenly know more about some slice of human life than anyone else. Maybe you’ll see something about eating habits, sex, shopping, or driving patterns.
A few of the folks you have aggregated will inevitably get an insane lift from being hitched to you, and they’ll create even more excitement. An early investor in your fund will get superrich superfast, or a user of your free service will earn a windfall from sudden exposure. This will happen to only a tiny token number of people, though. It is really you, the proprietor of the Siren Server, who will benefit above all others.
At first, all you’ll have is rewarding network effect. That means that people will benefit from using your server because other people are using it. A virtuous cycle causes more and more people to use your offering. That’s not enough, however, if you want to build a world-class, persistent Siren Server. In addition, you have to inject some sort of punishing network effect.
Make Others Pay for Entropy
Once both rewarding and punishing network effects are taking hold, another crucial task is to make sure that risk is being radiated out to other people and institutions, and not accruing to your server. Sites like Pinterest invariably demand that users click through an agreement that places all responsibility for copyright violations or anything else squarely on the user.
If people are paying money to use your server, don’t accept any of it directly if you can possibly avoid that. You should be a broker between buyers and sellers to the degree that’s possible. You can then earn commissions, placement fees, visibility fees, or any number of other fees yet to be conceived, but without taking any responsibility for the actual events that took place.
Make both buyers and sellers click through agreements that make them, not you, take on all liabilities. These click-through agreements are the grandiosely verbose descendants of the Zen koan about a tree falling in a forest that no one hears. No one will read them, so they are very unlikely to be tested in a legal proceeding. No one wants to read them, not even lawyers. Some lawyer at the Electronic Frontier Foundation or some such place might occasionally be able to make it through one of them, but that is rare. Since they are unread, they basically do not exist, except for setting the basic rule everyone understands, which is that the server takes no risks, only the users of the server. The ideal is for click-through agreements to remain unread until your server becomes so huge that it’s scary.
This principle applies doubly if you are running a Wall Street fund on your server instead of a Silicon Valley startup. The ideal Siren Server is one for which you make no specific decisions. You should do everything possible to not do anything consequential. Don’t play favorites; don’t have taste. You are to be the neutral facilitator, the connector, the hub, but never an agent who could be blamed for a decision. Reduce the number of decisions that can be pinned on you to an absolute minimum.
What you can do, however, is pattern how other people make decisions. You can get people to have less privacy or organize a business around coupons, but you never get into the middle of any specific event within the pattern template you’ve created for other people to use.
Bills Are Boring
It isn’t free to run a Siren Server. You will need to hire some of those fabled PhDs from MIT or Caltech sooner or later, and pay the storage and connectivity bills. For the Silicon Valley startup variety of Siren Server, this brings up the question of monetization. It isn’t polite or cool to think about monetization very early in the game. Have some faith, man! Information always turns to money, somehow, sooner or later.
Money usually doesn’t flow much into a Siren Server in its earliest phases, but fortunately they’re cheap to run. You can outsource much of the nontechnical heavy lifting that might come along to the peasants who populate Mechanical Turk and similar services. “The entire cost of running this business on the human side is incredibly low,” says Keith Rabois, chief operating officer of Square, an up-and-coming Siren Server that hopes to become the router of choice for consumer credit cards.1
Coattails
Fleeting success sporadically flares for a lucky few players on the sidelines of Siren Servers, as if by magic. These rags-to-riches tales are 21st century echoes of the famous Horatio Alger tales of the 19th century, in which unlikely underdogs worked hard and found great success—except the hard work part is no longer a given.
No end of ridicule has been directed at Horatio Alger stories ever since they first appeared, because they build false hope. They are deceptive; even when the original story is true, it is vanishingly unlikely that any particular individual will find sim
ilar success by pursuing similar strategies. Horatio’s algebra offered horrible ratios. Theater can’t replace a functional economy. An economy can’t grow authentically if it is too much like a casino.
