There is a class of organizational failure that is specifically dangerous because it looks like nothing at all. The sprint is green. The dashboards are healthy. The team is shipping. And the mechanisms that will eventually produce the breakdown are working silently, right now, in the things nobody is watching. Not because anyone chose to ignore them — but because the standard toolkit of technology leadership is designed to surface what is already failing, not what is in the process of failing.
The five patterns in this article are not the dramatic failures. They are the quiet ones — the slow compounding of neglected upstream work, the invisible growth of error under organizational insularity, the framework that outlived its usefulness twenty quarters ago, the imperceptible drift from building something useful to performing the appearance of building something useful, and the gradual dissolution of the shared rituals that hold distributed effort together. None of these announce themselves. All of them are correctable — if they are first seen.
What Your Engineering Organization Refuses to See Upstream
Peter Attia, in Outlive, opens with Bishop Desmond Tutu’s parable: “There comes a point where we need to stop just pulling people out of the river. We need to go upstream and find out why they’re falling in.” [[Untitled 9]] The parable is applied to medicine, but its structure maps precisely onto how most technology organizations manage their systems, their teams, and their technical foundations. They pull people out of the river — fight fires, fix production incidents, perform heroic deployments, staff up when teams are overwhelmed — and they never go upstream to find out why people keep falling in.
The mechanism that keeps organizations downstream is simple. Attia: “Someone who drops dead of a heart attack did not just get sick an hour earlier. The disease was working inside them, silently and invisibly, for decades.” [[Untitled 9]] Technical debt, team burnout, architectural brittleness, and security vulnerabilities all follow the same timeline. They do not appear. They compound. By the time they are visible enough to register on a dashboard, the majority of their damage has already been done in the years of silent accumulation that preceded the crisis.
Attia’s framing for proactive versus reactive medical practice applies with striking precision to engineering organizations: “Medicine 3.0 places a far greater emphasis on prevention than treatment. When did Noah build the ark? Long before it began to rain. Medicine 2.0 tries to figure out how to get dry after it starts raining. Medicine 3.0 studies meteorology and tries to determine whether we need to build a better roof, or a boat.” [[Untitled 9]] Most engineering organizations are running Medicine 2.0. The investment in prevention — the technical health reviews before the debt becomes acute, the team load conversations before the burnout becomes a resignation, the architectural assessments before the system becomes unmaintainable — is consistently deferred in favor of the treatment the current crisis requires.
A curated leadership note from 02-21-2024 distills the cost in engineering terms: “the cost of software is the cost of change.” [[02-21-2024]] And: “designing love/excellence in, means designing noise out.” [[02-21-2024]] The organizations that design excellence in — before those things become critical — are practicing upstream engineering. Those that design reactively are treating the river. A note from 2022-08-04: “Keeping the habit alive is a powerful act. It’s easier to stay in shape than to get in shape. It’s easier to keep a house clean than to get it clean. Many days it may feel like you are treading water, but maintaining your progress saves your future self a great deal of work.” [[2022-08-04]]
Why Isolation Guarantees Error Accumulation
A 2025 reflection note makes the structural claim plainly: “We are terrible at recognizing bias and weakness in ourselves. But we are really good at recognizing it in other people. Thus, I am your best source of self-correction, and you are my best source of self-correction. This idea was supposed to be the engine of democracy. The degree to which we are isolated and withdrawn, is the degree to how likely we are to fall to self-deception.” [[2025-01-20]]
This is not a call for more feedback mechanisms. It is a structural observation about the mechanism of self-correction. You cannot correct what you cannot see. You cannot see what is invisible because it is everywhere around you — the water the fish does not know it is in. A 2024 note frames it: “the story of the fishes, don’t know what water is, because its around them they having nothing to differentiate it with.” [[2024-01-15]] The architectural assumptions your senior engineers have held for a decade are not visible to those engineers. They are visible to anyone who has not held them — the new hire, the client’s engineer, the competitor who built a different system. But only if those perspectives are structurally invited rather than filtered out.
