In 1977, at the Great Lakes Shakespeare Festival in Cleveland, Ohio, a theatre director named Dan Sullivan screamed at a hungover twenty-year-old actor who had just delivered a disastrous rehearsal. The actor was Tom Hanks. Sullivan's instruction was three sentences long: show up on time, know the text, and have a head full of ideas. That was it. No motivational speech. No feedback sandwich. Three rules, delivered at volume, to a young man who had failed all three.
Forty-three years later, Hanks stood at a podium to accept the Cecil B. DeMille Award at the Golden Globes and told the world that those three rules were the most important lesson he ever received. Not from a film school. Not from a director on a hundred-million-dollar picture. From a man running a regional theatre in Ohio who refused to tolerate laziness. "You have got to show up on time," Hanks said, "and you have to know the text and you have to have a head full of ideas. Otherwise, I can't do my job."
I have been thinking about those three rules for months. Not because they are profound -- they are obvious. I have been thinking about them because they describe, with painful precision, everything the technology industry has collectively decided not to do.
The Word Itself
The word rehearsal comes from the Old French rehercier, meaning to harrow again -- to re-plough soil that has already been ploughed. This is not a metaphor about repetition. It is a metaphor about preparation as physical labour. You take the ground apart. You break it up. You turn it over. Then you do it again. The soil does not improve because you walked across it. It improves because you worked it.
The technology industry has lost this understanding. Preparation is not skimming a document five minutes before a meeting. It is the deliberate, effortful act of working material until you have internalised it -- until you can respond to it, challenge it, build upon it. Goethe understood this in 1803 when he wrote his Rules for Actors: rehearsal conditions must match performance conditions. The performing arts have known for centuries that preparation is not optional. It is the work.
A Sequence, Not a List
What Sullivan gave Hanks was not three independent virtues. It was a sequence. Each rule unlocks the next.
You cannot know the text if you do not show up on time. Showing up late is a statement -- it says the text does not matter, the other people in the room do not matter, and the work itself can wait. You cannot have an idea if you do not know the text. Ideas that emerge from ignorance are not ideas. They are guesses dressed in confidence. And confidence without preparation is the most expensive thing in technology.
The jazz analogy proves the point. Miles Davis hired virtuosos precisely because they already knew the text so thoroughly that they could depart from it with purpose. The freedom came from the preparation, not instead of it. Without that substrate of deep knowledge, you have nothing to adapt, nothing to depart from. You are improvising from zero, and improvisation from zero is not jazz. It is noise.
This sequential logic -- Rule One enables Rule Two enables Rule Three -- is also a diagnostic. When things go wrong, you can trace the failure backwards through the chain to find where it broke. And once you start looking, you see it everywhere.
Then Like Now
In 1935, Boeing's prototype B-17, the Model 299, crashed on takeoff at Wright Field in Dayton, Ohio. The aircraft was not faulty. The pilot -- an experienced test pilot named Ployer Peter Hill -- simply forgot to release the elevator lock. He did not know the text. Boeing's response was not to blame the pilot. It was to invent the pre-flight checklist: a system designed to ensure that every human being, no matter how experienced, would be required to know the text before they performed. The checklist did not replace competence. It enshrined it.
That was a Rule Two failure. The pilot showed up. He was ready to fly. But he had not worked the material -- had not harrowed the soil -- and the aircraft fell out of the sky.
That same company, Boeing, decades later shipped the 737 MAX without disclosing the MCAS system to pilots or airlines. Three hundred and forty-six people died. The company that invented the checklist -- the company that understood, in 1935, that knowing the text is a matter of life and death -- decided that schedules mattered more than preparation. They stopped knowing their own text. Worse, they ensured their pilots could not know it either, because the text was hidden from them. That is not a Rule Two failure. That is a Rule Three failure -- the leadership had no idea worth the name, only a schedule and a share price.
