jebmo’s posthaven

When I first read Paul Graham’s essay How to Make Wealth, I was quickly seduced, like, I am sure, many young corporate cubicle-dwellers who think themselves more clever than their employers, and found myself day-dreaming about quitting my job and forming a startup with a catchy url and the $50 Million I would get for selling it to Google. At the time, I was able to overlook the flaws and inconsistencies in the argument because the inspirational nature of the piece was so compelling. On closer examination, however, I find that underlying Graham’s motivational pep talks for the under-appreciated geek is a tenuous philosophical position that demands proper refutation. Whether this essay will succeed in righting Graham’s injustices remains to be seen. There are so many problems with Graham’s thinking that it is difficult to organize a focused response. Thus, with this essay, I intend mainly to introduce the reader to an alternate line of thought, lest they be so overwhelmed by Graham’s crusades that their mental capacities for logical reasoning are badly warped.

The principal function of Paul Graham’s essay “How to Make Wealth” is, as in so many of the thoughtful writings available on his website, to inspire. His words are a manifesto for the creative hacker trapped in a cubicle. His main argument is that computer programmers have a talent for creating wealth, and are exploited by a corporate system in which they are not justly rewarded for their work. The startup offers an alternative that embodies greater freedom and greater reward. Graham appeals directly to the deeply held insecurities of intelligent introverts. We want to believe that there is a world that will worship us like gods, not cram us into ever smaller featureless boxes and treat us like cogs.

As an inspirational piece, Graham’s essay is undeniably powerful. Yet, for a great many of his points he drifts into territory that is clearly not his specialty and the flawed assertions that he makes subverts his main message and makes him look foolish. Graham’s understanding of economics is woefully pedestrian, un-researched conjecture. His conceptualization of the economics of exchange relies on a stubborn insistence and blind naive trust that the free market is the ultimate solution to which his audience must subscribe. It is distressing that one could allow arrogance to so cloud his judgment that he would embarrass himself by making such claims in an area of study he is clearly lacking any authority on which to speak.

To understand why Graham, and the many others he has inspired, make these mistakes it is necessary to look at the nature of introversion that develops among computer programmers, and how this peculiar characteristic can interfere with their judgement on matters of the social world.

1. Introversion

Introversion is a curious and wonderful trait: the ability to create and inhabit a mental utopia devoid of the complications of external reality is a truly liberating experience. Free from the ordinary trappings of social norms and customs, those lucky enough to dwell in inner space are rewarded far more generously than those whose sense of accomplishment must be subject to the approval of others. On the other hand, the profound disjoint between one’s inner life and the outer world—the locus of the introvert’s insulation—is also dangerous in that it can lead to the most ridiculous assumptions about how the world works.

No class of person better illustrates what it means to be a paragon of introversion than the computer programmer. The nature of the work itself very nearly demands that a rich and complex inner world exists wherein intricate structures of abstract thought can roam and romp freely. It is no coincidence that the psychology of those attracted to programming includes people who feel very comfortable inside their own heads, and rarely more than a short distance beyond. It is also no surprise that many of them strongly prefer their inner fantasy world and the absolute power they wield within it as opposed to the nasty, brutish, and short external world in which we are all powerless, enslaved by animalistic needs, and waiting for death.

Introversion comes in many forms, but the form that typically develops among hackers is a particularly potent and tantalizing one. The mind of a hacker must have at its core a logic factory. This provides the ability to construct massively complex structures from the basic elemental components of logical axioms, operators, and truth-statement evaluators. But this is not enough. A hacker must also have developed advanced visualization tools to navigate within the logic factory at incredible speeds. One must be able to zoom in to an infinitesimal point, understand its purpose, then quickly zoom out to galactic scale and know how each part in the factory contributes, leviathan-like, to the grand design. This requires the mind to be fluid, aware of multiple simultaneous perspectives and dimensions, to constantly challenge the impetus for order and explore the unseen paths that only a revolutionary will find. If you’ve ever been there, however briefly, where any of this made sense, you know you would give anything to go back.

This world that the hacker inhabits, while a rich source of internal delight and personal inspiration, is generally a piss poor model from which to form opinions of the outside world. Nature is not deterministic, people do not behave like cellular automata, and economic systems do not spring forth from algorithms found in textbooks by Don Knuth.

