The Long Goodbye

It was always going to be a book.

You showed up.

That is not nothing. That is, in fact, everything. You have limited hours in a day and a nearly infinite supply of things competing for them. You chose to spend some of those hours here, week after week, on a subject (bureaucracy, institutional failure, the ways organizations systematically destroy the very value they were built to protect) that most people would rather not think about at all.

You not only thought about it. You engaged with it. When Unforgiven ran a few weeks ago, it described how the Bank Secrecy Act conscripts banks into becoming a forward-deployed compliance apparatus, and how non-profits like the Southern Poverty Law Center exploit this to get rich by diffusing accountability. People recognized the mechanism. They had seen it. They had lived inside it. Readers connecting a model to their own is what this publication was for.

Some of you I have met in person. I spoke about bureaucracy at the Dirtnap Conference. The conversations that followed with investors, operators, and people who had spent careers trying to understand why institutions fail the way they do confirmed what the blog had been telling me for years. This subject matters, the models are useful, and there is a much larger audience for the serious version of this argument than the conventional wisdom would suggest.

I am profoundly grateful for the readership, the engagement, and the caring about a problem the world persistently underestimates.

What the blog was for

The purpose of Closest Point of Approach was never to complain about bureaucracy. That is too easy and too cheap. The purpose was to build models. Predictive ones. Models that would let investors, leaders, and anyone who operates inside a large organization see what was coming before it arrived.

We examined bureaucracy as a delegation problem: accountability passed from hand to hand until no one holds it. We looked at it through one-way and two-way doors, asking why organizations with every incentive to experiment choose instead to calcify. We traced it through universities, banks, the civil service, tech giants, non-profits.

And in Other People’s Money, the post that perhaps most directly drew on my prior career as an investor, we examined what happens to risk when the people making decisions are insulated from its consequences. The bureaucrat spending institutional capital. The fund manager deploying someone else’s savings. The regulator whose career survives the failure of the policy she mandated. When the downside belongs to someone else, the incentive to be careful evaporates. The model that post built is one of the load-bearing walls of the book.

We also made the case that bureaucracy is more like cancer than a problem. You do not cure cancer. You put it in remission. There will always be some background level of bureaucratic noise because human beings populate organizations and cannot fully separate their personal interests from those of the entity they serve. The best you can do is catch it early, treat it aggressively, and remain vigilant. Once it has metastasized and planted its hooks in the culture, you are doing chemotherapy on a patient who needed surgery three years ago.

Those models were built here. That is what the blog was for.

What the book is for

The book is what happens when you impose a spine on the models.

The posts were the laboratory. The book is the conclusion. It is the moment you ask not just what you observed, but what it all adds up to. And there is a final section that has never appeared here.

It begins with a provocation: we have outgrown bureaucracy.

Bureaucracy was a technology. It was the solution that industrial civilization developed for the problem of coordinating large numbers of people across complex tasks in an environment where information was scarce and verification was expensive. Given those constraints, it made sense. Hierarchy. Standardized process. Trust established through credentials and titles rather than real-time data. It was never elegant. But it worked, after a fashion, for the world it was designed for.

That world no longer exists. That’s one reason why I ignored much of the theoretical writing on the topic. I saw it as outdated.

The digital complexity that has accumulated over the past forty years manifests itself as the volume and velocity of information, the pace of change, and the intricacy of the systems we now depend on. It was never something bureaucracy was capable of managing. It was overwhelmed before it knew it was overwhelmed. The form-filling and committee-sitting and memo-writing continued. But the world had already moved past them.

Now something more fundamental is happening. When agentic AI can verify claims, audit transactions, and execute decisions at machine speed, the assumption that underpins bureaucratic process, i.e., that trust must be established through elaborate human ritual because verification is costly, begins to collapse. If the cost of verification approaches zero, what do you need the apparatus for?

Can AI help us develop the trustless organization? What would that look like?

There are live experiments happening in the marketplace right now, and the view is still maturing. But it raises something more profound than an organizational design problem. When humans live inside a cybernetic system, in genuine, continuous interaction with machines that can think, decide, and act, what is the nature of human agency inside that system? What does accountability mean when the decision-maker is not fully human? What does institutional self-interest look like when the institution can automate its own expansion?

The bureaucracy posts were the diagnostic. This final section is the prognosis. One day we may look back at bureaucracy the way we look at the human appendix: a vestigial structure from an earlier stage of evolution, costly to maintain, occasionally dangerous, and destined for obsolescence.

That argument required a book. It could not be a post.

What happens here

New posts stop. Each week I will republish one prior post, not in the order they were written, but in the order that builds the book’s argument. Each republication will note where that post sits in the larger structure. Think of it as a curated reread: the experiment, organized.

If the book does what I believe it can do, the conversation that follows will be larger than anything we could have had here. I hope you will be part of it.

As we restructure the tools with which we do business, we also need to re-engineer our organizations. This is perhaps the more meaningful question.

What if we can develop a zero trust architecture for the company or the non-profit or the agency, one in which we never trust, but always verify? One in which accountability is automatic and transparent.

There’d be no stopping a company like that.