Attia’s account of Ignaz Semmelweis in Outlive is the engineering culture case study for what organizational isolation does to the person with the right answer: “His observations were most unwelcome. His colleagues ostracized him, and Semmelweis died in an insane asylum in 1865.” [[Untitled 9]] Semmelweis had the correct answer about handwashing and infection. His colleagues’ consensus was wrong. They continued killing patients for fifteen years to avoid the discomfort of external correction. The dynamic is identical in engineering cultures that have been successful long enough to stop listening — to customers, to adjacent teams, to their most recent hires. The hierarchy enforces the consensus. The consensus is the water. Nobody sees the water.
From the mixed quote collection in Untitled 12: “By judging others we blind ourselves to our own evil and to the grace which others are just as entitled to as we are.” — Dietrich Bonhoeffer. [[Untitled 12]] Applied to engineering leadership: the organization that judges external input as irrelevant — too junior, too far outside the domain, insufficiently technical — has closed the only mechanism by which its accumulated errors can be corrected.
The Model You’re Teaching That’s Capping Your Team’s Thinking
A 2025 reflection note describes a pedagogical pattern that extends far beyond education: “Its almost completely false but we teach it to kids, because it trains the mind/perspective, gives them skills so they can get the skills to get to the place where they can get the propositional knowledge.” [[2025-01-20]] The Bohr model of the atom is the canonical example. It is almost entirely false. It is an extraordinarily useful teaching scaffold. And in classrooms that mistake it for reality, it permanently caps the thinking of everyone taught it. The note adds: “you dont receive an insight, you participate in a self-organizing process.” [[2025-01-20]] The scaffold is not the destination. It is the apparatus that makes the destination reachable.
Engineering organizations are full of Bohr models. Sprint velocity as a proxy for team capability. Story points as an input to delivery forecasting. The two-pizza team size heuristic. The microservices-by-default starting assumption. Each was a useful simplification of a genuinely complex real-world phenomenon. Each has been institutionalized in enough organizations to qualify as an industry-wide scaffold. And in every organization that has confused the scaffold for the building, the scaffold is now capping what the team can see.
Attia identifies the same structure in medicine: “the notion of a continuous march of progress from Hippocrates’s era to the present is a complete fiction.” [[Untitled 9]] And: “Medicine 2.0 takes the findings of evidence-based medicine and… the problem is that no patient is strictly average.” [[Untitled 9]] The randomized controlled trial — evidence-based medicine’s foundational tool — was designed to find what works on average across populations. Its application to individual patients is a scaffold that has been mistaken for a treatment protocol. The scaffold that made population-level science possible is actively preventing the individualized treatment that actually works.
A 2023 leadership note names the organizational imperative: “to build the groundwork for detailed explanation in ways that people need to understand, make it as mundane as possible.” [[2023-07-06]] Start with the scaffold; teach toward the reality. The scaffold trap is not in having simplified frameworks — it is in forgetting they are simplified. A 2024 note offers the most practical illustration of what scaffold-as-reality produces: “I thought she was accusing me of murdering their cat, when really she wanted to make sure i washed my hands.” [[2024-01-15]] The operating model was completely wrong. It made the actual question impossible to answer correctly. The misidentification was not stupidity — it was the logical application of the wrong frame to a real situation.
The Difference Between Being Useful and Being Visible
A curated leadership note contains one of the sharpest formulations of a distinction most engineering leaders have never made explicit: “as a man are you optimizing for service not attention.” [[02-21-2024]] The Hormozi formulation is more blunt: “Hormozi mantra: ‘Fuck happiness. I will be useful.’” [[02-21-2024]] These are not the same orientation, and they do not produce the same organizations.
The difference is directional. A 2025 note names the structural error in attention orientation: “The construct of meaning, is typically is egocentric. Instead, how do I create meaning for other people.” [[2025-01-20]] The engineering team whose roadmap decisions are shaped primarily by what will be impressive in the next board presentation — the all-hands review, the quarterly leadership report — is attention-oriented. Every individual decision may be technically defensible. The aggregate is a product shaped for the benefit of the producers, not the users.
