Vannevar Bush Dreamed the Internet in 1945. What Founders Should Actually Take From It.

Vannevar Bush did not predict the internet the way a fortune-teller predicts a lottery number. In July 1945, writing in The Atlantic Monthly, he described a specific problem — the collapse of human judgment under the weight of too much recorded knowledge — and proposed a machine to solve it. The machine was never built. The problem has never been worse. For founders and operators navigating a world of infinite dashboards, real-time feeds, and AI-generated summaries, Bush’s eighty-year-old argument is the sharpest strategic memo you have not read.

Who was Vannevar Bush, and why does 1945 matter to a founder in 2025?

“As We May Think” is a seminal 1945 essay by Vannevar Bush, an American engineer who directed the U.S. Office of Scientific Research and Development during World War II, outlining a vision for tools to extend human memory and manage the growing volume of recorded knowledge through associative indexing rather than rigid hierarchies.¹ Bush was not a theorist operating at the margins. As chairman of the National Defense Research Committee and later director of OSRD, Bush coordinated the activities of some six thousand leading American scientists in the application of science to warfare, and was in effect the first presidential science advisor.² He had spent the war years inside the most consequential decision-making apparatus in modern history. He understood, viscerally, what it meant to act under pressure with incomplete and overwhelming information.

The essay was first published in The Atlantic in July 1945, before and after the atomic bombings of Hiroshima and Nagasaki. Bush expressed his concern for the direction of scientific efforts toward destruction rather than understanding, and articulated a desire for a sort of collective memory machine with his concept of the memex, believing it would help fix these problems.³ The timing is not incidental. A man who had just helped build the atomic bomb was asking, in public, whether humanity had the cognitive infrastructure to handle what it now knew. That is a founder’s question. It is the question every operator faces when their company’s data complexity outpaces their team’s ability to act on it.

What was the memex, and what did Bush actually argue?

At its core is the Memex — a hypothetical desk-sized device equipped with microfilm storage, rapid selectors, and user-created “trails” of linked information — to enable nonlinear access akin to the brain’s associative recall, an idea that prefigured hypertext systems and digital libraries.⁴ The hardware description — microfilm reels, electromechanical controls, a desk-sized cabinet — is the part that ages poorly. The conceptual architecture is the part that ages not at all.

Bush’s central claim was about indexing logic. When data of any sort are placed in storage, they are filed alphabetically or numerically, and information is found (when it is) by tracing it down from subclass to subclass — a method Bush considered artificial and inadequate.⁵ The human mind, he argued, does not work that way. It works by association: one concept triggers another, a trail forms, context accumulates. As conceived, the Memex was capable of making permanent associative links in information.⁶ Those permanent, user-defined trails were the invention. Not the microfilm. Not the desk.

Bush’s vision of associative trails — paths linking related information items through user-defined associations — foreshadows the hyperlink structure underpinning the World Wide Web, where documents interconnect via clickable links to facilitate non-sequential exploration.⁷ But the hyperlink, as it was eventually implemented, solved only half of Bush’s problem. It made information traversable. It did not make it navigable for decision-making. The distinction matters enormously for founders.

The intellectual genealogy: how Bush’s idea became the internet

The chain of influence from Bush’s essay to the modern internet is unusually direct and unusually well-documented. Doug Engelbart, who later became a pioneer in the development of interactive computing and invented the mouse, got part of his inspiration by reading Bush’s article while waiting for a ship home from the Philippines in 1945.⁸ That encounter was not casual. Keeping the memex in mind, Engelbart began work that would eventually result in the invention of the mouse, the word processor, the hyperlink, and concepts of new media for which these groundbreaking inventions were merely enabling technologies.⁹

On December 9, 1968, Engelbart demonstrated the result. The 90-minute live demonstration featured the introduction of a complete computer hardware and software system called the oN-Line System or NLS, which demonstrated for the first time many of the fundamental elements of modern personal computing, including windows, hypertext, graphics, efficient navigation and command input, video conferencing, the computer mouse, word processing, dynamic file linking, revision control, and a collaborative real-time editor.¹⁰ The audience gave him a standing ovation. The computing establishment largely moved on and built something different.

