Discourses on Big Tech

Discourses on Big Tech

Prudent men are accustomed to say, and not by chance or without merit, that however wishes to see what has to be considers what has been: for all worldly things in every time have their own counterpart in ancient times.That arises because these are the work of men, who have and always had the same passions, and they must of necessity result in the same effect.

— Niccolo Machiavelli, Discourses on Livy

There’s a thriving mini-industry in publishing books about What Machiavelli Would Do To Climb The Corporate Ladder. Some of these are fun, but many of them are less about Machiavelli and more about the idea of Machiavelli. Which is really too bad. There have been power-hungry psychopaths throughout history, and as long as there have been such people, they’ve had enablers. If that’s all Machiavelli is, he’s rather boring, just another amoral hack who happens to be unusually good at his day job. That’s the impression of Machiavelli that you’ll get from The Prince, but at around the same time that he wrote that book, he was also working on The Discourses, a very different work promoting a very different worldview.

The Prince is—or at least, can be read as—a selfish person’s guide to achieving maximum power. Discourses on Livy is something different: it’s a selfless person’s guide to creating sustainable institutions, knowing that there are plenty of Prince-reading types lurking in the background. So while The Prince talks about how an amoral monarch might take power and deal with threats, The Discourses talks about how a republic that preserves personal liberty can prevent itself from being seized by, say, an amoral monarch.

The fact that he wrote both of these books raises the important question of what, if anything Machiavelli really believed. There’s a different mini-industry there, from people trying to parse exactly what his true beliefs were and what he was just saying for effect or shock value. Was he secretly a fan of human rights, perhaps the creator of modern liberal thought? Was he mostly a nationalist, who praised republics because the Roman republic had united Italy, a project he saw as valuable? Was he as bad as he sounds at his worst—“Therefore a wise lord cannot, nor ought he to, keep faith when such observance may be turned against him, and when the reasons that caused him to pledge it exist no longer. If men were entirely good this precept would not hold, but because they are bad, and will not keep faith with you, you too are not bound to observe it with them,” or the classic “[A] question arises: whether it be better to be loved than feared or feared than loved? It may be answered that one should wish to be both, but, because it is difficult to unite them in one person, it is much safer to be feared than loved…”

If it were an easily answerable question, there wouldn’t be so many books trying to answer it, but one interesting hint observed by Leo Strauss is that Machiavelli opens both The Prince and The Discourses with a dedication: The Prince to an existing ruler, Lorenzo de’ Medici, and The Discourses to two of his friends, saying “I have chosen, not those who are but those who from their infinite merits deserve to be Princes,” and noting that people who dedicate their books to princes often flatter them and ignore their flaws. Treat the dedications as a dashed-off note from a busy writer who just finished his manuscript and wants to get it out the door, and that’s trivia. But treat them as meaningful and the implied meaning is that Machiavelli thinks The Prince is a sort of poison pill, the kind of advice that is liable to get a ruler overthrown and replaced by, say, some of Machiavelli’s friends who have been reading sounder advice.1 Treat The Prince entirely unironically and it’s an act of moral bulimia: swallowing a bunch of bad principles, and throwing them up undigested onto the page. This may make Machiavelli an early example of accelerationism.

The Mini-Medici and Wannaborgias who read The Prince and treat it as good advice are falling for a wonderful meta-Machiavellian trick—they’re getting the fun and punchy advice that tricks them into overstepping their bounds and losing everything.2

And should we care what Machiavelli has to say? Sure, there’s a fun frisson from drawing analogies between poisoning a rival and preventing them from getting a promotion, or arranging an excommunication and getting someone’s name trending on Twitter. But are the analogies really valid? There are two ways to answer this, one looking forward from the 1510s and the other looking backward.

The forward-looking argument is: we’ve been reading Machiavelli ever since, and half a millennium of scholars (and cynical politicians) can’t all be wrong.

