The undetermined future is Solnit's most radical epistemological claim and the foundation of her entire political philosophy. The future is not a destination toward which humanity travels along a route that can be mapped; it is a space of possibilities—some visible, most not—shaped by every choice, institution, norm, and small act that the present contributes to it. The space is genuinely open. This openness is not a deficiency in our predictive models but the actual structure of reality. The desire for certainty—for the expert who knows, the model that predicts, the narrative that resolves—is the desire to close the space of possibility, to convert the genuinely open into the comfortably determined. But the closing is itself a political act, because the people who claim to know the future shape it not because their predictions are accurate but because their certainty displaces the participation of everyone else. If the experts know, there is nothing for the non-expert to contribute.
Solnit's argument builds on historical cases where outcomes that looked impossible became actual because people acted without foreknowledge. No analyst in 1988 predicted the Berlin Wall would fall in 1989—the Wall's permanence was embedded so deeply in geopolitical structure it was invisible. A confused press conference, bureaucratic miscommunication, and a crowd showing up at the border produced a cascade no one designed and no one could stop. The permanent thing was not permanent. The assumption organizing an entire civilization's understanding of what was possible turned out to be contingent. The lesson is not that good things happen unexpectedly (though they do) but that the most important changes are, by definition, the ones the existing framework cannot predict—because if the change could have been predicted, it would have been assimilated into the framework rather than reshaping it.
Applied to AI, this means the consequences currently dominating discourse—job displacement, productivity gains, creative disruption—are the legible consequences, the ones that fit existing categories and can be debated with existing vocabulary. The consequences that will actually determine the trajectory are the ones that do not yet fit any category—the Berlin Wall moments, the cascades no one designed, the possibilities no framework contains. Solnit cannot tell us what those will be. Neither can the builders at the frontier, the researchers studying impacts, or the regulators attempting governance. The significant future is the one that cannot be predicted, which means the posture of certainty—whether optimistic or pessimistic—is always a form of blindness.
The undetermined future is not an invitation to fatalism ("nothing can be known, therefore nothing can be done") but to participation. Because the future is open, the institutions built now, the norms established now, the governance frameworks created now will shape which possibilities become actual. The labor movement did not know whether organizing would produce the eight-hour day, but the eight-hour day was produced because they organized. The undetermined future is the future in which organizing matters—not because it guarantees success but because success is possible only if the organizing happens.
The concept of the radically open future has roots in existentialism (Sartre's "existence precedes essence"), pragmatism (Dewey's experimental method, James's pluralistic universe), and the philosophy of history that rejects both Hegelian teleology and Marxist determinism. Solnit's version is distinctive in its empirical grounding—she documents not philosophical arguments but historical cases where the future turned out to be more open than the present's categories could contain. The fall of the Berlin Wall, the end of apartheid, the development of AIDS treatments, the Zapatista uprising—each was impossible until it happened, and each happened because people acted on the possibility without knowing whether their action would matter.
The AI discourse, Solnit argues, is structured by the opposite assumption: that the future is determined either toward triumph (the accelerationist narrative) or catastrophe (the doomist narrative). Both narratives close the space of possibility and thereby eliminate the ground on which agency operates. Solnit's insistence on the undetermined future is the philosophical foundation for her insistence on participation: show up, build institutions, tend norms, make the daily choices that accumulate into trajectories—not because the outcome is guaranteed but because the outcome is genuinely uncertain and therefore genuinely depends on what people do.
The Future Is a Space, Not a Destination. Treating the future as a point toward which history moves along a trajectory is a category error. The future is a field of possibilities shaped by present choices, institutions, norms, and the accumulated small acts demonstrating that alternatives exist.
Certainty Produces Passivity. If the accelerationist is correct that AI will democratize capability, there is no need to fight for institutional structures ensuring distribution. If the catastrophist is correct that AI will concentrate power, there is no point in building governance alternatives. Both certainties eliminate the space where participation matters.
The Significant Change Is the Unpredictable Change. The consequences that fit existing frameworks are the ones that will be assimilated into existing power structures. The consequences that reshape frameworks are the ones no existing vocabulary can contain—which means the most important AI consequences are the ones nobody has anticipated.
Prediction as Political Act. The claim to know the future is always a claim to power, because certainty displaces the participation of everyone who is uncertain. The futurist, the strategist, the model—all perform the same function: closing the space of possibility by claiming it is already known.
Participation Makes Futures Actual. The undetermined future becomes determined not through the unfolding of a pre-written script but through the institutional choices, daily practices, and organized efforts of people who cannot see the outcome and act anyway. The acting is not secondary to the outcome; it is constitutive of it.