Bend the chimney; move the firewood

May 31, 2026

Here’s the story, which comes from the Book of Han (first century BCE):

A man visits a friend’s house and notices the kitchen stove: the chimney runs straight up, so sparks shoot out freely, and a pile of firewood is stacked right against it. He tells his host plainly — bend the chimney so it doesn’t throw sparks, and move the firewood somewhere else, or you’ll have a fire. The host says nothing and does nothing.

Soon enough, the house catches fire. The neighbors rush in and beat back the flames, and several of them get singed and burned in the process. The fire is put out, the house survives. Grateful, the host throws a banquet to thank everyone who fought the blaze. He seats the most badly burned men in the places of highest honor and rewards the rest according to how scorched they got. But the guest who’d warned him in the first place isn’t invited at all.

Someone at the feast finally speaks up: “If you'd listened to that man about bending the chimney and moving the firewood, you'd have had no fire and no expense. Instead you honor the men with burnt heads and scorched brows as your guests of honor — and the one who gave you the advice that would have prevented the whole thing gets nothing.” The host is embarrassed, and sends for the original guest.

The four characters — 曲突徙薪 qū tū xǐ xīn, literally “bend the chimney, move the firewood” — became shorthand in Chinese for taking preventive action before a disaster, and the story as a whole is a standing indictment of how people reward the dramatic cure and ignore the quiet prevention. I keep returning to this story because it clarifies three failures that organizations routinely mistake for virtue.

First, they confuse truth with transmission. The early warner was right, but correctness alone did not move the institution. That is not merely a failure of persuasion. It is a failure of uptake.

A serious organization cannot depend on truth arriving through charisma, seniority, persistence, or luck. It has to build a path by which reality travels from the person who sees it to the person who can act on it.

A warning is not complete when it has been spoken. It is complete only when it has been received, understood, assigned an owner, and allowed to change the plan. Otherwise “speaking up” becomes a ritual that protects the institution from guilt without protecting it from disaster.

Second, they confuse rescue with excellence. The banquet seats people according to their burns, and we do the same. We promote the engineer who pulled the all-nighter to save the launch, not the one who designed the system, a quarter earlier, so the launch could not fail.

This is not only unfair. It is structurally dangerous. Heroism is visible. Prevention is not. The sacrificed weekend can be narrated; the weekend that never had to be sacrificed disappears. The person who drags the project back from the edge looks indispensable. The person who kept it from reaching the edge looks, at best, merely competent.

But institutions become what they reward. If recognition gathers around the rescue, then crisis becomes the path to status. The organization begins to manufacture the very emergencies it later congratulates itself for surviving. It calls this grit, ownership, urgency, or commitment. Often it is only bad design with better lighting.The same principle applies to software.Keeping an implementation simple is not a retreat from seriousness. It is bending the chimney. In a business with daily burns, there is almost no chance of arriving at clean architecture by declaration. The system will reflect the conditions under which it is built: interrupted attention, urgent fixes, partial knowledge, and people making local decisions under pressure. Simplicity matters because it gives the organization something it can still understand, repair, and gradually regularize after contact with reality. It prevents future burns by leaving fewer places for fire to hide.

The second lesson is this: a serious institution must learn to honor prevention as much as recovery, because wherever emergencies become the main theater of excellence, emergencies will multiply.

Third, they confuse ignored foresight with personal disappointment. The people who saw it coming do not merely feel slighted. They learn something about the institution. They learn that reality has no standing until it becomes damage.

Being right and ignored corrodes differently from being wrong. To be wrong is to revise your judgment. To be right and ignored is to revise your faith in the place around you. It teaches you that truth does not matter when it is still only a warning. It matters only later, when it can be counted in missed deadlines, burned money, lost people, broken systems, or bodies on the floor.

That lesson does not produce loyalty. It produces resentment first, then withdrawal. The clear-sighted stop arguing. Then they stop warning. Eventually they stop seeing on the organization’s behalf at all.

The third lesson is this: when an organization repeatedly ignores the people who can see what is coming, it does not merely waste their insight. It trains away the very perception it needs to survive.

The propensity to lead without being given the authority to do so is what, in organizations, distinguishes people that can innovate and break free of the constraints that limit their competitors. Innovation is all about leadership, and leadership is all about innovation. The rarity of the one is a direct result of the rarity of the other.

Innovation is a subject whose talk:do ratio is even more out of whack than that of leadership. Upper management in most companies talks a good game on innovation. The party line goes something like this: “We need innovation to survive. It is so important. Its importance simply cannot be overstated. No sir. Innovation is reeealy, reeealy important. And innovation is everybody’s job. In fact, it is probably the most important part of everybody’s job. Listen up, everybody: Get out there and innovate.” Oh, and by the way,

• Nobody is given any time to innovate, since everyone is 100-percent busy.
• Most innovation that happens anyway is distinctly unwelcome because it requires accommodating change.
• Real innovation is likely to spread beyond the realm of the innovator, and so he or she may be suspected of managing the organization from below, a tendency that upper management tends to view with great suspicion.