The backwards-looking argument is that Machiavelli made arguments citing the Roman Republic, Classical Greece, medieval Florence, and the contemporary Ottoman empire. If there’s something that’s true for Gaius Marius, Philip of Macedon, the Medici, and Sultan Bayezid II, it’s probably a valid statement about the human condition. Machiavelli makes many modern points with older analogies, which makes perfect sense: he’s using Livy’s writings on ancient Rome to give advice to Renaissance city-states, so some of it is stripped down to a timeless enough essence that it can be applied straightforwardly to modern concerns. He is, for example, obsessed with the idea that mercenaries are strictly inferior to other kinds of armies, and notes that a prince needs to be good at war above all else. Machiavelli, in other words, was one of the first management thinkers to warn against outsourcing one’s core competency. He also has some unkind words about environmental assessments, zoning, and the FDA, arguing that “One council or magistrate should not be able to stop the actions of cities.” Given the information he has, his example here comes from Venice, but if he were writing today he’d probably go on at length about historic laundromats or the dreadful threat of shadows.

The paradox of first principles thinking is that it’s very hard, but by definition anyone can do it. So if you reject tradition and decide to go your own way, the most effective way to do it is—see what other people have done, because philosophy has a multi-millennia head start and many more brain-years than you. So if you don’t read the classics, you’re opting, by default, into another tradition, one tested more by theory than by practice.

Machiavelli cares about power. Simultaneously thinking about very minor maneuverings—who gets to control which strategic fort?—and setting things up for the bigger question: is there such a thing as “Italy?” And, if there is, how do we ensure that it’s not ruled by France, Austria, or Spain? This dynamic is familiar to anyone who has tracked one of the big tech companies: following Uber circa 2017 meant bouncing back and forth between “What do unit economics look like in suburbs compared to dense cities?” and “When and how will they automate driving, eliminating tens of millions of jobs, and completely reshaping cities?” In early 16th century Italy, people who cared about power would try to get involved in government, as Machiavelli did, and they’d try to shape outcomes according to their goals. Today, people who care about power learn to code.

Starting Up

The ongoing question Machiavelli tries to answer in the Discourses amounts to: what can we learn from early Rome? Specifically, what can we learn about building durable institutions from a country that fought wars constantly and faced nonstop internal power struggles, but still managed to expand over time?

One of the first points he raises is: is it better for a city to start somewhere pleasant or somewhere harsh? Modernizing this question, he’d ask: is it better to start in a high-margin business with good returns on capital, or a low-margin business buffeted by cyclical swings? Or, more directly, do you want to look for the next Google or the next Amazon? He weighs arguments for both sides—the advantages of living somewhere nice are obvious, but harsh terrain does tend to toughen people up. That maps nicely to the Amazon/Google comparison, too: since they first flipped the switch on AdWords, the question for Google has been just how big they can get, not whether or not they’d be profitable. Amazon went through a long period during which its survival was vividly in doubt: in 2000, a bond analyst repeatedly published research arguing that the company could be close to insolvency, and this research was taken seriously at the time: Amazon’s stock dropped 20% and the analyst was soon hired by a hedge fund. He turned out to be wrong, of course, but this wasn’t obvious at the time. In 2000, Amazon was a retailer with a challenging cost structure, fighting its way through a recession; the fact that it was run by someone who Warren Buffett considers “the most remarkable businessperson of our age,” was not part of the argument then, and surviving the dot-com bust is in fact one of the reasons Bezos gets that honor.

Ultimately, Machiavelli comes down on the side of 1) living somewhere nice, and 2) using the force of law to keep people from getting lazy and decadent. For all his craftiness, sometimes the smartest thing he does is refuse to overthink it: of course the obviously easier path is easier, and of course side effects have solutions to them. In modern terms, he’s advocating for a corporate culture that consistently acts scrappier than it needs to be. At any company, you can basically run into three kinds of constraints: dealing with outsiders, dealing with insiders, and not having enough money. His basic argument is that the dealing-with-insiders problem—corporate infighting, slow turnaround time, bad documentation, etc.—are in principle solvable, while the external problem will always be outside of your control. There are some companies that develop internal pathologies because they’re making enough money not to care about a decaying internal structure, and that’s always a difficult problem to solve, because so many people are unmotivated to solve it. Google seems to have gotten better at this under Ruth Porat’s tenure as CFO, by applying more discipline to individual projects and disclosing more about how each part of the company performs. This doesn’t completely solve the easy money anomie of running a dominant search engine, but it’s a start.