A note in Untitled 12 / Alex Hormozi: “You’ll build a bigger brand making 0 content and accomplishing shit for 5 yrs, then making content about what you did. No one cares about your opinion until they know it’s worth something. Back up your talk with walk.” [[Untitled 12]] Organizations that build the most enduring reputations do so by delivering value for long enough that the reputation arrives uninvited. Those that pursue the reputation directly — the press releases before the product, the thought leadership before the delivered value — are attention-optimized. They frequently have excellent communications and chronically struggling products.
A 2023 leadership note / John Adams: “Fire them with ambition to be useful.” [[2023-07-06]] Not ambition to be recognized. Not ambition to be promoted. Ambition to be useful — specifically useful to the people the organization exists to serve. The 02-21-2024 note names what this looks like in practice: “achieve something great in the agency of others.” [[02-21-2024]] And: “It is the ego that cares about whether or not people think or care about you.” [[02-21-2024]]
What Disappears When You Cancel the Ceremony
A 2023 reflection note makes a claim that most technology leaders would instinctively dismiss and that, on examination, is structurally precise: “The hole we are in as a society is due to the collapse of ritual both private and collective. One of the things that gives children’s safety, security and the ability to explore is the establishment of predictable rituals. You want happy children? Feed them at the same time everyday, sleep at the same time everyday, same ritual everyday.” [[2023-07-06]] The mechanism is not sentimentality. It is structural. Predictable shared rhythm creates the psychological safety from which exploration becomes possible. Remove the rhythm and you remove the safety. Remove the safety and you remove the exploration.
Applied to engineering organizations: the retrospective, the sprint demo, the planning ceremony, the weekly team sync, the quarterly architecture review — these are not inefficiencies to be eliminated in the name of velocity. They are the mechanism by which a distributed group of people remain a coordinated team rather than a collection of individuals executing work in parallel. The same 2023 note names what disappears when the rituals are removed: “What have we lost by losing the sabbath? we threatened the integrity of the family, We stopped doing anything communal together that is oriented towards the higher good. We sacrifice all the time, instead of stopping on Sunday, we exhaust ourselves in sacrifice.” [[2023-07-06]] The organization that cancels its ceremonies to move faster has made the same trade: it sacrifices the structure that makes the sacrifice meaningful, and replaces communal orientation toward the higher goal with isolated, exhausting, disconnected effort.
The 02-21-2024 note names the leadership question that ritual answers: “What is our drumbeat to whose drum are we marching?” [[02-21-2024]] In the absence of shared rhythmic practice, every team member’s own internal drumbeat diverges. The coordination costs that accumulate are far more expensive than any ceremony. A 2022 daily note names the individual discipline that underlies collective ritual: “Have a routine that is regimented and monotonous.” [[2022-11-01]] Not creative. Not surprising. Not inspiring — regimented and monotonous, which is the property of all reliable infrastructure. And from 2022-08-04: “Keeping the habit alive is a powerful act.” [[2022-08-04]] The ceremony does not need to be profound to be necessary. It needs to be kept.
So, What Is Your Engineering Organization Not Watching?
Five questions — one for each pattern — that surface where these dynamics are operating.
First: What is currently accumulating upstream in your engineering organization — silently, invisibly — that will become visible as a crisis in eighteen to thirty-six months? Not the technical debt you know about. The kind you have not yet discovered because you have not been looking upstream.
Second: Who is the Semmelweis in your organization right now — the person with the correct answer who is being managed out by the consensus? Every mature organization has one. The answer is not always welcome. The question is whether the structure for receiving it exists.
Third: Which of your current frameworks is the Bohr model — an almost completely false but extremely useful simplification that your teams have stopped treating as a simplification? The scaffold you have been teaching has become the ceiling. The question is whether you know which one.
Fourth: Is your engineering leadership optimizing to be useful or to be visible? The tell is in the priority stack when both are competing for the same resource.
Fifth: What ceremony did you cancel in the name of speed — and what coordination function is now being performed by nothing? The coordination collapse from missing ritual does not appear immediately. It compounds over weeks and quarters, invisible until friction makes it suddenly obvious that teams are no longer moving together.
The breakdown does not arrive. It was already there. The question is how far upstream you were willing to look.