Thus, there is a certain morbid fascination involved in watching a hacker talk politics because many of them have developed utterly ridiculous ideas that do not seem worthy of their otherwise keen logical intellect. Here we find that someone with an obviously intricate and brilliant mind, capable of amazing feats of artistry and ingenuity with a computer, can simultaneously hold the most idiotic and brain-dead assumptions about society, economics, and politics. Hackers are particularly susceptible to the fallacy that because they are brilliant software developers, their views on everything else must also be equally brilliant. Although this is sometimes true, in the vast majority of cases, hackers appear positively child-like when it comes to non-silicon matters. Introversion, the source of the hacker’s strength, is also a limiting factor because it blinds its holder with a false confidence, and the appearance that just because things ought to be true, that they are in fact true. It is with this lens that we may begin to see how Graham has come to some of the more fanciful conclusions in his essay.

2. The Deification of the Programmer

One of Graham’s early contentions in his essay is that money does not equal wealth. Money is a finite medium of exchange used to move wealth around, but wealth, the object of human desire, can be generated by applying labor and talent to some raw materials. Graham’s formulation is a rather pedestrian rehash of Adam Smith’s account of specialization and exchange and it appears that Graham’s own intellectual development on economic matters appears not to have evolved much beyond the 18th Century. Sadly, this yields an agonizingly myopic view of social structure that neglects any treatment of the role of power, particularly how manipulation of economic need can distort market forces.

In deriving the political views of the hacker, Graham asserts that programmers, as craftsmen, create wealth directly: “they literally think the product”. As creators, they also directly observe the vast differences in productivity that separate the best programmers from the merely mediocre. The wisdom afforded this privileged class of wealth creators by these two observations, Graham concludes, must lead them to a libertarian philosophy: if “the top 5% of programmers probably write 99% of the good software,” then why shouldn’t those top 5% be justly rewarded for the wealth they have directly created?

It is amazing how many otherwise intelligent people have been seduced by this remarkably simplistic line of thought. While tantalizing in its apparent logical completeness, there are actually several critical errors in the above reasoning which render Graham’s conclusions baseless. The first is the idea that measurement of things like quality and success can be objective, perfect and fair. These are not objective facts, they are highly contextual and can be manipulated by power struggles, charisma, clever marketing, or outright fraud.

Value is a social construct, and the measurement of the quality of software code is as much a consequence of marketing and sales efforts as anything that is emitted on the command line. Due to these contingencies, while it is certainly true that wealth is being created in there somewhere, it is simply not possible to attach a direct quantitative relation between the work a programmer is doing and the exact amount of wealth that results from it.

A second critical assumption being posited in Graham’s essay is that one person’s direct contribution can be disentangled from that of others. In contrast to the programmers, the craftsmen-wealth-creators, Graham identifies that “Many of the employees (e.g. the people in the mailroom or the personnel department) work at one remove from the actual making of stuff.” So what exactly do they contribute to the wealth generation process? Does Graham imply that without these others working at “one remove”, the programmers could still create the wealth they do? Without the human resources team to coordinate their medical benefits, would the programmers be as productive? Without the legal team to fend off frivolous lawsuits brought by patent trolls, would the programmers still have enough time to create their wealth after first reviewing the IP implications of every line of code they write? What about the operations departments who keep the power flowing, the networks running, the servers humming, what fraction of programmer productivity is the result of their hard work? Finally, without the sales and marketing teams, how else will code find it’s way to the top of the industry?

Graham implies that once the programmer geniuses are done with their work, they can sit back, fold their arms, and watch as their creation magically spreads throughout the land while dump trucks full of cash line up outside the door. Back to reality, thinking of all the factors that are involved in the success of a product; the work of the programmer—the creation of the actual product itself—may very well be the most insignificant part of the entire process.

3. Value is a method, not an attribute

Any Python programmer should be able to grasp the distinction I am trying to make with this claim. What Graham and others who follow his logic want to believe is that value is a known objective quantity that is attached to an object. In reality, it is a function that involves many factors only one of which is the labor that is used as an input. Since value is a social construct, it is exceedingly dangerous to allow ourselves to be governed by the fiction that value exists independent of competing forces and the machinations of power. Unfortunately, Graham makes this mistake throughout his essay and it severely undermines the legitimacy of his claims.

In his introductory section “The Proposition”, Graham describes the economics of working at a startup as a trade-off between time and intensity of work: “Instead of working at a low intensity for forty years, you work as hard as you possibly can for four.” Although he never says this payoff is guaranteed, he doesn’t deny it either, and implies that if you really are as good as you think you are, then by rights, you will accrue exactly what nature and the market deem you to be worth.

