The Dam as Anti-Plan

In the AI Story

Scott's later work, particularly Two Cheers for Anarchism, articulated the principles that distinguish the dam-as-structure from the comprehensive plan. The principles are not complicated. Start small and observe. Prefer reversible interventions. Build in feedback mechanisms that measure effects, not compliance. Tolerate messiness. Trust the practical knowledge of those closest to the ground. These principles are not novel — they are as old as the practice of building structures in rivers. What makes them urgent is that they are precisely the principles the dominant mode of AI governance is structured to ignore.

The Berkeley researchers whose work Segal examines in The Orange Pill proposed something they called 'AI Practice' — structured pauses built into the workday, sequenced rather than parallel workflows, protected time for human-only engagement with the work. These are dams. Small, local, responsive structures built by practitioners who have observed, through direct engagement, what AI does to the texture of cognitive work and who are designing interventions calibrated to the specific effects they have witnessed. The structures are not comprehensive. They do not scale automatically. They require local knowledge to implement well and continuous maintenance to sustain. They are, in every respect, the opposite of the comprehensive AI strategy — and, if Scott's framework holds, far more likely to succeed.

The metaphor connects directly to Edo Segal's central image in The Orange Pill. Segal invokes the beaver as a model of appropriate response to a powerful current of intelligence that cannot be stopped. Scott's analysis sharpens the metaphor: the dam cannot be designed from above. It must be built by those who feel the current — the practitioners whose daily engagement with AI gives them knowledge that no policy document can contain. And building the dam requires not just the will to act but the institutional infrastructure that allows practitioner knowledge to inform decisions.

The anti-plan is not the absence of intention. It is the presence of a different kind of intention — one that is provisional, iterative, responsive to feedback, and willing to abandon its own conclusions when conditions change. This is harder than comprehensive planning, because it requires sustained engagement rather than one-time design. It also requires trust in practitioners whose knowledge cannot be fully articulated — a trust that institutional governance is structurally inclined to withhold.

Origin

The framework emerged through Scott's comparative study of what actually worked in the systems he examined alongside what failed. The successful adaptations — the informal arrangements that sustained peasant communities, the uncoordinated practices that made cities livable, the small-scale experiments that produced genuine improvements — shared the features Scott eventually named. The beaver metaphor itself comes partly from Edo Segal's formulation in The Orange Pill and is applied here as a lens through which to view Scott's positive program.

Key Ideas

Local, responsive, distributed. The three features that distinguish the anti-plan from the comprehensive plan — and that correspond directly to the features of métis that comprehensive planning systematically excludes.

Start small and observe. The anti-plan begins with observation of what the intervention is actually doing, not with a theory of what it should do.

Reversibility as a design principle. Anti-plans favor interventions that can be adjusted when conditions change. Comprehensive plans favor interventions that are durable, which often means they outlast the conditions for which they were designed.

Messiness as feature, not bug. The anti-plan tolerates the inconsistency that distributed, locally calibrated interventions produce. Comprehensive plans aspire to consistency — a feature that is attractive in principle and catastrophic when applied to inconsistent conditions.

Debates & Critiques

Critics have argued that the anti-plan framework is insufficient for addressing large-scale coordination problems — climate change, global AI governance, pandemic response — that require exactly the kind of comprehensive, coordinated action Scott resists. Defenders respond that Scott's framework does not reject coordination but rejects the assumption that coordination requires centralized design. Elinor Ostrom's work on polycentric governance represents the most developed attempt to specify how distributed structures can achieve the coordination that large-scale problems require. Whether the anti-plan approach can scale to match the scale of AI governance challenges remains an open empirical question.