Mandelbrot
The Accountant
On day 143, the plane was accepted by the Army Air Corps and ready for flight. We had beaten the schedule of 180 days.
That shop of crates and tent canvas became the first Skunk Works, and the man who assembled it out of stubbornness and salvage was Clarence “Kelly” Johnson. He would go on to design the U-2, the Constellation, and the SR-71, the Blackbird that flew at three times the speed of sound while other experts still called it impossible.
For most of my life I filed Lockheed under one heading. A merchant of death. A legacy contractor grown fat on public money, the kind of place where nothing moves without a signature and every signature needs three more above it. That picture was not wrong so much as it was lazy. Kelly Johnson complicated it. Reading his account, I found not a bureaucracy but a violent allergy to one, run by a man who thought the whole apparatus of committees and inspectors and review boards was the enemy of anything worth building. There are two sides to most coins. I had been looking at one of them for a long time.
What Did I Get Out of It
The Tyranny of the Committee
Kelly ran the most secret and technically demanding programs in American aviation with teams a fraction the size of his competitors. Fifty people on the U-2. A hundred and thirty-five engineers on the SR-71, an aircraft where, in his words, everything had to be invented. He knew exactly what made the difference, and he named the villain plainly.
There is a tendency today, which I hate to see, toward design by committee—reviews and recommendations, conferences and consultations, by those not directly doing the job. Nothing very stupid will result, but nothing brilliant either.
The committee optimizes for the absence of blame, not the presence of brilliance. That is the whole trap in one line. Spread a decision across enough people and no single person owns it, which feels safe, and produces exactly the beige competence Kelly describes. I watch this happen in every organization that scales. The sign-off chain lengthens. The number of people who can say no grows while the number who can say yes shrinks. Everyone is consulted and nobody is responsible. His third operating rule was a direct assault on this drift.
The number of people having any connection with the project must be restricted in an almost vicious manner.
Vicious is the right word, and it is the word most organizations flinch from. Restraint on headcount reads as a constraint, but Kelly treated it as the enabling condition. Fewer people meant faster decisions, tighter control, and a morale that came from everyone being able to see their own hand in the result. The Outsiders CEOs understood the same thing about lean head offices. Small is not a compromise you accept. It is a weapon you choose.
Skin in the Game
Kelly flew in the aircraft he designed. Not as ceremony, but as method. He treated his own fear as a calibration instrument.
I have a philosophy that those who design aircraft also should fly them—to keep a proper perspective. The engineer knows where the quarter-inch bolts may be marginal, what the flaps are likely to do or not do. I figured I needed to have hell scared out of me once a year in order to keep a proper balance and viewpoint on designing new aircraft.
The designer who has to sit in the seat draws the bolt differently. This is skin in the game as an engineering practice rather than a philosophical argument. The person who bears the consequence of a decision makes a different decision. I think about this every time I look at a control that was signed off by someone who will never feel its failure. The signature is there. The exposure is not. Kelly refused that separation, and he made it structural rather than optional. He even reframed the entire test regime through the same lens.
I always have thought of flight testing as a method of inspection—to see how well you have engineered and built the bird.
Inspection, to him, was not a gate someone else operated after the fact. It was the builder discovering the truth about his own work while strapped inside it. The feedback loop was closed and personal. Contrast that with how most large systems separate the maker from the checker, the author of the number from the auditor of it, and you start to see why quality diffuses the moment responsibility does.
Inspection as Theater
The most uncomfortable passage in the book, for me, was a single arithmetic fact about a supplier.
At the Baird Atomic Company, which made the vehicle’s horizon sensor, Lockheed had 40 people inspecting, coordinating, and reporting. Yet Baird had only 35 people building the instrument. We resolved that situation by returning responsibility for the product to the vendor.
More people watching the work than doing it. I have seen versions of this ratio, and it always announces the same disease. When an organization loses trust in an outcome, its reflex is to add a layer of surveillance, then another to watch the first, until the cost of checking the thing exceeds the cost of making it and the quality has not improved by a single unit. Surveillance is not the same as assurance. Kelly’s fix was to invert the instinct entirely.
Push more basic inspection responsibility back to subcontractors and vendors. Don’t duplicate so much inspection.