Graham spends a great deal of time quantifying exactly how much more productive you could be as a superior programmer compared to the average employee, when these metrics involve fundamentally unquantifiable units. He ignores the fact there are so many mitigating and mediating factors between the work and its value, and instead blindly assumes that his calculations of productivity will translate into a greater payout for the smart hacker. Unfortunately, he has devised a model that poorly fits the data, since there is far more unexplained variance in the equation that he totally ignores.

The mistaken idea that value can be objectively quantified goes hand in hand with another critical error that Graham assumes to be true: that it is not only possible but easy to disentangle the value contribution of one person from another. This problematic idea is introduced in his second footnote: “Hypothesis: A company will be maximally profitable when each employee is paid in proportion to the wealth they generate.” Later in the section “Working Harder”, Graham writes:

“A company that could pay all its employees so straightforwardly would be enormously successful. Many employees would work harder if they could get paid for it. More importantly, such a company would attract people who wanted to work especially hard. It would crush its competitors.”

Although this sounds attractive at face value, upon closer reflection it becomes apparent that such a situation would be too easily subject to manipulation, and the result would be the complete opposite of Graham’s utopian claims. While the hackers are busy working on the hard problems faced by the company, the one who wins the prize will have chosen the only real hard problem remaining: how to convince the owners that he or she was the one responsible for all the value being created and to whom the windfall should be paid. Such a system would only reward the best schemer, the best at personal marketing and showmanship, the best ass-kisser. Any company that attempts such a vague “pay for performance” initiative would destroy itself because it would end up rewarding the wrong people, and all the right people would get fed up and leave.

In the section “What a Job Is”, Graham outlines the deleterious effects of mind-numbing institutionalization and how working for a large corporation is such an agonizing prospect for a bright hacker: “In a company, the work you do is averaged together with a lot of other people’s.” Later, he writes, “I think the single biggest problem afflicting large companies is the difficulty of assigning a value to each person’s work.” And yet, he did it so easily in his “back of the envelope” calculation that a superior programmer could be 36x as productive as the typical employee. How does this extra productivity translate to reward? Graham’s answer is that it can only happen in the context of a startup.

Let’s explore his thinking further: “A programmer, for example, instead of chugging along maintaining and updating an existing piece of software, could write a whole new piece of software, and with it create a new source of revenue.” Here is the crux of Graham’s assumption that programmers are the real engine of the value chain. He ignores the fact that without the infrastructure and ancillary components of the business, it isn’t so easy to simply translate that new piece of software into pure profit. Consider an alternative scenario: a sales team could sell an existing product to a new vertical. Now, the market for the company’s products has been expanded and incremental value has been added to the same product line. This requires no additional work from the programmers, and yet the sales team has now increased the value of the programmer’s work. Who should get the credit?

Graham’s discussion of the function of sales contradicts his claim that programmers are the wealth creators. “Salesmen are an exception. It’s easy to measure how much revenue they generate, and they’re usually paid a percentage of it.” This assumes that sales works in a vacuum, but in reality they also rely on the same infrastructure of the business to support their sales efforts. Sales needs a product team to design a solution to spec for their clients, an analytics team to provide data on its effectiveness, marketing to shape and package the value proposition, etc. Once you begin to consider how all these facets work together, it becomes impossible to disentangle the contributions of one group, whether sales or programming, from that of another. Yet Graham repeatedly overlooks this point: “Unfortunately, companies can’t pay everyone like salesmen. Salesmen work alone. Most employees’ work is tangled together.” This is wrong because nobody works alone, not sales, not marketing, not operations, and not the programmers. Graham is almost right: he should have said that everyone’s work is tangled together, but this would defeat his entire argument.

In the section “Measurement and Leverage”, Graham extends his explanation for the merit pay of salesmen to upper management. Here his ideological biases are revealed in their full glory. Graham argues that measurement and leverage are two necessary conditions for becoming rich. Given that executives have an easily determined metric—short-term profits—and the leverage to make command decisions, this apparently justifies their enormous pay packages. Graham so desperately wants to justify his own wealth as the righteous product of his own personal labor without acknowledging the effects of either luck or power, to which more sober commentators on wealth creation are quick to point.

“There is one other job besides sales where big companies can hire first-rate people: in the top management jobs. And for the same reason: their performance can be measured. The top managers are held responsible for the performance of the entire company. Because an ordinary employee’s performance can’t usually be measured, he is not expected to do more than put in a solid effort. Whereas top management, like salespeople, have to actually come up with the numbers. The CEO of a company that tanks cannot plead that he put in a solid effort. If the company does badly, he’s done badly.”

