Hope as practice is Rebecca Solnit's central philosophical contribution, developed most fully in Hope in the Dark (2004). Unlike optimism, which expects a good outcome regardless of action, and unlike despair, which expects a bad outcome regardless of action, hope recognizes that the future is genuinely undetermined and that this uncertainty is precisely what makes human participation meaningful. Hope is not passive expectation but active engagement—showing up, building institutions, tending relationships, asking questions—without the guarantee that the effort will succeed. In the AI transition, where both triumphalist certainty ("AI will democratize everything") and catastrophist certainty ("AI will destroy everything") produce the same practical result—passivity—hope becomes the operational alternative: the discipline of building dams, designing curricula, fighting for governance frameworks, not because success is guaranteed but because the undetermined future is the only kind worth working for.
Solnit developed the concept in the early 2000s, writing against the paralysis that had gripped the American left after the Iraq War invasion and the despair following George W. Bush's 2004 re-election. The distinction between hope and optimism emerged from her historical research into social movements: suffragists, civil rights organizers, labor activists, environmentalists. In every case, the people who changed history did not act because they knew their efforts would succeed. They acted because the outcome was uncertain, and the uncertainty meant their participation might matter. The 2016 update to Hope in the Dark—written after the 2016 U.S. presidential election—sharpened the distinction further, arguing that despair is a form of certainty as corrosive as naive optimism, and that both produce the same abdication of responsibility.
Applied to the AI moment, hope as practice means specific, granular engagement with the institutional structures that will determine how AI's power is distributed. It means the teacher redesigning her curriculum around questions AI cannot originate. It means the executive choosing to grow her team rather than converting productivity gains into headcount reduction. It means the parent creating spaces for boredom in a child's attention-saturated life. It means the community organizer demanding algorithmic transparency from municipal governments. None of these acts guarantees a good outcome. Each demonstrates that an alternative to the default arrangement exists, and demonstrations of possibility are how landscapes change—not through grand strategies but through the accumulation of small, local, often invisible acts that prove the alternative works.
Solnit's hope is not sentimental. It acknowledges the damage—communities displaced, attention colonized, depth eroded, meaning extracted. It refuses the inference from damage to defeat. The damage is real; the defeat is not yet determined. The gap between them is the space in which hope operates, and the operation is not emotional but practical: building the institutions, fighting the fights, showing up to the governance conversations, maintaining the dams that redirect the river's flow toward life. The practice does not promise success. It makes success possible, which is the only honest promise available in conditions of genuine uncertainty.
The phrase "hope in the dark" itself comes from a line often attributed (incorrectly) to Václav Havel but actually drawn from Solnit's synthesis of multiple sources: Havel's essays on dissent, Kafka's aphorisms, and the lived experience of activists who worked for decades without seeing results. The 2004 book was written quickly, in the months after the Iraq invasion, as a direct response to the despair Solnit was encountering in her communities. It was reissued in 2016 with a new introduction addressing the Trump election, and again found a wide audience among people searching for a vocabulary that was neither naive optimism nor resigned cynicism.
Solnit's formulation shares structural features with other frameworks developed independently: Joanna Macy's "active hope," Viktor Frankl's logotherapy (meaning-making under conditions of suffering), and Cornel West's "audacious hope" grounded in the Black prophetic tradition. What distinguishes Solnit's version is its empirical grounding—every chapter of Hope in the Dark documents a specific historical case in which uncertain action produced unexpected outcomes, from the fall of the Berlin Wall to the development of antiretroviral AIDS therapies to the Zapatista uprising.
Uncertainty as Precondition for Agency. If the future were determined, participation would be irrelevant. The genuinely open future is the only kind in which human choice matters, which makes uncertainty not a deficiency in analysis but the structure of freedom itself.
Action Without Guarantees. The suffragists organizing in 1848 did not know women would vote in 1920. Civil rights workers sitting at lunch counters in 1960 did not know the Civil Rights Act would pass in 1964. They acted because the outcome was uncertain, and the uncertainty made action meaningful rather than futile.
Hope Requires Work. Hope is not a disposition one possesses but a practice one maintains—daily choices about where to direct attention, what institutions to build, which fights to fight. In the AI context, this means showing up to governance conversations, redesigning educational systems, demanding transparency, building cooperatives—specific institutional labor without guaranteed outcomes.
Despair and Optimism Are Mirror Images. Both treat the future as determined—despair assumes it's determined toward the bad, optimism assumes it's determined toward the good. Both produce passivity. Hope is the third position that recognizes the outcome depends on what people do next.
The Darkness Contains Light. Dark times are not times without hope but times when hope is hardest to see—scattered in local acts, invisible to dominant narratives. The light exists in teachers redesigning curricula, executives choosing to grow teams, developers in Lagos building products, communities organizing for transparency. Seeing this light requires looking where the metrics don't measure.