by Jaron Lanier
In fact, conservatives have gone to endless lengths for decades to make this very point when it suits them. Since the Reagan era, a highlight of the conservative playbook has been to claim that lowering taxes raises tax revenues. Their claim is that lower taxes stimulate business growth independently of any other variables. That is precisely a claim that there can be more than one equilibrium.
This is the famous Laffer curve, which was promoted by one late 20th century president, Ronald Reagan, and ridiculed by another, George H. W. Bush, as “voodoo economics.”
It’s counterintuitive, no? On the face of it, lowering taxes should lower the amount of money brought in by taxes. A remarkable, decades-long, and maniacal public relations campaign has brought about a general atmosphere in which the idea is respectable.
While there are huge problems with the way the idea is understood and the way it has influenced policy, the ascendance of a nonlinear, systemic sensibility into popular folklore bodes well. If the public can “get” the Laffer curve, then the public can probably also gain a more honest and balanced sensibility of the nonlinear nature of the big challenges we face.
A serious attempt to find a Laffer peak, a long-term lower tax rate with higher revenues,* would have to be as experimental and long term as the quest to improve weather predictions. Maybe something about education levels, retirement rules, or even the weather would make all the difference. It would be as ridiculous to say a Lafferesque solution is impossible as it would be to say it is automatic or easy to find.
*There have been claims that the effect has already occurred briefly in special circumstances.
The Laffer curve was supplanted in early 21st century conservative economic rhetoric by a different curve, which is really just a straight line:
Both curves are hopelessly oversimplified. Recall your finicky shower knobs. If even your shower behaves in a complex way, surely the economy is also complex. Understanding it is more like the process of predicting the weather or improving medicine than it is like these smooth lines. Economics is a real-world big data problem, which means it’s hard. It’s not a phony big data problem of the kind being used to build instant business empires. That confusion is one of the great confusions of our historical moment.
The original Laffer curve had the merit of showing two peaks on either side of its valley. That betrayed an acknowledgment that there can be multiple equilibriums. The latest replacement, the absolute faith in austerity, doesn’t even acknowledge that. To accept it is to be completely hypnotized by the illusions of easy complexity.
It is senseless to speak in the abstract about whether the Laffer curve is true or false. It is a hypothesis about peaks and valleys on a landscape of real-world possibilities, and these might or might not exist. However, the possibility of existence does not mean that any such landmarks have been found.
Systems with a lot of peaks must also have a lot of valleys between the peaks. When you hypothesize better solutions to today’s way of dealing with complex problems, you are automatically also hypothesizing a lot of new ways to fail. So yes, there might well be ways to lower taxes that cause tax revenues to rise, or the economy to grow, but they will be tweaky and nontrivial to find.
To find that kind of sweet spot on the landscape requires a methodical search, which implies a certain kind of governmental actor, which is not to the liking of many of the people who most want taxes lowered. A government has to act like a scientist. Policy must be tweaked experimentally in order to “crawl on the landscape.” That means a lot of analysis and testing, and no preconception of how long it will take to get to a solution—or expectation of a perfect solution. Anyone offering automatic detailed foreknowledge of a genuinely complex system is not on the level. Cloud calculations are never guaranteed or automatic. It’s hard magic.
Keynes Considered as a Big Data Pioneer
The same argument that applies to taxation can just as well apply to employment. Keynes was offended by the sort of situation that can come up: People want to work, but there are no jobs. Builders want to build homes, but the customers are broke. Companies hold on to cash. Banks don’t lend. Homelessness rises as construction workers can’t find jobs. It would seem that all the necessary buyers, sellers, and financiers are waiting in the wings, and yet they fail to interact to cause a market to rise. This is the sort of stuck state that Keynes suggested should be prodded with stimulus.
Depressions and recessions can be understood as low hills on the energy landscape of an economy. If you have made it to the top of a low hill and you crawl around incrementally, you will always lose altitude. You seem to have already found the best state you will ever find. That is what a stuck state feels like. Holding on to money is better than lending it when the borrower is unemployed.
However, there might be a much higher hill to be climbed, just over a valley. An employed borrower could get the loan to buy the house from the developer who would employ the borrower. Keynesian stimulus is supposed to function as a kick that imparts enough momentum to bound across the valley up to a higher peak.
Keynes was an unapologetic financial elitist and had no interest in a quest for income equality or a planned economy. He simply sought a mechanism to get stuck markets unstuck. No one has proposed an alternative to his idea of a stimulus. The enduring nuisance is that someone has to guess about exactly how and when to aim a stimulus kick; this is just another way of saying you can’t have science without scientists.
Keynesian economics is an authentic form of big data science, which means it is hard and not automatic or instant. (It encompasses such ideas as the Laffer curve.)
The left is just as capable of falling into the fallacy of expecting uncomplicated results from a given economic strategy. There is no automatic correlation between social spending and social improvement, or between fiscal stimulus and fiscal improvement. Every stab at zooming from one hill through a valley to a higher hill on an energy landscape is an experiment without guarantees.
