The Role of a CEO - Part III - Being The Best CEO

The first piece in this series argued that a CEO's job is to convert the firm's specialized capabilities into compounding returns on the capital entrusted to it. The second argued that the work has to happen at the limit, because every meaningful axis of the role shows non-linear returns to operating at the extreme. That leaves the question this piece is meant to answer. What does it actually take to be the best at this job.

Why Look at the Extreme

There is a methodological point worth making before getting into specifics. The same argument that says CEOs must operate in the limit says, by extension, that the only useful place to look for instruction is at the people who are actually there. You do not learn how a curve behaves by averaging its midpoints. You learn by examining its asymptotes.

The median Fortune 500 executive has very little to teach, because the median, by definition, is where companies are slowly being eliminated. Reading another account of how an average large-company CEO ran a quarterly review is not going to move the curve for anyone running a real business. The useful examples are the ones at the edges, where the principles are visible because they are operating in their pure form.

Two of the clearest asymptotes in modern American CEO practice are Warren Buffett and Elon Musk. On the surface they look like opposites. They aren't. Each has identified the comparative advantage that lets them outcompete every other operator in their respective industries, and each has bent their entire CEO posture around it. The distance between their methods is where the framework underneath becomes visible.

Musk: The Centralized Operator Extreme

Musk's stated view of the role is uncompromising. The CEO of a technology company should reach the chair through engineering and design rather than finance or marketing. They should know the company soup to nuts. They should be technically deep enough to recognize when their team is wrong, when reality is telling them something the spreadsheet is not, and when a problem requires an answer that no one else in the building is prepared to give. He has described the role as eating glass, meaning that the CEO is the filter for the worst, hardest, most personally painful problems in the company. If the job feels comfortable, in his view, it is not being done.

The operating posture that follows from this is extreme centralization. Musk lives in the trenches of his hardest businesses. He runs hourly KPIs during critical phases. He picks a single North Star metric — cost per pound to orbit at SpaceX, autonomous miles at Tesla — and pushes every other decision through that filter. He spends most of his working hours on engineering and manufacturing problems rather than on management abstractions. His underlying claim is that without this depth, a CEO cannot accurately assess their team's work, and a leader who cannot accurately assess their team's work is, sooner or later, deceived by it.

His view of competitive advantage follows the same logic. In industries where the technology frontier is moving, what was protected yesterday is a commodity tomorrow. The only durable advantage is the rate at which a company produces new capability. He calls this building tech trees rather than moats. The pace of innovation itself, sustained over years, is what keeps competitors out, and the operator deep enough to drive that pace is what creates it.

Buffett: The Decentralized Allocator Extreme

Buffett's posture is, on almost every visible dimension, the opposite. He looks for three traits in a leader — intelligence, energy, and integrity — and he treats integrity as the hinge that makes the other two safe. He prefers businesses so structurally protected that, in his words, an idiot could run them, on the working assumption that one eventually will. He buys those businesses, installs or retains operators he trusts, and then leaves them alone for decades. Berkshire Hathaway is famously decentralized. Buffett does not run hourly KPIs on his subsidiaries. He barely runs quarterly ones.

His claim about the job is that the CEO's primary function, properly understood, is capital allocation. He has pointed out that a CEO who retains earnings equal to ten percent of the company's net worth annually will, after a decade, have personally directed the deployment of more than sixty percent of all capital at work in the business. Some CEOs turn retained dollars into fifty-cent pieces. The best turn them into two-dollar bills. Over enough years, this difference dominates everything else.

The operating constraint Buffett actually faces, then, is not running businesses. It is finding businesses worth buying at fair prices, and finding places to deploy retained capital at returns that justify retaining it in the first place. His daily work is a search problem under a discipline of patience. The skill being exercised is judgment about what to acquire and at what price, exercised slowly, across decades, and across legal entities.

His view of the moat follows the same logic. In industries where the technology frontier moves slowly, structural protection that exists independent of management's exertion is what compounds. Distribution scale. Customer habit. Switching costs. Regulatory position. Brand. Those are the assets he is paying for. The right CEO posture in that environment is to acquire those positions cheaply and protect them, not to invent new ones.

Same Axis, Opposite Ends

They are not arguing about how to run a company. Both have produced sustained, extraordinary returns over multi-decade horizons by leaning hard into the comparative advantage their industry actually rewards. The skill set Buffett brings to the chair — judgment exercised slowly under patience — would be malpractice at SpaceX. The operating depth and velocity Musk brings to his chair would burn capital at Berkshire's holdings. Each is the right answer for a different binding constraint, and each would be a worse CEO if they tried to run the other's playbook.

