On my way back from Australia, I had a chance to reflect on years of writing and wrote this piece as a consequence.
Over the years I have worked through nearly every format available to someone who thinks about defense and strategy for a living. I have given lectures and briefings, sat for interviews, written more articles than I can count, and published a number of books. From the outside these look like variations on a single activity: explaining what I think. They are not. Each one engages a different part of the mind, and each one produces a different kind of knowledge. Understanding those differences is why, in an era when the incentives all point toward shorter and faster, I keep returning to the book.
I have come to think of these formats less as a menu of options and more as a division of cognitive labor, each with its own discipline and its own failure mode. Knowing which one a given piece of thinking actually needs, not which one is fastest, or most visible, or most rewarded by the current incentive structure, is itself part of the work. What follows is an attempt to lay out that division as I have experienced it across four decades.
Lectures and Briefings Are Thinking Out Loud with Other People
A lecture or a briefing is fundamentally a social act of cognition. You arrive with a structure, but the room shapes what happens next. An officer asks the question you had not thought to ask. A skeptic in the third row exposes the soft joint in your argument before you have noticed it yourself. A briefing in a headquarters, where the audience carries the operational burden of whatever you are describing, is an unforgiving test: people who have to live with the consequences of an idea will tell you quickly whether it survives contact with reality.
What I value here is the jointness of it. Ideas get developed in the room, not merely delivered to it. Some of my most useful formulations were not written first and then presented; they emerged mid-sentence in front of an audience, because the presence of intelligent resistance forced a sharper phrasing than solitude ever would. There is a kind of thinking that simply does not happen alone at a desk. It needs another mind pushing back in real time, and a briefing room is one of the few settings that supplies that pressure on demand.
The limitation is equally clear. The insight is momentary and unanchored. You cannot hold thirty interlocking arguments in a room for ninety minutes. The format rewards clarity and compression, and it punishes the kind of extended reasoning that needs a hundred pages to do its work. A briefing that tries to carry too much collapses under its own weight; the audience remembers the one slide that overreached, not the nine that were sound. So the craft of briefing is really the craft of subtraction, deciding, under time pressure, which nine-tenths of what you know has to stay out of the room.
Interviews Are Disciplined Listening
Interviews are the other half of that conversational process, but with the direction of flow reversed. When I sit down with a squadron commander, a maintainer, a program manager, or a foreign chief of staff, my job is to build the conditions under which someone else’s expertise can surface in its own language. The thinking process here is inductive and patient. You listen for the detail that does not fit the official narrative, the offhand remark about sustainment that turns out to be the whole story, the frustration that reveals a structural problem no briefing chart will show.
That work produces raw material of a quality you cannot get any other way. But raw material is exactly what it is. A single interview tells you what one person sees from one position at one moment. Its meaning is latent, and it becomes visible only when set against dozens of other conversations, across services, nations, and years. The discipline here is resisting the temptation to treat a single vivid quotation as a conclusion. It is a data point, not a finding, and the difference between the two is patience.
Over time this accumulates into something like a longitudinal instrument. The same question asked of a maintainer in one service and a program manager in another, three years apart, starts to triangulate on something neither of them could see individually. That triangulation is invisible in any one transcript. It only exists in the analyst’s memory, or in the notes, until a larger format gives it somewhere to live.
Articles Serve Someone Else’s Thinking
An article does something quite specific: it takes one theme, narrows the aperture, and puts a clean input into other people’s thought processes. That is a genuinely useful function and I have no quarrel with it. A well-made piece of two thousand words can shift how a community frames a problem, and it can do so quickly, while the question is still live. There is a real satisfaction in that kind of timeliness in catching a debate while it is still open and giving it one useful turn.
But the constraint is real. An article assumes its context rather than building it. It borrows a frame the reader already holds, adds one increment, and stops. It is a contribution to an argument someone else is conducting. Over time I noticed something about my own articles: the best of them kept gesturing at a larger structure they had no room to build. The footnotes were doing the work the text could not, each one a compressed pointer to an argument that needed its own chapter and was not going to get one inside two thousand words.
I have also noticed that articles age differently from books. A good article is sharp precisely because it is anchored to a moment; that is its strength and its expiration date. Read five years later, it tells you what someone thought was urgent at the time, which is its own kind of historical evidence, but it rarely still functions as the argument it was written to make. A book, done properly, is built to survive its own publication date.