Company Culture, Machiavelli Style

“Princes should not complain of any sin that the peoples who they have to govern commit, for it must be that such sis arise either by his negligence or by his being stained with like errors.” Machiavelli knows that culture starts at the top: CEOs constantly set an example, good or bad, for what kind of behavior is tolerated and what kind is required.

I wrote about this a bit last year: Bill Gates was still writing code that shipped in 1989, when he was already a billionaire. This was not, in direct terms, a great use of his time, but was probably highly motivating to everyone else at Microsoft. Similarly, Bezos’ customer support emails have a pretty low direct ROI, even if the time investment was minimized by making their text either a “?” or the dreaded “!” And it turns out Machiavelli wrote about that, too: “Nonobservance of a law that has been made, and especially by its author, is a thing that sets a bad example.”

He advocates a very realistic approach to laws, as he does with just about everything else: "Policies are never wise if one can or should doubt their observance.” Companies can set ostensible goals, especially around things they insist they won’t do, but those goals only work if they’re explicit, and even then they only really happen when incentives are aligned.

If there’s a quick summary of a working Machiavellian culture, it’s one that starts at the top and works its way all the way down. CEOs, even if they have the enviable minimal-sleep mutation, can’t monitor every employee all the time, so at some point they have to delegate. And that delegation means setting up the right rules and also the incentive to follow them. An example is a cost-effective form of delegation, and works as a sanity check, too: the goal is not just to encourage everyone to follow the rules by setting a good example, but to reconsider a rule if following it would be too hard for the CEO.

Where the Machiavellian approach to corporate culture gets very fun, though, is when it stops working. If a CEO leaves not-entirely-voluntarily, it means there wasn’t manager-company fit, but it also means there was some kind of organized effort to make the switch. The broad term for such an effort is a conspiracy, a term that has negative connotations but that really just means any time a group of people work together in secret to accomplish some aim.

Conspiracies happen! And, true to his reputation, Machiavelli loves talking about them. The “Conspiracies” chapter is the longest section of the Discourses, and it’s full of practical advice and examples. But it’s worth taking one step back and looking at big tech companies and conspiracies with an even wider lens: Anticompetitive behavior is fundamentally a conspiracy—convincing another company to accept deal X in order to achieve secret aim Y. And as Machiavelli emphasizes, a conspiracy can’t succeed if it has too many participants. Since any defector could kill the conspiracy, every one of them can feel entitled to 100% of the spoils. Machiavelli argues that republics are more trustworthy than monarchies because republics simply can’t move fast enough for an opportunistic betrayal. Bigger companies with more formal checks and balances have the same problem—if it takes two weeks to get all the relevant decision makers onto the same conference call to hash things out, the chance may have passed.

So the classic IBM/Microsoft deal worked because IBM operated more like a republic, with a complex org chart, carefully delineated responsibilities, checks and balances, etc.; Microsoft was a lot closer to a monarchy, so when Gates saw an opportunity, he took it.3

What this means is that big companies, especially companies that dominate their market, already have a somewhat conspiratorial nature. They already know what words not to put in emails, so if they’re used to backchannel communications for one purpose, why not use them for another purpose? This might explain why Uber’s 2017 went the way it did: in a company that has to carefully compartmentalize what it says to drivers, regulators, and partners, it’s possible to compartmentalize what board members say to the CEO.