The instructive point is that they sit at opposite extremes of the same management axis — centralization — and both extremes work. Companies that fall between them, where the CEO is involved enough to slow operators down but not deep enough to actually drive the technical decisions, get the worst of both. They have neither the high-velocity execution of a CEO who is genuinely in the soup, nor the autonomous compounding of a CEO who has chosen good operators and stayed out of their way. This is the most common failure mode I see in operating companies, and it fails for exactly the same reason the in-between strategic position fails. The middle of any meaningful axis is the dead zone.

The first decision a CEO has to make, then, is which extreme their business actually requires. That decision should be driven by the binding constraint of the business itself, not by the temperament of the person in the chair.

The First Company I Ran

I lived in the dead zone. The first company I ran was at neither limit. Marketing was unfocused. Operations were sloppy. Costs were nowhere. It was a business slowly being eliminated in the way most businesses are slowly being eliminated, and as a first-time CEO I had to learn the trade in real time.

I learned the role in the order the business forced on me. Marketing first, because nothing else mattered if no one knew we existed. Then operations, because marketing surfaced demand we couldn't fulfill. Then cost, because operations exposed the margin we were leaving behind. Each binding constraint only became visible once the previous one was gone. The lessons were painful, and they came fast.

The business was not strong until each function, one at a time, was operating in the limit. Eventually we became a low-cost provider in the category. Then the thing flourished.

Years went by in an instant. I lost weight. The first two years were the most stressed I have ever been in my life. We made it happen, because nothing else was going to. That is what the work feels like when a CEO is operating at the limit on the constraint that binds in the moment. It isn't a posture. It's a price. I received the Andy Grove level of paranoia when within the first three months of acquisition I lost a third of my customers through no fault of my own or the sellers, and the mental framework of clawing out and having a daily disposition to solving this has never left my mind since.

Know Thyself, Then Build the Team That Completes You

The Greek maxim know thyself is older than business and more important than most of what passes for business advice. Before a CEO can honestly choose the posture their business requires, they have to honestly inventory what they bring to the chair and what they don't.

Most CEOs are good at one thing and have convinced themselves they are good at four. The first useful exercise of the role is the unsentimental admission of where you are weak. Not what you are working on. What you are weak at, today, and likely to remain weak at.

The decision tree from there is short. The binding constraint of the business is fixed. Your capabilities are partially fixed. The only variable is who you put around you. If the business demands operating depth at velocity and you are temperamentally an allocator, you do not get to choose to be Buffett. You hire a COO who can be Musk on your behalf, give them real authority, and stay out of their way. If the business demands patient capital deployment and you are an operator by instinct, the inverse holds.

Extreme talent recruitment is not an HR function. It is a CEO function, and it is the most direct lever the CEO has on whether the company actually operates at the limit. A CEO who is honest about their gaps and willing to recruit aggressively to fill them at the extreme has a chance. A CEO who hires comfortable people who reflect their own profile back at them does not.

The team has to be aligned on the same goal. Compounding free cash flow against the cost of capital, widening the moat, daily cutting cost with empathy but not sympathy, year after year, and taking those freed up dollars and reinvesting them into productive returns. Within that alignment, each seat has to be filled by someone willing and able to operate at the limit of their own function. Marketing in the limit. Operations in the limit. Finance in the limit. One function in the dead zone is enough to drag the whole business there.

Capital Allocation Is Always Applied Math

Capital allocation is one of the more misread parts of the role. It is conventionally treated as Buffett's domain, and Musk is treated as an operator. That framing is wrong. Both men are running capital allocation at the limit. They are running it on completely different time signatures, in different surface areas, but the underlying mathematics is the same.

Buffett's allocation runs on a horizon of years to decades. He retains earnings, scans for businesses to buy, places a large bet, and lets it compound. The conversion of a retained dollar into more than a dollar of future earnings happens slowly, and it happens by acquiring entities outside the parent company.

Musk's allocation runs on a horizon of weeks to quarters, and it happens almost entirely inside a single operating system. SpaceX retains its launch revenue and immediately deploys it into Starlink, then into Starship, then into the next adjacent capability. Tesla cycles capital from vehicles into energy storage, into autopilot, into compute, into robotics. Per-share value is created not by acquiring outside companies, but by reallocating internal capital, at high velocity, into the next technology adjacent to what already exists. Each adjacency, executed well, creates the surface area for the next one.

Both are doing the same applied math. A retained dollar is being converted into more than a dollar of future earnings. The variables are how fast that conversion happens, where the next dollar physically lands, and whether the CEO is making the decision deliberately or letting it happen by default. Every CEO is allocating capital, whether they recognize it or not. The only question is whether they are doing it at the limit of their thinking or somewhere short of it.

