How technologies harden into infrastructure and why the window for intervention is early.
Some technologies become a condition.
A technology gets used because it solves the problem in front of someone. That use becomes visible. Other people follow. Alternatives get harder or more expensive. Eventually the original design stops being evaluated at all.
Then other systems get built around it. At that point, changing course no longer means replacing one technology. It means rebuilding everything that came to depend on it.
I call this Systemic Topography.
The point of this framework is timing. We are very good at explaining lock-in after it has happened. That is usually when the evidence is clearest and the leverage is gone. I am interested in the earlier period, while the shape of the system is still being negotiated, often without anyone realizing that is what is happening.
Read the paper →Five stages describe the process.
The mechanism is not always the same. I use three kinds of terrain to distinguish what is hardening and where intervention is still possible.
The corner castings on a shipping container are eight steel fittings whose dimensions were fixed more than sixty years ago. Every crane, chassis, ship cell guide, and yard on earth now assumes them.
Robert Moses built the parkway bridges on Long Island low. Whatever the intent, the clearance decided for decades who could reach the beach, because a bus could not pass underneath. Physical dimensions outlive the arguments that produced them.
A system called COMPAS scores the likelihood that a defendant will reoffend. Once it was in front of a judge it became a fact in the room, and the system began generating the record that appeared to confirm it.
A version of this cycle is now moving unusually fast. The interfaces that agentic AI systems use to reach the rest of the world are being determined now, driven by the need for systems to identify themselves, request access, exchange information, transact, and act.
This looks technical at the use stage but does not stay this way. Protocols set because they are efficient today can become the terrain every company, institution, and machine has to cross tomorrow.
This is where the framework overlaps most directly with my investing and board work. I am close to emerging technologies before the ecosystem around them makes those design decisions expensive to revisit.
Systemic Topography is not a way to predict which technology wins.
It is a way to ask where we are in the process, what is becoming expensive to reverse, and therefore what kind of leverage still exists.
For an investor, that may mean distinguishing a durable interface from a temporary product advantage. For a technologist, it may mean recognizing that a protocol decision is more consequential than it first appears. For a public official or standards body, it may mean intervening before the practical cost of doing so becomes prohibitive.
The framework describes what happens to technologies and systems as they harden. The question I am working on now centers around people inside those systems.
As more of the world is shaped by defaults, recommendations, models, and automated decisions, which parts of judgment do we want to keep practicing ourselves?
What I'm working on now →Write to me about a company, a question, or a room you are organizing.