The same pattern destroyed the Space Shuttle Challenger. In 1986, engineers at Morton Thiokol knew the O-rings would fail in cold weather. They had the data. They knew the text. But when management pushed back, one executive told the lead engineer to stop being an engineer: "Take off your engineering hat and put on your management hat." That is a Rule Three failure in its purest form. The text was known. The data was clear. But the leadership's idea of what mattered -- schedule, optics, political pressure -- was rotten, and rotten ideas corrupt everything downstream. The engineers were told to unknow what they knew.
Captain Chesley Sullenberger, who landed US Airways Flight 1549 on the Hudson River in 2009, provides the counter-example. He said afterwards that "my entire life had been a preparation to handle that particular moment." He had 19,663 flight hours. The emergency lasted 208 seconds. Every hour of those 19,663 was rehearsal. Every hour was harrowing the soil again. When the moment came, Sullenberger did not need to think. He knew the text so deeply that his ideas -- the split-second judgements about altitude, speed, river versus runway -- emerged from preparation, not panic.
That is what the sequence buys you. Not certainty. Not a script for every scenario. It buys you the capacity to respond to the scenario you did not predict, because you have worked the ground so thoroughly that your judgement is calibrated and your ideas emerge from deep knowledge rather than shallow guessing.
The Diagnosis
Only 31% of software projects succeed -- a number that has barely changed in decades despite agile transformations, DevOps revolutions, and billions spent on tooling. Nobody knows the text because nobody wrote it. That is Rule Two, collapsed at scale.
The Cortex 2026 Engineering Benchmark found that incidents per pull request have risen 23.5% and change failure rates are up roughly 30%, despite -- or perhaps because of -- AI-assisted development. We are moving faster. We are not moving better. That speed without preparation is Rule Three dressed as progress: leaders confident enough to ship, too unprepared to notice what they are shipping.
The dysfunction reproduces itself. Who designed these systems? Leaders and managers who did not show up on time, did not know the text, and did not have an idea about what a sane working day should look like.
Dark Agile and the Misread Text
There is an irony here worth pausing on. Ron Jeffries urged developers to abandon Agile in 2018, coining the term "Dark Agile" -- the phenomenon of the manifesto being used to make developers' lives worse, not better. The manifesto said "while there is value in the items on the right" -- documentation, planning, processes -- but the industry read only the left side. It treated "working software over comprehensive documentation" as permission to abandon preparation altogether.
The text was misread because people did not know the text. A document about valuing careful reading was itself carelessly read. That is not irony. That is diagnosis.
What Preparation Cannot Do
There are moments when the text does not exist yet. Startups building genuinely novel products, teams exploring territory that has no map -- you cannot rehearse what has not been written. Sullivan's three rules assume a text worth knowing. The leader's prior obligation is to ensure one exists. But even in uncharted territory, the discipline holds. You can know the adjacent texts -- the market, the physics, the history of what others tried and why they failed. Sullenberger had never ditched an Airbus A320 in the Hudson. But he had rehearsed engine failures thousands of times. The text was not identical to the moment. It was close enough.
There is also the opposite failure: preparation as avoidance. I have seen teams rehearse endlessly, perfecting documents and architectures, never shipping, never exposing their work to the world. That is not knowing the text. That is hiding behind it. Sullivan's third rule -- have an idea -- is the corrective. Preparation without the courage to act on it is not preparation. It is delay wearing the mask of diligence.
So I am not arguing that preparation is unconditionally good. I am arguing that its absence is unconditionally bad. The difference matters.
The Challenge
Showing up on time is respect for other people's time and attention. Knowing the text is respect for the work itself -- its complexity, its history, its consequences. Having an idea is respect for the future -- a declaration that you have done enough preparation to contribute something that moves the work forward rather than dragging it sideways into another conversation about what we should have read before we arrived.
If we are leaders who send meeting invitations without agendas, we are telling our teams that their preparation does not matter. If we are architects who have not read the requirements document, we are telling our engineers that their time is worth less than ours. If we are CTOs who have not studied the failure modes of the systems we are shipping, we are Boeing in 2019, not Boeing in 1935.
That is not leadership. It is extraction.
Show up on time. Know the text. Have an idea.
The soil will not plough itself. Get up and harrow.
(Views in this article are my own.)