Two Views of the Moat, and Which Applies When

The same comparative-advantage logic explains the split on moats. Buffett operates in industries where the technology frontier moves slowly. A regional concrete business, a consumer staples brand, a railroad, an insurance underwriter — all of these have moats that exist in distribution, customer habit, regulatory position, and switching costs. Innovation pace is not the binding constraint in those businesses. Operating discipline and capital allocation are. Buffett has tuned himself to win at exactly that game.

Musk operates in industries where the technology frontier moves quickly. A space launch company, an electric vehicle manufacturer, an AI company — all of these face commoditization risk on any moat that does not advance. The pace of innovation, sustained, is the only durable protection. Operating depth and intra-system reallocation are the binding constraints. Musk has tuned himself to win at that game.

Each is correct for the field they have chosen to play on. The strategic error that ends companies is picking the wrong philosophy for the wrong industry. Building tech trees in a stable industry burns capital that should have been allocated elsewhere. Trusting in static moats in a moving industry is how a once-dominant business wakes up one quarter and discovers the moat dried up while management was managing.

The Principle This Gives You

Pulling these threads together, the working frame is this. Every business has, at any given moment, one or two binding constraints that determine whether it compounds or stalls. The CEO's job is to identify those constraints honestly, know themselves well enough to admit which they cannot personally close, recruit the talent that can operate at the limit of each function, choose the management posture the constraints actually demand, allocate capital deliberately at the time signature the business requires, and pick the moat philosophy that fits the industry's frontier.

None of those decisions is permanent. The binding constraint shifts as the company matures. The right posture at year three is rarely the right posture at year ten. What does not shift is the principle. Identify the constraint. Operate at the limit against it. Move to the next one.

What the Series Argued

It is worth restating, plainly, what the three pieces in this series have argued.

The first piece, building from Ricardo through Smith to Friedman, established the rule. A CEO converts the firm's specialized capabilities into compounding returns on the capital entrusted to it. That is what the role exists for.

The second piece, working from the calculus concept of a limit, established the posture. Every meaningful axis of the role rewards operating at the extreme and punishes operating in the middle, so a CEO must push the company, and themselves, toward the asymptote of strategy, cost, talent, and character.

This piece, using Buffett and Musk as the two cleanest examples of the role done at the limit, gave the answer to what it takes to be the best at it. Know yourself. Identify the constraint that actually binds your business. Recognize the comparative advantage you hold against the field. Recruit the team that can operate at the extreme of each function. Push to the limit the constraint demands. Nothing in the middle.

A CEO who has internalized those three things — the rule, the limit, and the binding constraint — has, in my experience, helped by frameworks for operating, and little use for most of what is otherwise written about the role; because much of it doesn't stem from the fundamentals then get to the extremes. The math underneath the job is also older and simpler than the literature suggests.

Maximizing the Self

The math underneath the job is, in fact, older than the literature suggests by a couple of millennia. It is older than business writing, older than the corporation, older than capitalism. What this framework actually describes, underneath the language of constraints and limits and capital, is an exercise in maximizing the self.

The Greek maxim know thyself was carved into the temple at Delphi for a reason. Aristotle took the maxim and built it into a complete ethics. He argued that the highest good for a human is not pleasure, not wealth, not honor, but eudaimonia. Flourishing. Achieved through the lifelong cultivation of arete — excellence — in character and reason. The full, active actualization of our uniquely human capacity for rational thought.

Aquinas finished the thought a thousand years later. A concept fully and clearly understood in the mind, in its purest form, leads to right action — a foundational concept in the renaissance. You cannot help but do the thing if you understand it fully. The understanding and the doing collapse into action.

"You can just do things." It is an American restatement of a two-thousand-year-old idea about the radical autonomy life actually provides to anyone willing to take it. Once you actually understand what the binding constraint is, what your comparative advantage is, what extreme the business demands, what your team needs to look like, the action follows. You can just do well-ordered things.

The CEO who has done this work is doing more than running a company. They are doing the human work Aristotle and Aquinas were pointing at. Taking the rational faculty seriously. Applying it to a real problem under real constraints. Acting on what it reveals. Eudaimonia in a suit and a P&L. Maximize value for society, and the investors who took the risk to create or own the company.

The reason the role is hard is the same reason that life is hard. Most people will not pay the price of seeing themselves clearly. Most people will pay the price living a life of quiet regret, not being free to act on the world as it is and avoiding uncomfortable issues rather than running towards them. The CEOs who do, in my experience, are among the most fundamentally alive people I have ever met. They are doing what humans are built to do — at the limit, on something that matters.

That is what the role actually is, properly understood. It is not a title. It is a sacrificial vehicle for human flourishing of self while creating value in the world.