An experimentalist’s attitude is the only way forward. Technologies of complexity must be approached in a measured way, with patience and fortitude.
However, the possibility that there might be higher peaks waiting to be discovered is also the prelude to hope, and the way out of our current knot of austerity and acquiescence to private spy empires.
CHAPTER 11
Narcissism
The Insanity of the Local/Global Flip
The most basic reason to doubt or fear Siren Servers is not that they’re unfair. Life is unfair, as my conservative friends never tire of pointing out. No, the problem is that Siren Servers eventually become absurd, because of the “Local/Global Flip.”
A Siren Server can become so successful—sometimes in the blink of an eye—that it optimizes its environment—changes it—instead of changing in order to adapt to the environment. A successful Siren Server no longer acts only as a player within a larger system. Instead it becomes a central planner. This makes it stupid, like a central planner in a communist regime.
The problem is not the fault of Google or derivatives funds or any of the other schemes. Instead it’s a dangerous temptation dangled by Moore’s Law—a temptation we must learn to resist.
Cheap networking facilitates exaggerated and rapid network effects. These engender failures of the classical economic models, which had been based on competitions between multitudes of players with distinct and limited information positions.
For instance, networked finance kept on pretending it could eject risk out into the economy at large, like a computer radiating waste heat with a fan, but it became as big as the system. The computer melted.
Similarly: Health insurance companies in America, by using cloud computer analysis to mostly insure people who didn’t need insurance, similarly ejected risk into the general system. But there wasn’t some giant vastness to absorb the waste. Instead, the economies in which finance and insurance could exist in the first place were weakened.
Alas, all
Siren Servers as currently construed are likely to eventually falter in similar ways.
Google might eventually become an ouroboros, a snake eating its own tail, unless something changes. This would happen when so many goods and services become software-centric, and so much information is “free,” that there is nothing left to advertise on Google that attracts actual money.
Today a guitar manufacturer might advertise through Google. But when guitars are someday spun out of 3D printers, there will be no one to buy an ad if guitar design files are “free.” Yet Google’s lifeblood is information put online for free. That is what Google’s servers organize. Thus Google’s current business model is a trap in the long term.
The Local/Global Flip also reduces the number of available business plans. Silicon Valley, which once seemed a portal to unlimited potential, now induces claustrophobia as so many distinct companies with different competencies and cultures must compete for the same global pool of so-called advertisers. It is pathetic that Google and Facebook, two companies offering very different services, already have to compete over approximately the same customers.
Siren Servers Think the World Is All About Them
To the owner of a Siren Server, it can seem as though that server has a godlike overview of events not only on the network, but also in the world at large. This is the fantasy of being able to accomplish global optimization. It is an illusion.
Facebook’s mission statement commits the company “to make the world more open and connected.” Google’s official mission is to “organize the world’s information.” No high-frequency trading server has issued a public mission statement that I know of, but when I speak to the proprietors, they claim they are optimizing what is spent where in “the world.” The conceit of optimizing the world is self-serving and self-deceptive. The optimizations approximated in the real world as a result of Siren Servers are optimal only from the points of view of those servers.
For someone who has scaled a peak, that peak becomes the known world. It becomes hard to remember that there might be other peaks. This helps to explain the prevalence of vain selective blindness in the assessment of peaks already scaled by Siren Servers. A derivatives fund manager will suffer the illusion that the fund has brought maximum optimization and risk management to the world. A social network owner will believe that his business is one and the same with an ideal society.
It’s easy to say what a shower is optimized for. You can state what temperature and pressure you would like to achieve. What is a market supposed to optimize for? In some abstract sense, a market ought to optimize for efficiency, but market efficiency is a subjective idea.
When it comes to Siren Servers, efficiency is a synonym for how well a server is influencing the human world to align with its own model of the world. This is just the big data way of stating the fundamental ambiguity of artificial intelligence. We can’t tell how much of the success of an AI algorithm is due to people changing themselves to make it seem successful. People have repeatedly proven adaptable enough to lower standards in order to make software seem smart.
If we are to adhere to the most bloodless abstractions, such as efficiency only as measured from a Siren Server’s point of view, then a more “efficient” economy would shrink compared to a less efficient economy. If robots are someday perceived to efficiently run the world, then little money will change hands and little investment will be made thereafter.
This economic dead end would be a stuck state on one peak on an energy landscape, a lowly little pathetic hill. If we succumb to the illusion that there can be only one peak, or a single equilibrium, then we might believe that any deviation from the top of this foothill would be a rejection of efficiency and rationality. But that would be a display of the same kind of mathematical illiteracy that has already poisoned our politics.
Whatever it is we want the mechanism of a marketplace to achieve for us, we will not find the highest peaks if we organize markets to radiate risk and become deterministic accumulators of power around a small number of dominant computing nodes. Too little is learned through that process.
