Reflections 19th April
The nature of invisible debt.
Growth is a ravenous beast; all the more so when the growth is demanded by those who don’t have skin in the game. Then it is all about appearances; a convincing story to tell, knowing people won’t check the substance.
It is the mentality and culture of Potemkin villages, something rapidly put together and just convincing enough to fool somebody passing by without stopping to look carefully. People have built entire careers creating corporations that are Potemkin villages, where instead of paint, it is a series of acquisitions and ambitious, if flawed, strategies that beguile the unwary. Caveat emptor: networks of privilege, easy debt, and golden parachutes.
There is something particular about debt incurred without the intention of repaying it. It is not recklessness or malign intent. It is calculation: that the bill will fall due after you have left the building, taken the bonus, and moved on. The debt becomes someone else’s problem. It is the sort of debt that administers the coup de grâce to a wounded company like Denby Pottery, where over two hundred years of craft history is likely to be brought to an end by debt with no accountability.
Finance is not the only sort of debt to be treated this way. It is as true for technical debt as it is for cultural, reputational, or geopolitical debt, and the same logic runs through all four. Borrow against the future. Dress the facade. Keep moving.
What we tend to forget is how old this pattern is, and how deeply it is woven into the fabric of moral life. The anthropologist David Graeber spent a career reminding us that the language of reckoning, redemption, and guilt is not a metaphor borrowed from finance. Finance borrowed it first. In Sanskrit, Hebrew, and Aramaic, the words for debt and guilt were, for centuries, the same word. The person who builds a Potemkin village and moves on before the facade falls is not merely making a financial miscalculation. They are doing something older and darker than that, and we are still fighting these battles without quite knowing it.
Technical debt is not a sign of failure but a fundamental characteristic of how complex systems evolve. It represents the gap between the messy reality of the present and the cleaner standards of the future. Taking on debt is often a strategic necessity: to hit a market window, test a concept, or keep pace with a competitor. The interest is paid in the currency of developer time and system stability. The most resilient organisations treat technical debt as a revolving door. They don’t aim for a perfect, debt-free system, which is often a mirage in a fast-moving industry, but focus instead on sustainable management. By allocating time to repay the principal through refactoring and documentation, they ensure that the cost of past decisions never outweighs their capacity for future innovation. The discipline is not elimination but maintenance: ensuring that the choices made yesterday don’t rob you of the ability to move tomorrow.
The IT estates of many large organisations, seen clearly, bear a closer resemblance to favelas than to planned infrastructure. Not chaotic, exactly. A favela has its own internal logic, its own improvised connections, its own surprising resilience, but the connections are unmapped, the dependencies undocumented, and the whole becomes increasingly difficult to navigate or extend. Each system was a reasonable response to a felt need at the time. Nobody designed the mess. It accumulated, decision by decision, acquisition by acquisition, until the organisation found itself living inside a structure it no longer fully understood. Like a favela, it persists not because it is good but because too much now depends on it, and the cost of removal is too high to contemplate.
Cultural debt accumulates more quietly and compounds more dangerously. It builds whenever an organisation consistently chooses speed, convenience, or appearance over the slower work of building genuine trust, honest communication, and shared understanding. Each shortcut is defensible in the moment, but the aggregate is corrosive over time. This kind of debt lives in the informal register of human obligation, in what people owe each other that cannot be put on a balance sheet or entered into a system. When it is violated, the language that surfaces is not financial; it is moral. The person who knew why a particular decision was made has left. The trust that would have allowed a difficult conversation to happen was never quite established. The habit of working through genuine difficulty together was never developed, so when the crisis arrives, the organisation discovers it does not have the relational reserves it believed it had. What looked like culture was closer to compliance, and what felt like alignment was proximity. Remove the proximity, through a restructure, a departure, or simply the passage of time, and what remains is what was actually there. What is missing is a form of mētis, invisible until it is no longer there.
Reputational debt is, perhaps, the most dangerous of the three organisational forms, because it is the only one that cannot be repaid quietly. Technical debt can be refactored. Cultural debt can be rebuilt through patient, sustained work that requires no announcement and no audience. Reputational debt, once it falls due, plays out in public. The gap between the story an organisation has told about itself and the substance behind it becomes visible simultaneously to everyone: employees, customers, investors, journalists, and regulators.
John Muir, writing about the natural world, observed that when you try to pick out anything by itself, you find it hitched to everything else in the universe. The same is true here. The reputations of organisations are hitched to their cultures, their cultures to their systems, and their systems to the decisions made by people who have long since left the building. When one thread is pulled, others follow. The Potemkin village always depended on the audience keeping its distance and maintaining its speed. Stop, and look at the back of the facade, and the picture changes entirely.