This claim that CEOs have risk which justifies their reward simply does not match up with reality. CEOs are no more at risk than the gentlemen soliders who lead their armies from the comfortable confines of a bunker well removed from the front lines. Whether the company succeeds or fails, the CEO and most of upper management will still be handsomely paid. Even in the rare occurrences when a CEO is ousted for poor performance, there are two key factors that mitigate their risk: first, they are paid so extraordinarily well that they can afford to be out of a job, and second, they travel in networks within which it is easy for them to secure another position of equal stature relatively quickly. Failed CEOs find new posts all the time. Consider how in Silicon Valley, being the former head of a failed startup is considered to be a badge of honor—it demonstrates that you have experience and will do things differently next time. Oddly, however, the honor appears to accumulate with each successive failed startup, hence the term “serial entrepreneur”.

There is a class dynamic at work here that Graham ignores. Upper management have cultural capital to which mere mortals do not have access. For ordinary workers, being laid-off through no fault of their own, to say nothing of being laid-off due to their own failure, is certainly not a badge of honor, and they find it much more difficult to parlay this experience into a positive spin. The CEO really has no downside risk at all. The explanation for their pay rates has nothing to do with measurement or leverage, but is determined purely by their location in the social structure, and by the ability of power to set its own rules.

Graham’s limited understanding of organizational dynamics leads him to more embarrassing claims in the section “Smallness = Measurement”. Here, he uses the analogy of the “ten best rowers” who, if you take them out of a large system and put them together with a shared goal, will necessarily be superior. But running a company is not as clear-cut as a boat race. There are multiple means for a company to succeed, and multiple strategies for getting there. Ten of the smartest people will not necessarily come up with the best strategy for success, let alone the best tactics for executing those strategies. The dynamics of team effort have a great deal of impact on a company’s success. ‘The smartest guys in the room’ was a phrase used to describe Enron, and certainly no one would want to emulate them today. There are numerous cases where the smartest people have made the dumbest decisions. Arrogance clouds judgment far more aggressively than introversion.

In elaborating on the startup trade-off, Graham writes “if you want to make a million dollars, you have to endure a million dollars’ worth of pain.” This is an interesting perspective on an industry that sees a great deal of failures. For the vast majority, you will endure the pain and get nothing. How can this be seen as an incentive for young people entering the work force? The answer, of course, is that Graham’s inspirational mantras whip up such fanatical devotion among his target audience that they overlook these obvious facts and are inculcated into a way of thinking that if they really are the best, then they simply need to put in the hard work and they will automagically succeed.

It is amazing how well this piece serves as marketing fodder for Graham’s venture capital arm, Y Combinator. His business model relies on convincing hordes of eager young hackers to sign over their surplus labor to his investors. With logic crafted to appeal directly to the introverted minds of recent computer science graduates, he has no shortage of cannon fodder lining up on his doorstep willing to eat Ramen and gleefully line the coffers of his investor’s portfolios.

In the section “Technology = Leverage”, Graham returns to his flowery appeal to the introverted mind by using some grand historical examples of technological innovation in the past—Florence in the 1200s, the Dutch in the 1600s—with the underlying implication that you too, if you are a genius programmer, can quit your job, found a startup, and be next in this line of magnificent historical transformations. As if writing a new ecommerce shopping cart application can be placed alongside these examples. Bonus points, I suppose, if you use an obtuse language to do so.

“Technology that’s valuable today could be worthless in a couple years.” This is Graham’s justification for why technology rewards fast movers. We do not need to travel far to find an example here. Graham’s own software was completely worthless two years after he sold it to Yahoo. It had to be rewritten because nobody could (or cared to) maintain a web application written in Lisp. How interesting then that Graham ignores this point when considering the value of programmer’s labor. Perhaps he was led to this conclusion upon reflection of the fate of Viaweb and how lucky he was to have sold at the right time. It turned out that the “wealth” he created was not so durable after all.

In the penultimate section, “Get Users”, Graham puts forth some advice to the young hacker on the best way to think about their startup. “I think it’s a good idea to get bought, if you can.”

Conveniently, he runs a company that will offer to buy you out, but only, of course, if you’re one of the really really smart ones.

4. Libertarianism != Meritocracy

Even if we were to overlook these deficiencies in the argument leading up to the conclusion, the most critical of errors Graham makes is the assumption that a world constructed in a libertarian framework would rightly and justly reward hard work with a proportionate amount of wealth. It remains a giant leap to assume that a libertarian world would operate in such an equitable fashion.