The net here is that it takes a bit of a rebel to help even the best innovation achieve its promise: rebel leadership. The innovator himself doesn’t have to be a great leader, but someone has to be.

What rebel leadership supplies to this process is the time to innovate—you take a key person away from doing billable work (this may constitute constructive disobedience on your part) in order to pursue a nascent vision—and the hard push for whatever reshaping the organization has to submit to in order to take advantage of the innovation.Timothy Lister and Tom DeMarcom, Peopleware: Productive Projects and Teams.

Early in my career, DeMarco and Lister’s Peopleware struck me as a revelation. The line that stayed with me was that leadership is all about innovation, and innovation all about leadership — that the rarity of one follows directly from the rarity of the other. As an existentialist, I found that argument almost embarrassingly seductive. It flattered the part of me that wanted to believe the situation was mine to bend. If I could not get an idea adopted, that was information about me: my patience, my timing, my channel, my ability to make value visible. So I built a practice around that belief.

I learned that patience is not a virtue one admires in the abstract. It is something one practices, often humiliatingly. Sometimes you say the same thing ten times, in ten meetings, until it stops being your idea and becomes the room’s idea, and you let it. I learned not to assume malice when ordinary stupidity, fear, fatigue, or busyness explains the failure just as well. I learned that the useful question is often not whose fault it is, but how to demonstrate value so plainly that refusal becomes embarrassing. I learned to change the avenue. Socialize an idea privately before you table it formally. Let a nudge do what a manifesto cannot. Seed the ground quietly instead of writing the heroic document nobody reads. The messenger has more degrees of freedom than he thinks.

All of this is true. And all of it can become a cop-out.

Because Peopleware was not only telling the messenger to become more skillful. It was saying something more severe: leadership is all about innovation. If an organization does not innovate, that is not, in the first instance, a fact about the cleverness, patience, or political skill of the people inside it. It is a fact about its leadership. “You failed to get your idea through” and “we built a place where good ideas die” are descriptions of the same event from opposite ends. Only one of those ends has the authority to fix it.

There is a quiet violence in telling people the failure is theirs — be more patient, find a better channel, demonstrate more value — when the structure itself was built to ignore them. Yes, cream sometimes rises. But “the cream rises” is not a strategy. It is what people say instead of having one. It describes the survivors and says nothing about everything good that drowned on the way up.

Now that I am the one with authority. The job is not to admire the burns. It is to build an institution where fewer people have to be burned in order to be believed.

A few things follow.

Leadership has to have skin in the game. Innovation cannot be everyone’s job and no one’s risk. If the people at the top bear no cost when good ideas die, good ideas will keep dying. An institution thick with rent collectors — people extracting from a position without staking anything on its future — has already decided against innovation, whatever its town halls say.

None of this means every warning is wisdom. Organizations receive too many signals to treat each one as prophecy. Judgment still matters. Prioritization still matters. But that is precisely why the process has to be explicit. A warning can be rejected. It should not be allowed to disappear.

The mechanisms exist. We simply refuse to use them. Basecamp’s Shape Up has a “betting table,” where the people with authority decide, explicitly and on the record, what the organization will actually bet the next cycle on. The ritual is not the point. The commitment is. Someone with power has to put a name beside what gets built and what gets ignored, instead of allowing good ideas to evaporate into ambient busyness.

There is a difference between giving people voice and giving truth power. Many organizations confuse the two. They create surveys, retrospectives, open-door policies, and feedback rituals, then congratulate themselves for listening. But a warning has not been honored because it was collected. It has been honored only when someone with authority is forced to answer for what happens next.

The minimum discipline is closure. If someone raises a serious warning, the organization owes an answer: accepted, rejected, deferred, owned by whom, revisited when. Silence is not a decision. It is the most cowardly form of decision. It preserves managerial freedom by destroying institutional memory.

Surfacing is not a hope. It is a designed act.

And the retrospective has to be taken seriously, with teeth. In the old story, the host is merely embarrassed. He sends a belated invitation, and the matter ends. That is not accountability. It is etiquette.

We now live in a world where almost any institution can have near-perfect recall. Every warning, every dismissal, every “we’ll get to it,” every quiet act of managerial evasion can be logged and retrieved. When the house burns, we can know exactly who was told to bend the chimney and chose not to.

The early warner deserves the seat at the head of the table. And the owner who was warned, shrugged, and let the house burn through willful negligence should face the fire too. Not embarrassment. Consequences. Otherwise the banquet simply repeats. The scorched are honored, the prophet goes hungry, and somewhere today a chimney is still throwing sparks onto dry wood while someone who sees it perfectly clearly decides it is not worth saying again.