And this kind of conspiring has its attractions: “For to be able to make open war on a prince is granted to a few; to be able to conspire against them is granted to everyone.” But conspiracies don’t run themselves! In fact, they are in one sense self-defeating: Machiavelli points out that if you’re disgruntled, and you share your view with someone who feels the same way, they’re mostly relieved that they’re not alone, and they’re less likely to want to change things. (And, of course, if you reveal your opinion to someone who doesn’t feel that way, you’ve made a big mistake.)

Sometimes, conspiracies can fail without completely ruining the perpetrator, if they’re talented enough. Machiavelli gives the example of a Carthaginian politician who planned to poison the entire Carthaginian senate at a wedding feast and assume power. The senate decided that he was popular enough that they couldn’t punish him, so they reacted by setting a limit on wedding feast expenses, instead. This kind of story is just about the least likely to leak in a corporate context, since it’s embarrassing to everyone involved—not to mention being a serious violation of The No Asshole Rule—but it does seem clear that some very high-functioning sociopaths at the companies that need such people will get away with things that would otherwise get them fired.

What Machiavelli ends up describing is that companies with very effective founders, and with strategies that ultimately give them unassailable control over their target markets, are surprisingly fragile internally. Top-down management combined with internal information asymmetry means that it’s easy for the CEO to miss what’s going on internally without strenuous effort, and means that management slip-ups can quickly spiral into legitimacy crises. This portion of the book would have been very different if 16th century Florence had discovered the wonders of dual-class stock.

Founders and Succession Problems

If there’s one obsession that runs through the Discourses, it’s the question of what happens after a very good leader leaves the scene. At one point, Machiavelli seems to argue that succession is the one fundamental question that needs to be answered: “It is seen that two virtuous princes in succession are sufficient to acquire the world, as were Philip of Macedon and Alexander the Great.”

Machiavelli knows that the natural state of any institution is to decay unless it’s fixed: “It is a thing clearer than light that these bodies [i.e. principalities or companies] do not last if they do not renew themselves.” That is a pressing concern for many of the biggest tech companies right now: some, like Microsoft and Alphabet, have gone through CEO transitions, which have generally gone quite well—in both cases, most of their market cap was created under CEOs who were hired well after founding and promoted from within. Amazon is early in its transition, but their current CEO previously ran their most profitable unit, and one that demonstrated many of the most Amazonian business behaviors. Netflix has a co-CEO, and Salesforce is likely to have a new CEO soon ($, The Information).

Many of the top tech companies are very much reflections of their founders, for better or for worse. And trying to maintain that is a challenge; the last thing Microsoft in 2021 wants to be is a company that embodies the culture of Microsoft in the mid-90s. At the same time, they don’t want to lose the more permanently valuable parts of their culture.

Machiavelli’s answer here is somewhat ambiguous; he definitely seems to prefer republics over monarchies, even though he finds monarchies more interesting. But he also knows that the big changes are generally executed by one person, not by a committee. The solution is to go through cycles of renewal, driven by individuals who end up in charge during a crisis: “If a republic were so happy that it often had one who with his example might renew the laws, and not only restrain it from running to ruin but pull it back, it would be perpetual.”

And that’s a hard model to implement. How does a company that’s calm and boring recruit and promote someone whose jobs is to shake things up? And if they can’t recruit that person, how can they bring them on from the outside and give them enough legitimacy to actually execute on their plans?

The succession problem remains fundamentally hard to solve, but Machiavelli points to meta-stabilizing factor: crises that don’t break an institution will tend to fix it instead. “The Veientes believed that by assaulting the disunited Romans they would conquer them, and this assault was the cause of the union of the latter and of their own ruin. For the cause of the disunion of republics is usually idleness and peace; the cause of union is fear and war.” So that’s the dark hope of any founder-dependent company that’s trying to figure itself out after the founder leaves: hopefully things will get bad enough to prompt a refounding before they get so bad that the business collapses entirely.

What’s The End State for a Company?