In the midst of discussing the hacker and its discontents, Graham notes, though without any empirical substantiation, that “so many of the best programmers are libertarians.” Let’s explore this further and try to find out how he is led to this silly claim. Graham makes this observation in a discussion of the programmer as craftsman. Programmers create. Their labor leads directly to a tangible product. They also recognize the vast differences in productivity among programmers. Graham assumes that this explains why so many programmers are libertarians. But why? Why would a disparity in productivity lead people to want a system with less or zero regulatory oversight?

Presumably, Graham is implying that those who believe themselves to be at the top end of the curve would want an economic system in which they would be appropriately rewarded for their superior efforts. He assumes that a more libertarian society would necessarily be more meritocratic. But this assumption has been proven false time and again, based on everything we know about the nature of power and corruption. Historically, societies with laissez-faire economic policies have been associated with rigid, hierarchical social structures with negligible social mobility. In a “free” society, what reason would those in power have to give lots of money to a productive hacker, when they could simply use coercion? What guarantees would you have that the definition of productivity would be fairly evaluated or that everyone would agree upon the metrics for productivity? An antagonist could manipulate the rankings via fraud or deception and without the means to stop them, an honest worker would not stand a chance. Another alternative is that, collectively, everyone else would simply decide that what programmers do is tantamount to sweeping floors and changing light bulbs, and suddenly programmer salaries would begin to resemble those of janitorial engineers.

Graham is by no means the first of his kind to fall victim to the false consciousness of laissez-faire rhetoric. How do so many programmers become deluded by this free-market ideology? This is precisely the dangerous potential side-effect of the hacker’s introversion: they begin to think that the world does (or should) operate like a computer. Having spent so much time in the logic factory of their mind, they buy in to the fallacy that the real world should somehow emulate the electronically deterministic one: that compensation is a linear function of value, merit can be objectively determined, markets are free, and everyone plays by the same rules. Stop laughing.

It is hard to fault someone for wanting to receive fair compensation and appropriate recognition for their hard work. The introverted hackers feel especially put upon by a society which tends to reward flashy salesmen in fine suits and fast cars. Unfortunately, the hackers’ introversion has bled them of any empathy, so they cannot see anything but the work they are doing, and cannot see the value that anyone else is contributing. Thus, with a history of not faring too well in social competitions, the insecure hacker seeks a redistribution of wealth based on metrics they they alone define. To assume that everyone else should judge the value of their work the same way they do is not romantic but naive.

The problem shared by libertarians, free market pundits, and laissez-faire capitalists is this assumption that we can all agree on standards and that human beings will tend to do only good things if left alone. This fallacy flies in the face of every shred of evidence we have about the nature of power. Graham rightly points out that wealth is not a zero-sum game. Power, on the other hand, is most definitely a zero-sum proposition, and those who have it will stop at nothing to retain it and quash any perceptible threat. Power is a black hole—it consumes everything and yet remains a perpetual void. We see instances of this even with the meager controls in place in our society today. Loosen or remove those controls and power, unchecked, will accumulate until nothing remains. Anyone or anything in its path will be destroyed.

The final section “Wealth and Power” is the first time that Graham even mentions power in this context. This section is so painfully amusing to read that it is almost not fair to critique it because it was clearly written as an after-thought to cap off an already lengthy essay. “A great deal has been written about the causes of the Industrial Revolution.” Surprisingly, however, it can all be summed up in a few quick paragraphs that basically argue that the triumph of Western civilization is all owed to nerds who wanted to get rich. Feudalism and communism were just bullies on the playground, and any government intervention preventing nerds from accumulating more wealth will only lead to ruin.

Graham concludes: “The same recipe that makes individuals rich makes countries powerful. Let the nerds keep their lunch money, and you rule the world.” Sadly, the bottom line is that unless you are among a tiny handful of the elite, a laissez-faire economic system would enslave you, not liberate you. The only people who will enjoy the fruits of a truly free society are those with the power to keep you in line. Given more freedom, those in power will first cut off the access to power, to prevent anyone from challenging them. They will immediately go for the jugular. There is no compromise with power and no means to share it. Thus, the power elite will readily command vast armies of people and resources to ensure that you stay down. This will be their first priority. Their second priority will be to extract labor from you in the manner that serves them best. In a perfectly libertarian society, the bullies would not only take your lunch money, they would murder your family, burn your house, and leave you for dead by the side of the road. This is your free society. Enjoy.