There are actually two types of 21st century Horatio Alger story that spark off of Siren Servers. One type is the occasional “viral” success. This might occur on YouTube, for instance. Every now and then a sympathetic individual will achieve visibility, and perhaps even income. But in the longer term, such people generally must seek to leverage Internet visibility in a gambit to find success in old media, which is being shrunk by the Internet at the same time. For instance, TV talent shows like The Voice attract YouTube stars, even though their audience on YouTube might have been larger. At least old media, even in its decline and with all its problems, still offers a career path.
A more disturbing version of deceptive viral success is the occasional charity effort on sites like Reddit. A sympathetic figure in need will reach the hearts of a large audience and get some help, usually in the form of many small donations. On the one hand, in each case this is wonderful, and yet it’s ultimately a way for people to feel good while having achieved nothing, in statistical terms.
However, viral success is small-time compared to another, more rarefied kind of rags-to-riches tale. When a Siren Server is on the rise, in a honeymoon phase, a small-time player might just score a once-in-a-lifetime spectacular lift.
The Web itself went through a honeymoon phase around the turn of the century, which is now remembered as the “dot-com bubble.” During this time there were weird and wild successes that motivated a stampede of hopefuls. My favorite story was of a young woman who drove up her credit card debt and then created a website asking for donations from strangers to pay off her debt, for no reward and with no real explanation. It worked, but as I recounted in Gadget, that was only because her timing was accidentally perfect. She caught the wave at the right moment, and none of the many copycats who followed her could duplicate her success. Of course it would be absurd to think that any could.
Her success was due to the fact that she had a rapt audience for that fleeting magic moment when a network is gaining its network effect but before all the hapless scammers of the world rush in to dilute the radiated benefit of it. She was like the first person to arrive in California for the gold rush, when the gold was still visible, strewn on the surface.
During the honeymoon phase of a newly successful Siren Server, a lucky few people will typically be gifted with astounding, deceptive success. Their stories will be celebrated, creating a distorted popular perception of opportunities.
For instance, musician Amanda Palmer launched a Kickstarter campaign in 2012 that became legendary. She stated a goal of raising a hundred thousand dollars to support a new release and tour, but instead raised more than a million dollars. There were other similar tales during that honeymoon year for Kickstarter. (Good for her! I would like to believe this type of success will become unremarkable, but as things stand, I won’t hold my breath. World, please prove me wrong!)
An interesting psychological phenomenon, when a modern Horatio Alger–like hero hits the jackpot, is that she might succumb to the illusion, for a moment, that she has achieved Siren Server status herself. Palmer promptly asked for free labor from musician fans on the same tour that had been luxuriously funded by those same fans. Needless to say, the Mechanical Turk effect came into play, and free labor presented itself. In response to a torrent of criticism from professional musicians, Palmer relented and announced a plan to pay musicians. (Good for her, once again! But this doesn’t suggest a societal solution.)
There is always a tale of someone making it big or changing their life because they caught a digital wave at just the right moment. If only there were enough of those moments generated by the regime of Siren Servers to support a society.
The Closing Act
How do Siren Servers die? We don’t know as much as we will someday soon, since the phenomenon is still new. One can imagine Wal-Mart being overwhelmed by Amazon, for instance, simply because Amazon is even more computational than Wal-Mart.
Amazon, by using superior computation, might potentially piggyback on Wal-Mart’s legacy of supply chain optimization, and essentially aggregate Wal-Mart’s efficiencies into its own. It wasn’t Amazon that brought about all the cheaply available goods, but by having the best spy data at a given time, Amazon might become the concern that benefits the most from them.
Maybe some other Siren Server related to self-driving vehicles will come along and be able to make itself meta to Amazon and then steal whatever advantage Amazon builds up, since Amazon depends on vehicles to deliver goods to consumers. Competition becomes mostly about who can out-meta who, and only secondarily about specialization.
We do know that Siren Servers can die. It happened to Lehman Brothers.
They are not like classical monopolies that can become so entrenched that a regulator might choose to break them up. How would you break up Facebook? Into one for fake hot babes and another for political organizing? The idea is absurd.
