Increasing Returns and the New World of Hindsight

The theory of increasing hindsight states that the more you grow, the more you (think you) know.

“Increasing returns doesn’t argue that bigness is wrong. It doesn’t argue that Microsoft is evil. It’s simply an economics that concerns itself with what happens in markets where there are advantages to market volume and market share. The theory itself is neutral.” — W. Brian Arthur

W. Brian Arthur is too nice. But I’m neither professional economist nor reputable so I don’t have to be. I also know a thing or two about building social networks that don’t take off. Could I have been the next Mark Zuckerburg? Unfortunately, time won’t tell, because as far as I know, when John Cusack (I want to say Cossack) and his gang of underachieving rabble-rousers got together to go skiing, that was just a movie.

Hot Tub Time Machine (2010)

No, I’m not inviting you to a hot tub party. I am, however, about to present the final nail in the coffin for any would-be stragglers who fail to see that Silicon Valley represents the largest threat of sustained monopoly the world has ever seen. That nail is called non-ergodicity, and for all you who are unfamiliar with modern physics, this will be a real eye opener.

Whether a system, such as a mixture in a beaker, a food web, or the economy is ergodic depends on the amount of time it takes to return to a previously experienced state. If it takes more time than physically allowable, i.e. before the Universe ends, then you’ve got yourself a non-ergodic system. Basically, the Universe will not make all possible proteins, all possible art, or all possible technology. It will only produce a small subset of each, and this means a lot when it comes to matters of prediction.

It turns out that for sufficiently complex systems, such as the biosphere, the econosphere, and the human brain, history is unrepeatable over such timescales, meaning that there are vastly more possible system permutations than there is physical time to access any one system in its entirety (simulating 1 second of human brain activity takes 82,944 processors circa 2016, and that says nothing about consciousness, emotion, or art).

If you had a time machine, however, you could conceivably witness the differences in the way your system plays out over its various histories. Imagine a new past: familiar empires would be unseated, popular figures would vanish; history would be virtually unrecognizable. Given that the consequences of this realization are intimately related to our understanding of economic and social welfare, it is no wonder that few people in positions of influence and power admit to be aware of this basic statistical property of the world; to do so would be their undoing.

Take a swanky cocktail party, for example. Mash about its component parts: its people, the drinks, the decor. Mix them in every conceivable way. Take note of the sights, the sounds, the signals, the meanings behind each conversation, the relationships formed as a consequence of every stare, the children, grandchildren, and great grandchildren that result from that single beautiful moment, the unending possible lives that each of these people may lead, as well as each and every other life touched by their various serendipitous excursions into the great unknown world beyond those glimmering fer forgé through which our esteemed party-goers first arrived.

It turns out you could never possibly explore the full space of contingent events (My apologies, Ken) at that party, even if you lived for an amount of time equal to the sum total lifespan of every human being ever born. By contrast, take a barrel of milk and churn it vigorously. After your hands cramp, you’ll very quickly realize that no matter how may ways you shake it, you’ll never witness the emergence of anything more sophisticated than a thick slab of butter. A cocktail party, therefore, is more chaotic and vastly less ergodic than a stick of Danish Creamery.

“History enters when the space of the possible is vastly larger than the space of the actual.” — Stuart A. Kauffman

In effect, a non-ergodic system is subject to history because knowing everything you could possibly know about its component parts will tell you very little about how its future will unfold. In the same vein, knowing everything there is to know about its present state will tell you very little about what could have been in place of what did actually occur (the shadow of the past not knowable). If only we had a time machine!

“We’re gonna make like Hitler president or something. We cant do this!”

Failure to acknowledge non-ergodicity may have been the single biggest mistake the Economics profession has ever made, but that hasn’t stopped incredibly influential people from using this rather unfortunate misfeature of a small class of erudite mathematical models to claim they know how to predict the future better than the rest of us. The presence of ergodicity allows us to predict an equilibrium. Therefore, it is useful in constructing classical market-clearing models of the economy, where prices converge and economic returns are diminishing in scale (and these models are very effective at understanding the relationship between manufacturers of physical goods, such as canned pork or soybeans, their employees’ wages, and competitors’ market strategies), but classical equilibrium models have essentially nothing to say about growth, innovation, and the emergence of monopoly power, especially when it comes to high-technology.