The Book Is Where the Connections Live
A book gives me something no other format does: the intellectual space to connect a much wider range of ideas and data, and to argue a comprehensive set of themes rather than a single point.
This is not a matter of word count. It is a different cognitive process. In a book I have to decide what actually depends on what. I have to test whether a pattern I observed in Australian force design genuinely illuminates something in the North Atlantic, or whether it merely sounds similar. I have to carry a line of argument across chapters, which means the argument has to survive its own extension and many do not. Some of my most valuable discoveries have come from watching a favored idea fail at chapter seven, when it had looked entirely sound as a conference paper.
That failure is not a setback; it is the point of the exercise. An idea that only works at two thousand words, in isolation, unconnected to the rest of what you know, was never as sound as it looked. A book is the only format that forces the idea to answer to everything else you believe, and a surprising number of ideas that survive a briefing room do not survive that wider interrogation. I would rather find that out in manuscript than in print.
A book also forces integration across the other formats. The interviews become evidence. The briefing questions become the objections I am obliged to answer. The articles become components inside a larger frame that finally explains why each of them mattered. Nothing else I do performs that synthesis. Without it, I would have a career’s worth of accumulated observations and no structure holding them together, a filing cabinet, not an argument.
And a book has a duration that matches the phenomena I study. Military transformation, alliance adaptation, the shift toward autonomous systems, the reindustrialization of Western defense and these are not events. They are processes measured in decades, with internal contradictions that unfold slowly. A format that resolves in fifteen minutes or two thousand words cannot hold a decade-long process in view long enough to see its shape. Only a format with enough duration of its own can match the duration of the thing being studied.
The Discipline of Writing Under Contradiction
What a book demands, more than length, is a tolerance for holding contradictory evidence in view at the same time without resolving it prematurely. A briefing has to land somewhere in ninety minutes; an article has to land somewhere in two thousand words. A book can afford to say: here is what points one way, here is what points the other way, and here is why the tension between them is itself the finding. Some of the arguments I most trust in my own work are the ones that took the longest to stop wobbling where I kept two apparently incompatible sets of field observations on the desk for months before I understood that they were both true, in different parts of the system, at different points in the process.
That kind of patience is structurally unavailable to shorter formats, and it is exactly the kind of patience that complex institutional change requires. Alliances do not transform cleanly. Programs do not fail or succeed in a single reportable event. The honest account of most of what I study is contradictory in real time and only resolves into a pattern in retrospect which means the format doing the writing has to be able to sit with that unresolved state long enough for the pattern to emerge.
Why This Matters More in Dynamic Change, Not Less
The common objection is that a world moving this fast has outrun the book. Events change under your feet; by the time a manuscript is printed, the situation has moved. I think this gets the problem precisely backwards.
Rapid change is exactly the condition under which comprehensive thinking becomes indispensable. When everything is in motion, the scarce commodity is not information. There is a torrent of it and the need is to structure: a framework that tells you which changes are noise and which are signal, and how the pieces relate. Fast formats are very good at registering change and very poor at interpreting it. They tell you that something happened. They rarely tell you what kind of thing it was.
Books do the interpretive work. They build the frameworks that later let you read the daily flow of events quickly and well. Every sharp judgment I have been able to make in a briefing on short notice rested on structural work done slowly, years earlier, in a book manuscript where I had the room to think something all the way through. The speed people admire in a good briefer is usually borrowed speed, capital drawn down from a much slower investment made somewhere else, and a book is where I make that investment.
There is also a quieter reason the book matters more, not less, right now. When the volume of fast commentary goes up, the marginal value of one more fast take goes down, but the marginal value of a framework that actually organizes the flood goes up. A field saturated with hot takes is starved for structure, not starved for content. That is the gap a book is built to fill, and it is a wider gap today than it was when I started doing this work.
So I keep writing them. Not out of attachment to an older form, and not because the other activities are lesser, each does something the others cannot.
But the book remains the only place where I am required to connect everything I have learned and defend the whole of it at once. In a world of accelerating change, that discipline is not a luxury.
It is the thing that makes the rest of the work intelligible.