Individual Siren Servers can die and yet the Siren Server pattern perseveres, and it is that pattern that is the real problem. The systematic decoupling of risk from reward in the rising information economy is the problem, not any particular server.
Stories Are Nothing Without Ideas
This book proposes a grand future story in which the pattern of Siren Servers will be superseded by a more inclusive new pattern. But even today, it would be a mistake to only see chaos and meaninglessness in the crazed energies of a networked world.
The endgames of contests between Siren Servers are not meaningless. Siren Servers are not interchangeable. While they all share certain traits (narcissism, hyperamplified risk aversion, and extreme information asymmetry), they also represent particular, more specialized philosophies. The requirements of being a Siren Server leave enough room for variation that contests between them can also be collisions of contrasting ideas.
Facebook suggests not only a moral imperative to place certain information in its network, but the broad applicability of one template to compare people. In this it is distinct from Google, which encourages semistructured online activity that Google will be best at organizing after the fact.
Twitter suggests that meaning will emerge from fleeting flashes of thought contextualized by who sent the thought rather than the content of the thought. In this it is distinct from Wikipedia, which suggests that flashes of thought be inserted meaningfully into a shared semantic structure. Wikipedia proposes that knowledge can be divorced from point of view. In this it is distinct from the Huffington Post, where opinions fluoresce.
In all these cases, unquestionably big ideas are at play. The designs of these sites are embodiments of philosophies about what a person is, where meaning comes from, the nature of freedom, and the nature of an ideal society. When Master Servers die, the associated ideas can be suppressed for long periods of time, which is as close as ideas get to death.
The blog TechCrunch keeps a grim count of failed Silicon Valley efforts, called the Deadpool.2 In it we find not only would-be Siren Servers that died, but also an early hint at how ideas might be lost along with them.
For instance, Google tried to get a new Siren Server going, to tremendous fanfare, called Wave.3 It proposed that conversations between people could be highly structured from the start, to make the content of the conversation more valuable later. That would mean that meaning in natural language would be preserved even if everything said had to be fit initially into a particular, tree-shaped data structure. That in turn suggests a level of meaning in human conversation that is more orderly and tiered, more Chomskyan, than has ever been isolated by researchers before. It’s a major assertion about what meaning is, or might become. (I am skeptical that the idea is correct, but that is beside the point.)
Since the effort was about big business data instead of big science data, we can’t say that the idea s
uffered a Popperian* disqualification. Instead the idea was attached to a server that failed. With the death of a Siren Server, a distinct sensibility concerning human meaning and how we communicate with one another became effectively dead and unexplored for now.4
*Karl Popper was an Australian philosopher who famously described how science never achieves absolute, eternal truth, but instead gets closer and closer to truth by disqualifying false ideas. Mathematics, on the other hand, does include a concept of absolute, eternal truth, because of proofs.
A networked story is just as much a contest of ideas as was the Cold War, which served as a standard of meaning for Kushner and Fukuyama. Story lives, and the future is not random.
FIFTH INTERLUDE
The Wise Old Man in the Clouds
THE LIMITS OF EMERGENCE AS AN EXPLANATION
In 2012, the University of San Francisco, a Jesuit institution, themed its recruitment campaign on the idea that Christianity is like Facebook. One of the slogans was “Our CEO mastered social networking 2,000 years before Mark Zuckerberg was born.”1
There’s something to the comparison, and I find that worrisome. Each institution became powerful in an unconventional way. Each network created a center of power that bypassed territorial and political boundaries, and existed on its own plane. Each became what might be called a “social monopoly,” engaging in social engineering on a grand scale.
That’s not to say that bad things will necessarily happen in a social monopoly. They can achieve breathtaking large-scale social good. The Catholic Church unquestionably educates many millions of the poor, heals many millions of the sick, stabilizes many millions of families, and comforts many millions of the dying; in 2012 Facebook dipped a toe into the waters of social engineering by increasing the rolls of organ donors with a simple tweak of its user experience. By putting the option to donate right in front of people, many more people embraced that option.