FOURTH INTERLUDE
Limits Are for Muggles
THE ENDLESS CONVERSATION ABOUT THE HEART CARTEL
Thirty years ago I had the good fortune to encounter Marvin Minsky, MIT professor and one of the founders of the artificial-intelligence approach to computer science. Marvin was astonishingly gracious and generous to me, yet another young weirdo to be taken under wing. Around his table I heard early voicings of the tropes that would dominate Silicon Valley and then the world decades later.
In the early 1980s, Cambridge, Massachusetts, was still a rotting place, the way most of urban America was rotting from the 1970s. Seemingly bombed-out buildings and wretched street life pressurized the gaps between buildings at MIT. Inside those buildings a ferocious new flavor of intellectual life crouched and glowed. The nerd assault on everyone else’s reality was just beginning.
One night Marvin was expounding on the economics of artificial hearts over dinner. We dined at his sprawling, deliciously messy house, over in elegant, suburban Brookline. Piles of books, academic journals, and magazines coated everything, including what were probably multiple grand pianos, judging from the shapes of paper mountains. Amazing artifacts of 20th century science poked through as landmarks . . . parts of notorious robots, telescopes, some of the earliest digital musical instruments. The scent of aging paper and machine oil. A maze was all that was actionable in what had probably once been quite a large house. Lovely old wooden walls could sometimes be spied through narrow canyons.
Hopefully I can be forgiven for paraphrasing one of Marvin’s provocations, decades later: “Each billionaire with a heart problem should spend a billion dollars on an artificial heart. Research should be concentrated in a giant project, like the moon shot or Los Alamos. Fill a small city with top scientists and engineers to make the first really good artificial hearts for some rich guys. Sure there are some interesting projects already going on . . . but small-scale efforts are taking much too much time. Spare no expense! Get it right! Once there’s a single working model, the prices will collapse like they always do. Not long after that, everyone will benefit. What’s killing millions of people is that we’re so squeamish about letting rich people be rich.” Marvin’s eyes had an amazing glint when he was mischievous.
A retort, probably from some long-haired lefty student of the era, might have been: “Wouldn’t there be an artificial heart cartel? What’s to stop heart blackmail? You’d become an indentured servant just to stay alive.”
“No, that wouldn’t happen, and for the same reason it didn’t happen with computers. There’s more money to be made selling many millions of cheap things than a few expensive things.”
“But money is just a means to the end of power. Controlling the flow of artificial hearts would be a more direct means to the same end.”
“Same could have been said, and indeed was said about computers. Once there’s one artificial heart there will be many, so don’t worry about cartels. Someone will make a cheap one, just like someone made a personal computer.”
“But if the government hadn’t sponsored the start of computer science, computers might have been much more tightly controlled by the first companies.”
“Look, even if there was a cartel, it wouldn’t last forever. The bottom line is that the sooner the technology for a reliable artificial heart is created, the sooner people will benefit from it, especially ordinary people. The delay from your social squeamishness is going to waste much more time than it would take for a heart cartel to dissolve. Why do you want those millions of people to die so you can have your perfect society. Is it worth it?”
On and on the conversations would go. Finally talked out, we didn’t go to sleep, but coded until dawn.
THE DEADLY RISK OF NOT BEING A SHAPESHIFTER
We don’t really know what technology can achieve. Glinting at us from the horizon is a fantastic vista of a hea
venly future where anything might be achieved. We can’t tell how much is mirage. Just considering that some techie scenario is impossible might prevent us from discovering how to do it. We must not acknowledge limits. Limits kill.
The feeling of being a techie on the verge of escaping limits is ecstatic, manic, and irresistible. Not only did I feel it intensely, but I also learned to convey it to others. I wove visions of what Virtual Reality would be like in my talks in the 1980s, and into my patter as we gave demos in the lab. I could make people vibrate with excitement.
In Virtual Reality you could craft any world, any scenario. This idea of “any” is treacherous and deceptive, but I didn’t yet know that. I still love creativity and expression, and especially wild free expression, but I know that meaning comes from struggle with constraints. Meaning is when creativity has high stakes. Ultralight, friction-free moments in life are wonderful, but not as figure, only as ground.
The very idea of the computer is that it’s the “general” machine, in that “any” program can be run. That turns out not to be so in practice, even though we often can’t help ourselves and still pretend it is. As we run our lives through computers more and more, we must reconcile ourselves to the illusions and truths of the digital “any.”
THE FIRST MUSICAL “ANY”
My first encounter with the allure of overcoming finitude came long before I got involved with computers. Instead I succumbed to an obsession I developed in my early teens with the work of a composer named Conlon Nancarrow. This was not a likely event. I grew up in a tiny town in an obscure part of the country, southern New Mexico, and this was long before the Internet’s debut. It was hard to even be exposed to the pop culture of the day, much less anything obscure. And yet, somehow I came across a reel-to-reel tape of the man’s music and became absolutely mesmerized.