The insight runs against everything a control function is tempted to do. You do not manufacture quality by stacking inspectors on top of a supplier who no longer feels accountable. You manufacture it by handing the supplier the responsibility and the consequence together, then removing the safety net that let them stop caring. Adding controls can subtract ownership. Measuring harder can corrode the thing you were trying to protect, the same way chasing the metric breeds cobras. Kelly’s answer was not fewer standards. It was fewer hands, each one holding more.
Simple, But Redundant
Two of Kelly’s rules sit in apparent contradiction, and the contradiction is the interesting part. The first is the one everyone quotes.
I reverted to my old Skunk Works axiom, “Keep it simple, stupid.” The more complexity, the more potential for problems.
Simplicity was not an aesthetic for him. It was risk reduction. Every part you add is a part that can fail, a joint that can crack, a system that can behave in a way no one modeled. Strip the design to what the mission requires and you strip out the failure modes with it. So far, so clean. But the same man wrote something that looks like its opposite.
Redundancy in systems since then became a mania with me. With everything we build, we make sure that we can relight, restart, and keep flying.
Simplicity and redundancy pulling against each other. I sat with this longer than I expected to, because I could not immediately reconcile the two. Then it clicked. He was simple where failure was survivable and redundant where failure was fatal. Complexity was a cost he refused to pay for elegance and gladly paid for the ability to relight an engine at seventy thousand feet. The design is stripped to the bone. The failure path is built three deep. That is not a contradiction. It is robustness applied with judgment about which edges of the envelope actually kill you. Most portfolios, and most control frameworks, get this backwards. They carry redundant complexity in the normal case and none at all in the tail.
The Boy from Ishpeming
Kelly was the seventh of nine children in a poor family in Michigan. His father laid two thousand bricks a day. His mother took in washing on a scrub board, and Kelly picked up and delivered the laundry in his wagon, or on a sled in winter. From twelve he paid his own room and board. The origin of one of the century’s great engineers was a small mining town and a gift from a steel baron.
Next to my father, I credit Andrew Carnegie with being the most important influence on my early life through the library he had donated to Ishpeming. I went to the library almost every day. It opened a whole new world to me.
The Carnegie library returned to that iron-ore town a richer resource than the ore it had shipped out, and it reached exactly the child it was built for. There is something worth sitting with in the compounding here. A fortune built partly on Ishpeming’s iron came back as a building full of books, and one of those books turned a poor boy into the man who would design the fastest aircraft ever flown. The decisive input was not money. It was access, met by a child already reaching for it.
I was 12 years old that I would be an aircraft designer. My whole life from that time was aimed at preparing for that goal.
He knew at twelve and never wavered. He reworked every problem in a calculus textbook on a summer vacation, decades after passing the course, just to keep the capability sharp. What grit looks like when it is fused to a single aim early and never negotiated with. The passion did not appear before the skill. The two grew together over a lifetime of deliberate work, which is closer to how mastery actually forms than the story we usually tell about finding your calling.
Who Is This For
This is a book for builders and operators, for anyone who runs a team and suspects the process has quietly become the product. If you have watched a lean, fast group calcify into a slow one as it grew, Kelly will show you what was lost and name the mechanism. His fourteen points read like a specification for keeping speed and ownership alive under scale, and they cost nothing to read and a great deal to actually implement. He said so himself. He spent years trying to convince other companies to adopt the method, and almost none did, because they wanted the results without paying the price in delegated authority and restricted headcount. That honesty is the book’s best feature.
It is not for someone looking for a management framework to bolt onto an existing bureaucracy. The whole point is that you cannot. The method requires trust, authority concentrated in one person, and a customer willing to shrink their oversight to match. Half-measures produce nothing.
What the book changed in me was smaller and more personal than any of its lessons. I came to it carrying a tidy contempt for the industry Kelly served, and he did not resolve that contempt so much as expose how little thought I had put into it. The weapons he built are what they are, and I have not softened on that. But the man was not the caricature I had assembled from a distance. He was a poor kid from a mining town who loved machinery, refused to lie about a design, and built extraordinary things with tiny teams and no patience for the theater of control. Two sides of a coin. I had spent a long time examining only the one that confirmed what I already believed.