“CREATIVE MONOPOLY means new products that benefit everybody and sustainable profits for the creator. Competition means no profits for anybody, no meaningful differentiation, and a struggle for survival.” — Peter Thiel (From Zero To One, 2014)

You can’t predict the minutiae. While it is possible to predict the emergence of the concept of a video game as a class of possible future phenomena, a general attractor upon which predictable technological advances converge — science fiction authors have been doing this forever — you would never be able to predict, ahead of time, that one of video games’ penultimate successes would comprise an adventure where two Dinosaur-riding Italian brother plumbers go on a quest to save a Princess from a mutant warlord Turtle. And yet the Super Mario Brothers permeate modern pop-culture with such ferocity that many would be hard pressed to imagine a world without them.

If this is the case, if detailed prediction really is impossible, even in principle, then why are owners of powerful monopolies convinced that their visions for the future are more prescient than the supposed distributed wisdom of the crowds on which their enterprises are founded? Why do they start foundations instead of looking for ways to make their crowds more intelligent? Why don’t they consider setting up direct cash transfers for the poor to create new consumers? Why don’t they aggressively advocate for an increase in the minimum wage? What makes them comfortable dragging out anti-trust litigation for an entire decade while the rest of us are forced to fight for every inch?

If the philanthropic organizations of our new technological Tsars can’t predict the future with greater precision than the market can as a whole, if their clout is merely an intellectual illusion, then it stands to reason that if these monopolies had not existed in the first place, the personal fortunes of their owners would have been absorbed by the market anyway. So what’s the actual point of a Chan Zuckerburg Initiative, an organization with defined goals and mandates? A Bill & Melinda Gates Foundation, as opposed to tax breaks for the middle class? Are these kinds of directed investment instruments actually just salient social signals designed to preserve power in the face of public scrutiny over their owners’ plutocratic ambitions? Perhaps the Monopoly class is genuinely and innocently confused.

“Canada, your support of The Global Fund to Fight AIDS, Tuberculosis and Malaria is so remarkable I just had to give Justin Trudeau a hug.” — Bill Gates

It’s very easy to look like an asshole when you make public attempts to deny the obvious happiness shared in a photo like the one pictured above. But I just can’t help play the skeptic. People are fucking hyperboles (we perceive the present as vastly more valuable than is rational — see Hyperbolic Discounting). If the new Tsars of Silicon Valley are as every bit cunning as I think they are, they probably realize this as well. By expending a small fraction of their controllable wealth on socially desirable causes, it would be easy to convince the masses that their ambitions were benevolent, effectively downplaying the risk of government interference in their own enterprises.

In an increasing returns world, non-ergodicity has a lot to say about differential market share. Firms that manage to leverage positive network externalities early in the innovation cycle tend to capture more and more of those externalities as time goes on. As a consequence, competition is crowded out, and a monopoly forms. This feeds back into product quality, giving credence to the illusion that a firm’s software, or perhaps its community, is better than anything else out there (or anything else that could have conceivably emerged if VCs hadn’t thrown their entire lot behind a single play).

The truth is that potential successful competitors are usually snuffed out before they even have a chance to experiment. How many Zuckerburgs could have been? While Arthur claims that bigness is neither inherently good nor evil, I must, from the bottom of my heart, disagree. Too much bigness is profoundly evil. Monopolies crush consumer surplus, corrupt governments, stoke inequality (which is disastrous to not only people’s physical and mental well being; it engenders political instability as well); worst of all, they encourage plutocracy. Nobody worth their salt actually believes that a government founded on the personal values and objectives of a society’s wealthiest individuals could ever be a just one for very long (not even billionaires).

