As We May Think
Source: https://www.theatlantic.com/magazine/archive/1945/07/as-we-may-think/303881/ ↗
Point zero. Bush imagined the Memex in 1945 — a machine for augmenting human memory through associative trails.
Every hyperlink, every wiki, every recommendation system is a partial realisation of this essay.
Read it not as prediction but as the clearest statement of what personal computing was supposed to be for.
Central argument
Bush argues that the dominant bottleneck in human progress is no longer our capacity to produce knowledge but our ability to access and connect it — scientists are drowning in a literature they cannot navigate. His solution is the Memex: a personal desk-sized device storing books, records, and communications, navigable through 'associative trails' that mimic the way the human mind actually works, by association rather than by index. The central thesis is that machines should augment cognition by externalising and sharing these associative structures, not merely store data.
Critique
Bush's vision is stubbornly individualist — the Memex is a personal device, and trails are built and shared voluntarily by a solitary scholar. This misses the social and political dynamics of how knowledge is actually constructed, contested, and gatekept: who builds the trails, whose associations are treated as authoritative, and what gets excluded are questions the essay never raises. Eighty years of platform design have shown that associative architectures, left ungoverned, amplify existing power asymmetries rather than dissolve them.
Why it matters for product
Bush's framing of 'associative trails' is a precise diagnostic tool for product discovery failures: most internal knowledge systems — roadmaps, research repositories, decision logs — are indexed hierarchically and therefore die as organisations scale, because no one can find the reasoning behind past decisions. A CPO can use this essay to challenge why their organisation builds search when it should be building trails — persistent, human-authored links between insight, decision, and outcome that survive team turnover and compound over time.