Machiavelli frequently gets credited with founding realism as a school of political theory. He wasn’t the first person to say some of the sorts of things he said, but as Strauss points out he was the first to say them in his own name. There are Machiavellian lines in Plato’s Republic, but Plato puts them in the mouths of characters like Thrasymachus, so he’s not necessarily endorsing them. The way to sum up realism is: institutions are made out of people, not rules. As a corollary: the two broad approaches are 1) a good system for incentivizing whoever is in charge, or 2) a good system for selecting the right people to be in charge.

And the Republic comparison is not far off in another sense: Machiavelli wrote about principalities, i.e. independent polities, and the ends to which they could be directed. His audience would have identified as Florentines. People today often identify themselves primarily in terms of where they work—“What do you do?” is a standard icebreaker. The polis that was the subject of historical political science is not at America scale or even a modern city’s scale. But just about company scale. Florence had a population of around 50,000 in Machiavelli’s day, compared to Facebook’s 59,000 employee count at yearend 2020.

For a political entity, the big concerns start with survival and end with maximizing human flourishing. (The easiest-sounding one is usually whichever one you’re not struggling with at this moment.) That doesn’t map to every company, but it does map nicely to the dominant tech companies, many of which have monopoly-like characteristics. Whatever their core business is, it’s hard for someone else to replicate, which means that the company is responsible for figuring out what version of that business the world has to deal with. Google’s job is not just to maximize the net present value of future free cash flow (“Survival” is keeping that number above zero), but to figure out, to the best of its ability, what a search engine should be. Facebook has to do the same for social networks, and Apple has a large share of the responsibility for deciding what smartphones are and what should come after them.

Countries have natural boundaries determined by geography, language, and culture (with language and culture often shaped by those same geographic barriers). Companies, too, have natural boundaries, more defined by where they can get an above-average return on investment and where they’re stuck getting a market return or below. One subtext of The Prince—made explicit in the last chapter—is that there’s such a thing as Italy, and that the various city-states ought to stop squabbling, and that’s a partial theme in the Discourses, too. But that plan has natural limits; it doesn’t imply that Italy should then proceed to take over the world.

So the Machiavellian view of the big tech companies is that they should fight to defend their quasi-monopolies, but only do that in order to earn the privilege of having them in the first place. His analysis of Rome points out that they had a very fortunate transition early on, from the conquering Romulus to the law-giving Numa. And that’s likely the post-founder state of big tech: they start out taking over a market, and then spend time figuring out how they can deserve to keep it.

Further reading: Aside from the Discourses and The Prince, it’s also worth reading Harvey Mansfield’s Machiavelli’s Virtue and New Modes and Orders, the latter of which is a chapter-by-chapter analysis of the Discourses. Strauss’ Thoughts on Machiavelli is very heavy reading. Philosophy Between the Lines is a great view of esoteric writing, with examples from Machiavelli. And Eric Voegelin has some thoughts, especially on the broader context of what was going on in Europe at the time that Machiavelli was writing. James Burnham’s The Machiavellians is also good, on Machiavelli’s style and goals.

(Disclosure: I own shares of Amazon and Microsoft.)

A word from our Sponsors

Managing bills shouldn’t require sacrificing your investments. With M1 you can borrow against your investment portfolio to pay the bills and keep your money in the markets.

M1 is the all-in-one Finance Super App that allows you to invest, borrow, and spend your way, and M1 is fully integrated to make your money movements smoother and your financial life easier. Plus, it's free to download and get started. M1 is for savvy investors, like readers of The Diff, who want to take control of their finances and commit to a long term strategy.

M1 is not for day trading. It automates away day-to-day interactions so you can concentrate on the bigger picture. You can follow curated portfolios, schedule investments, automatically reinvest dividends, and rebalance your portfolio in just a few taps.

M1 is the Finance Super App: easy to use, award winning, and Yours to Build.

Investing in securities involves risks, including the risk of loss. Borrowing on margin can add to these risks. M1 Finance LLC, Member FINRA/SIPC.