The same logic, extended to its fullest reach, describes what is happening to the structures by which nations and societies hold themselves together. The debt here is not paid in developer hours or damaged brand equity. It is paid in something far heavier: in lives, in the slow erosion of trust between peoples, in the fracturing of alliances that were assumed to be self-sustaining because they had endured for so long. The rules-based international order, the solidarity of democratic alliances, the credibility of institutions built painstakingly after the catastrophes of the twentieth century: all of these required continuous investment in things that rarely appear on any plan. Genuine political will. The willingness to bear costs in the present against future stability. The patient maintenance of relationships that only reveal their value when they are tested.
That investment was deferred, repeatedly and across many administrations and many countries, by people who calculated, not always wrongly in the short term, that the structures would hold without it. The facade was convincing enough, and nobody was stopping to look carefully. Now the reckoning is arriving, not as a choice but as a consequence, and the debt is being paid by people who had no say in its accumulation. This is not just a political observation. It is the same structural pattern, at civilisational scale. The logic of the Potemkin village does not change when the village becomes a world order.
What is genuinely new is not the debt. It is the instruments now doing the looking. For most of the period in which these debts accumulated, the people who incurred them and those who might have exposed them were operating within the same human systems, subject to the same professional courtesies, the same institutional blind spots, and the same quiet incentives not to look too carefully. Auditors who depended on client relationships. Consultants whose next engagement depended on the last one having gone well. Boards that lacked either the time or the technical literacy to interrogate what sat beneath the surface. The facade held not because it was convincing to everyone, but because the people with the means to look carefully rarely had compelling reasons to stop and do so.
That condition is changing faster than most organisations have registered.
Tools like Anthropic’s Mythos, mapping the architecture of a codebase with systematic thoroughness, have no career to protect and no relationship to preserve. They work out what is there and where the fault lines lie. They do not know that the system they are documenting was someone’s defining project, built under pressure in circumstances that felt exceptional at the time. They simply surface the debt, dispassionately and at scale, in a form that is legible to people who previously had no means of seeing it.
Review platforms have done to employer brands what Mythos does to codebases: aggregated honest testimony at scale, anonymously, in a form that persists and is searchable by anyone who cares to look. The gap between the culture described in the recruitment brochure and the culture experienced on the inside was always there; what changed was the instrument. Supply chain mapping tools, increasingly AI-assisted, are doing the same to sustainability claims, tracing the provenance of assertions that were previously unverifiable because nobody could follow the thread all the way back. The story was convincing enough for anyone not to stop to look carefully. The looking is now automated, and it does not tire.
This pattern has historical precedent. Socrates worried, in Plato’s Phaedrus, that writing would create the appearance of knowledge without its substance, that people would seem to know things they had merely read, becoming hearers of many things while learning nothing. The printing press, centuries later, revealed how much institutional authority rested not on genuine legitimacy but on the controlled scarcity of information. Each new capability has a way of exposing what the previous arrangement was quietly depending upon. The current moment is no different in kind, though it may prove to be in scale and speed.
What the instruments are finding, then, is not simply code that needs refactoring. Beneath the technical layer lies cultural debt: the accumulated cost of years of choosing appearance over honest reckoning, of managing an illusion of culture rather than building it, of filling roles rather than developing people. Beneath that, reputational debt: the gap between the story and the substance. And beneath all of it, running deeper than any single organisation or institution, the geopolitical debt: obligations deferred, relationships undermaintained, structures assumed to be permanent that were, all along, dependent on investments that quietly stopped being made.
The Potemkin village always depended on the audience keeping its distance. The instruments we have developed no longer do. There is, increasingly, nowhere to hide.
The question worth sitting with is not whether the debt will be called in; that process has already begun, in ways most organisations and most governments are only just starting to understand. The harder question is whether those now facing the reckoning still possess the capacity that the work of underpinning actually requires: patient, local, judgment-laden, and relational, the kind that cannot be automated, cannot be acquired through a restructure, and cannot be dressed up to look like something it is not.
The instruments we have built may yet prove to be our salvation rather than our nemesis, if only we choose to repay the debts we have built up to those who were given no choice other than to finance them.
OpenAI’s recent proposal for an industrial policy for the intelligence age, published this week in Noema Magazine, is a striking case in point. One of the very companies whose instruments are exposing the depth of accumulated debt has now produced a thirteen-page document arguing that capitalism as currently configured cannot manage what is coming, and calling for wealth funds, redistributive taxation, and a new social contract.
Whether this represents genuine reckoning or a more sophisticated form of facade management is a question worth sitting with carefully. The gap between the story and the substance, after all, is precisely what the instruments are now trained to find.
I will come to the OpenAI proposal in my next post. In the meantime, I would invite you to read it for yourself and ask one simple question:
Is this the sound of underpinning work beginning, or is it the sound of paint being applied to a very large and rather elegant Potemkin village?
