In 1945, before personal computers, before the mouse and decades before the public Web, Vannevar Bush described a desk that could help a person move through information by association rather than by filing cabinet.
He called the imagined machine the memex.
It was not a digital computer and it was not the internet. Bush pictured microfilm, screens, mechanical controls and a personal archive. Yet one idea inside the machine still feels strikingly modern: a user could connect pieces of information into “associative trails” and later follow those trails again.
The problem was information overload before the digital age
Bush was an engineer, MIT professor and one of the most powerful American science administrators of World War II. As head of the Office of Scientific Research and Development, he coordinated large parts of the U.S. wartime research effort. By 1945 he was thinking about a different problem: science was producing knowledge faster than people could organize and use it.
His essay “As We May Think,” published in The Atlantic in July 1945, began from that bottleneck. The difficulty was no longer simply producing records. It was finding the relevant relationship inside a growing mass of records.
Libraries and indexes usually forced information into categories chosen in advance. Human memory worked differently. One thought called up another. A fact reminded us of an experience, which led to a person, which led to another document. Bush wanted a machine that supported that associative movement.
The memex was a mechanical vision, not a hidden laptop
Modern retellings sometimes turn the memex into a prototype computer. That goes too far.
Bush imagined a desk-like device using microfilm storage, viewing screens, controls and photography. The user could keep books, records and communications inside it, add notes, and rapidly retrieve material. The mechanism belonged to the technology of the 1940s.
The unusual part was how the user would organize the material. Instead of relying only on a fixed subject index, the person could link one item to another and create a trail through several documents. That trail could be saved, revisited and potentially shared.
What mattered was not the storage medium. It was the idea that a reader could build a path through information.
Links changed the unit of organization
A traditional index answers a question like: where is the document filed? Bush’s trail asks a different question: what should I see next?
That shift is subtle. It moves organization from containers to relationships.
Later hypertext researchers, including figures such as Douglas Engelbart and Ted Nelson, developed far more concrete digital systems. Tim Berners-Lee would eventually create the World Wide Web with its own architecture, protocols and goals. Bush did not design those systems, and it would be inaccurate to draw a straight engineering line from the memex to the Web.
But MIT histories and later accounts of hypertext repeatedly point to Bush’s essay as an important intellectual precursor. He gave technical culture a memorable picture of information connected by user-created paths.
He was imagining an interface for thinking
The memex was personal. Bush did not describe a global social network. He described a machine close to one person’s work: their books, records, notes and trails.
That makes the idea feel less like a prediction of the internet and more like an early attempt to design an external memory system.
The user would not merely retrieve a document and stop. They would annotate it, connect it to another record, preserve the connection and continue building a network of context. In modern terms, the value came from metadata created through use.
This is why the idea still resonates with note-taking tools, knowledge bases and research software. The hardware is obsolete; the interface question is not.
The same year, Bush was also redesigning science policy
1945 was important for another Bush document. At President Franklin Roosevelt’s request, he prepared “Science — The Endless Frontier,” a report arguing for continuing federal support of basic research after the war. The National Science Foundation later emerged from a broader political process shaped in part by that debate.
The two documents addressed different systems, but they shared a concern: knowledge is not useful merely because it exists. Institutions and tools determine whether people can find it, extend it and connect it to new work.
One system concerned national research funding. The other concerned an individual’s relationship with information.
The durable lesson is not prediction but structure
It is tempting to celebrate old visions only when they seem to predict modern products. That can distort what made them valuable.
Bush did not foresee today’s Web in technical detail. He did something more useful: he identified a structural mismatch between human thinking and rigid information systems, then proposed a mechanism that made relationships first-class objects.
That is still a productive design question. When a tool helps people collect more information, does it also help them recover the path that made the information meaningful?
Storage keeps facts. A trail keeps context.
In 1945, Bush imagined that difference using microfilm, levers and screens. The materials aged. The problem did not.
Sources
- The Atlantic — Vannevar Bush, “As We May Think”
- MIT Lemelson — Vannevar Bush and the memex
- MIT CSAIL — computing history and the memex
- Library of Congress — Vannevar Bush papers
- National Science Foundation — history and “Science — The Endless Frontier”
- Featured image — Vannevar Bush, U.S. Office for Emergency Management, via Wikimedia Commons. U.S. federal public domain.
SIGNAL TITANS
Leave a Reply