Elsewhere

I joined Dwarkesh Patel's The Lunar Society podcast for a wide-ranging conversation on optionality, the Great Stagnation, secret societies, and more.

Sticky Wage Growth

The Economist has a decent piece on why this bout of inflation is not like the 1970s ($). Most notably:

In 1976 more than 60% of American union workers were covered by collective-bargaining contracts with [Cost-of-Living Adjustment] provisions; by 1995, the share was down to 22%. A paper published in 2020 by Anna Stansbury of Harvard and Mr Summers argued that a secular decline in bargaining power is the “major structural change” explaining key features of recent macroeconomic performance, including low inflation, notwithstanding the decline in unemployment rates over time.

If 100% of wage growth responds to inflation, then the expected inflation rate next year always starts at this year's rate and then moves in response to growth, fiscal policy, and supply shocks. But if wages don't automatically adjust, then the inflation rate is noisier; some people get raises, and some people get effective pay cuts.

The Coal Boom

Coal has been a surprising growth industry this year; right now Peabody Energy shares are up 459% year-to-date. The optimism isn't around a long-term improvement in coal economics, but about short-term energy demand and low elasticity of supply ($, WSJ), which has pushed coal prices and coal companies' margins up. That effect is particularly strong for the stocks, because they're priced with the expectation that the business will go away fairly soon—which makes boomtime economics in any given year a large component of the full value of the company. Right now, Peabody trades at 4x estimated EBITDA over the next year (which also means it started the year trading at under 1x current EBITDA estimates!), so every good quarter is actually a big deal, even if the price increases are transient.

For an earlier Diff writeup on coal, see here ($).

Equity and Collateral

One of the biggest growth areas in consumer finance is unsecured lending that's underwritten at the transaction level, i.e. buy-now-pay-later. Another interesting trend—and topic of a future Diff piece—is the shift from owning to renting. One area that hasn't gotten as much rethinking is how to offer financial services to people who don't have high net worths, but do have some assets that can easily be valued. Specifically, car owners: Otto, a fintech startup, has just raised funding to provide vehicle loans. Car loans are a fun market, because like many lending products, they're ultimately driven by customer acquisition cost; adding a loan to an existing car sale is an easy way for car dealers to increase margins, but for many borrowers with poor credit, the need for a loan doesn't always coincide with when they buy their car. Making the market for car credit more liquid after the car is sold is a tougher proposition than just adding a term to an already-existing contract to buy a car, but it does give some borrowing opportunities to customers who would otherwise be paying wildly higher interest rates on their loans.

Building an Apple Car vs Appleifying a Built Car

Apple is working on adding more car controls to CarPlay: "The company is working on technology that would access functions like the climate-control system, speedometer, radio and seats..." One way to look at this is that the automotive business is increasingly a data business, which means that companies that want to enter it will need to start collecting car data early. Part of Apple's long-term hardware advantage is the company's ability to tightly integrate its products (and make it hard for competitors to do the same); if Apple's goal is to use the iPhone to drive car sales the way it previously drove Watch and AirPod sales, this is a good start.

Diff Jobs

We've had a great first week with Diff Jobs, and are pleased with how many interesting responses we've gotten. Some interesting open opportunities:

• Business analyst, experienced with Looker, who wants to work at an under-the-radar unicorn.
• Forward deployed engineer-style role at a company that's applying novel technology to a problem that's a multi-billion dollar operating expense at numerous companies.
• A company that's in the education space—but in it to fix it—is looking for multiple marketing and product roles, as well as a chief of staff.
• A payments-related company is looking for data scientists and machine learning engineers in Europe.
• On the crypto/Web3 side: if you've been doing security in that space for a while, we have a client who would be very interested to talk to you.

If you're interested in applying, please reach out with the form below: we'll contact you, talk through your background, and see where we can find a fit. (If we don't find something right away, don't worry—we're onboarding many new companies right now, and some of the companies we've talked to have hiring needs further in the future.)