The problem with trying to convince seemingly benign and credible persons that they are committing invisible atrocities while they pursue their groundbreaking humanitarian ambitions is that a strong combination of luck, power, and hindsight prevent them from considering any possible alternative history that does not result in their own inevitable market victories (the present is obviously and indisputably deserved):

“If your product requires advertising or salespeople to sell it, it’s not good enough: technology is primarily about product development, not distribution.” — Peter Thiel (From Zero To One, 2014), [conflating increasing returns for superior technology].

By contrast, in a world of constant or diminishing returns, these individuals are forced daily to confront the reality that competition is the only way to present a viable road map for the future of humanity. When given choices, consumers do happily vote with their wallets. Contrary to Thiel’s stark characterization of an ultra-competitive toy economy in which nothing ever happens because the average person is a hyper-intelligent Vulcan, the world does not have to end up that way.

The period from 1945 to 1960 was marked by the largest increase in gross national product per capita the US had ever witnessed in a quarter-century. GNP doubled from $200 billion to $500 billion in 15 years. Unfortunately, a World War, and the fiat reduction of inequality that followed was the only mechanism that allowed for Americans to indulge in such an unprecedented level of shared prosperity for so long. There was more wealth than people to share in it (at least before the end of that then ensuing Baby Boom).

Thiel and other Silicon Valley moguls do well what politicians the world over are regarded for executing with precision: inventing straw men to justify their personal biases. The truth is that those at the top of the socioeconomic ladder rarely know why or how they got to where they are because they can’t compare their present state of affairs with what could have otherwise been unless, of course, they had access to either a time machine or the next best thing: an advanced quantum computer (see Quantum Annealing).

When an individual whose success derives primarily from monopoly power goes on to claim that they know where we as a society ought to head next, they can’t actually say so in a way that is falsifiable. Unfortunately, in their hubris they manage to drive out the alternatives upon which their wins or losses could have otherwise been fairly assessed and compared. Bold claims by high-tech entrepreneurs about the future, then, can’t possibly be scientific because in the absence of credible threats from potential competitors there is no way to prove their claims wrong. When Mark Zuckerburg suggests that his private foundation is better suited to achieve a particular social goal than the sustained efforts of a horde of selfish but mostly rational American consumers, this is actually pseudo-science.

Unbridled unilateral growth all but eliminates natural experiments in business-making, so how can we actually be certain that the technical advances of the most successful firms today comprise the best there is, or the best that could be? There’s nothing to compare their achievements to. But even if non-ergodic systems produce more alternate histories than ergodic ones, why should you listen to what I have to say? It’s not like I can show how these shadow pasts would have worked themselves out in detail anyway, right? Well, if prediction can no longer serve as a barometer of our success then what remains? I think the simple answer is that we need to tell more stories and listen when they are told. If everyday people don’t come to control the narrative then technology and those few who own it most certainly will. This has never played out well.

Otto van Veen — Lycurgus Demonstrares the Importance of Education

Lycurgus of Sparta, brother to the King, cast himself into exile upon news of the birth of his dead brother’s son. Lycurgus could have easily usurped the throne, but he wanted to send his Kingdom a message: that justice rather than power, however benevolently executed, is the only way forward. Despite his quasi-fictional status, this Lyconic myth paved the way for centuries of actual Spartan Kings to come: humility, temperateness, and stoicism were the callings of their time. Only through the preservation of these virtues could a sustainable society be forged, a tiny nation state, in fact, that resisted monopolization by its enemies for roughly 700 years.

No serious economist claims that we should endeavor to live in a world of near perfect market competition, where everybody knows everything they’ll ever need to know to make perfect decisions at every moment in their life, the result of which is that nobody makes any profits, and we all die of starvation. But to live in a world which more closely resembles that straw man compared to one in which rampant technological inequality drives extreme sociopolitical enmity between millions of otherwise capable yet unemployed citizens is not, as Thiel and other Silicon Valley dreamers would have us believe, an irrational social pathology: anathema to progress.

It is the only moral way forward.