Nearly every major forecasting failure of the last twenty years shared the same hidden mistake.
Economists missed the financial crisis. Energy planners missed the electricity demands of AI.
Epidemiological models struggled with cascading behavioral responses during COVID. Political experts repeatedly failed to anticipate populist realignments.
The problem was not simply bad data or insufficient computing power. It was deeper: treating open systems as if they were closed ones.
We keep assuming the future behaves like a chessboard — fixed rules, known pieces, stable boundaries — when many of the systems that matter most behave more like clouds: evolving, generative, and capable of changing the game itself while we are still playing it.
And increasingly, our media environment makes this conceptual mistake worse. We demand “the hook,” “the takeaway,” “the three-point summary” precisely when reality is becoming least compressible.
The distinction matters more than most analysts acknowledge.
Some complex systems approach what we might call the closed ideal. Metabolic networks, power grids, digital protocols, and certain supply chains operate within relatively bounded possibility spaces. The components are largely known. The governing rules are comparatively stable. The interactions may be astronomically numerous, yet enumerable in principle. Uncertainty is real, but it is epistemic: we may not know everything, but we know the space in which things can happen. Better models, more data, and faster computation genuinely help.
Go is a closed system in exactly this sense. With roughly 10¹⁷⁰ legal board positions, it dwarfs chess by orders of magnitude, yet the rules are fixed, the board is finite, and the objective unchanging. When AlphaGo played Move 37 against world champion Lee Sedol in 2016 — a move so unexpected that Sedol briefly left the room — it was a genuine achievement: a system navigating an astronomical but bounded space and discovering patterns no human had seen.
AlphaGo succeeded because the board stayed fixed. History rarely does.
Many real problems behave as if the board could expand mid-game.
Open systems — economies, wars, pandemics, technological ecosystems, organizations, societies — do not merely evolve in time. They evolve in structure. New actors emerge. New couplings form. Cascades of second-order effects alter the conditions governing everything that follows. The crucial variables are not simply unknown. They do not yet exist.
This is the domain of contingency: the recognition that multiple futures remain genuinely possible, and that the future is not merely discovered but partially created through action, accident, innovation, and the interactions of agents attempting to anticipate one another.
The history of forecasting is largely a history of underestimating this.
Who imagined, at the invention of the automobile, that America would end up with more cars than drivers — or that this would generate not only new industries and infrastructures but entirely new ways of organizing cities, families, commerce, and daily life? The automobile did not merely satisfy existing demand for transportation. It transformed the space of what was possible.
Technologies do not merely solve problems. They create new possibility spaces.
Who imagined, after two decades of essentially flat per-capita electricity demand, that AI data centers consuming the power of mid-sized cities electricity demands of AI would suddenly rewrite the energy equation — straining grids, reviving dormant power plants, and forcing utilities to abandon forecasts underlying billions in capital planning?
Generative AI itself is an open-system event. The technology is not simply improving existing workflows; it is reorganizing labor markets, education, information ecosystems, and even energy infrastructure through interactions few forecasts anticipated only a few years ago.
The transition was not a deviation from trend. It was a transformation of the system generating the trend.
The 2008 financial crisis offers perhaps the starkest example. The models that failed were not naive. They were sophisticated, well-funded, and staffed by some of the most quantitatively capable people in the world. What they assumed — implicitly and structurally — was that the correlations governing normal times would continue to hold under stress. They treated a system capable of endogenous transformation as if it were merely complicated.
When the structure itself shifted, the models had no language for what was happening.
The models failed not because they were too mathematical, but because they mathematized the wrong kind of reality.
The forecasters who missed these shifts were not careless. They were applying closed-system thinking to an open-system world — treating the future as computable within a known space when what was actually unfolding was a transformation of the space itself.
Hindsight hides this from us. Once events occur, we construct narratives that make outcomes appear inevitable. The unrealized alternatives disappear because they leave no traces. We forget how open the world once was, how many paths remained available, and how different the future looked to intelligent observers living inside uncertainty rather than looking backward through it.
The philosopher Michael Oakeshott called this the “abridgment” of history: the reduction of contingent unfolding into tidy causal sequence. What disappears in the process may be what mattered most — the genuine openness of the situation and the real possibility of other outcomes. Historians sometimes call this “backshadowing”: narrating the past as if it had always been moving toward the present.
And we increasingly do the same thing to the future.
We project forward from the present as if the structure of the present were fixed — as if we were navigating a known board rather than one capable of expanding, contracting, or changing its own rules entirely.
Clocks and clouds capture these two orientations toward complexity.
Clocks are complicated. They contain many interacting parts, precisely engineered and operating in predictable ways. A skilled clockmaker can disassemble one and reassemble it. Given the current state, the future state is, in principle, calculable. The metaphor suits much of classical physics, large portions of engineering, and a surprising amount of economic modeling.
Clouds are complex. They consist of innumerable interacting elements, but the interactions are nonlinear, adaptive, and generative of emergent properties that cannot simply be inferred from the parts. You cannot take a cloud apart and reconstruct it. The future state is not merely unknown; it is, in a meaningful sense, undetermined. Small perturbations ramify. Structure evolves. What the cloud is at any moment cannot be separated from how it is changing.
Most of the problems that matter most — political orders, technological transitions, ecological systems, financial markets, geopolitical alignments — are clouds that we persist in treating as clocks.
The persistence is understandable. Clock thinking offers something cloud thinking cannot: tractability, precision, and the comfort of definite answers.
But precision attached to the wrong model is not clarity.
A clock can be optimized. A cloud must be navigated.
The appropriate response to a clock problem is a better forecast. The appropriate response to a cloud problem is something harder to name — call it orientation: a stance toward uncertainty that preserves optionality, attends to weak signals, and remains capable of revision when the structure itself begins to shift.
Which brings us back to the hook.
“The hook,” “the thesis,” “the key takeaway,” “the three-point summary” — this is clock thinking applied to ideas. It assumes that arguments can be compressed without distortion: that the essential content can be separated cleanly from its form, and that structure is merely packaging.
Sometimes this is true. Some arguments are genuinely summarizable. The periodic table is not diminished by compression. A well-defined mathematical theorem can often be stated more elegantly than the derivation required to prove it.
But some arguments are not summaries of themselves.
The unresolved tensions are not failures of compression. They are the substance of the argument.
Strip them away, and what remains is not the same idea made more accessible. It is a different idea — reshaped to fit the available space and the attentional demands of the moment.
This is the paradox of compression in an age of open systems: the ideas we most need — the ones that preserve contingency, resist premature closure, and keep multiple futures visible — are often the ideas least compatible with the forms through which modern discourse now travels.
The medium increasingly selects against the message.
Compression itself can erase contingency.
None of this is an argument against clarity. It is an argument for calibration: recognizing what kind of problem one is facing before deciding what kind of cognitive tools to apply.
Not all systems reward prediction equally. Not all realities compress equally. Not all arguments survive reduction intact.
In a closed system, the right response may be a better forecast.
In an open one, the deeper task is maintaining orientation while the landscape itself changes.
The danger is not uncertainty itself. The danger is false closure.
Some realities can be summarized. Others can only be navigated.