Meanwhile, J.C.R. Licklider formulated the earliest ideas of a global computer network in August 1962 at BBN, in a series of memos discussing the “Intergalactic Computer Network” concept.¹¹ As Licklider wrote: “Men will set the goals, formulate the hypotheses, determine the criteria, and perform the evaluations. Computing machines will do the routinizable work that must be done to prepare the way for insights and decisions in technical and scientific thinking.”¹² This is Bush’s argument in a different register: the computer is a decision-support instrument, not a content-delivery system. Building on the ideas of J.C.R. Licklider, Bob Taylor initiated the ARPANET project in 1966 to enable resource sharing between remote computers.¹³

Ted Nelson, who coined the term “hypertext” in 1967, describes Bush’s article as describing the principles of it.¹⁴ And when Tim Berners-Lee sat at CERN in 1989, he discussed the problems of information loss about complex evolving systems and suggested a solution based on a distributed hypertext system, which eventually became the World Wide Web.¹⁵ Berners-Lee referenced Bush’s associative trails as one of the early inspirations for linked documents on the WWW.¹⁶

Every major architect of the internet’s intellectual foundation — Engelbart, Licklider, Nelson, Berners-Lee — was working, consciously or not, inside the problem Bush defined in 1945. What they built was magnificent. What they did not fully build was the associative trail system Bush considered the point of the whole exercise.

What the internet became versus what Bush intended

The web that emerged in the 1990s solved the distribution problem with extraordinary elegance. Any document could link to any other document. Any person with a browser could reach any node. The information explosion Bush worried about became, if anything, more explosive — and the retrieval mechanisms remained largely hierarchical and keyword-based, precisely the indexing logic Bush had called inadequate in 1945.

Modern web browsers and search engines operationalize similar selection processes, allowing users to follow contextual paths through indexed content. In personal knowledge management systems, tools like Roam Research and Obsidian revive Bush’s associative principles through bi-directional linking, where notes automatically connect based on shared terms or explicit references, forming dynamic graphs of user-generated trails.¹⁷ These tools have a devoted following among knowledge workers. They remain niche. The dominant paradigm — the feed, the notification, the dashboard — is the opposite of what Bush proposed.

The feed is a reverse memex. It delivers information in chronological or algorithmic order, without regard for the associative context of the person receiving it. It is, in Bush’s terms, an artificial index imposed on a mind that works by association. Many studies show that when people are given more information than they can handle, they suffer from negative effects such as confusion, poor judgment, and bad decisions.¹⁸ Information overload can occupy numerous cognitive resources and damage decision quality.¹⁹ Bush diagnosed this in 1945. The research literature has been confirming it for decades. Founders are living it every morning when they open their laptops.

The founder’s translation: decisions, not feeds

The practical implication of Bush’s argument is not that founders should use different software. It is that they should think differently about the relationship between information and action. Bush was not building a library. He was building a decision instrument. The memex was designed to serve a researcher who needed to move from evidence to judgment — quickly, reliably, without losing the thread of prior reasoning.

Consider what that means operationally. A founder who has built a memex-style information environment has done three things. First, they have identified the specific questions their business must answer repeatedly — pricing, hiring, market positioning, product sequencing — and built trails of evidence that bear on those questions. Second, they have linked new information to existing trails rather than consuming it in isolation. Third, they have made their trails shareable, so that operators across the business can navigate the same associative structure rather than each building their own from scratch.

Most founders have done none of these things. They have built notification systems and called them information systems. The distinction is the difference between a memex and a feed. One is designed to support judgment. The other is designed to deliver content. Bush understood the difference. The industry he inspired largely forgot it.

This is not a technology problem. In Science, The Endless Frontier, his 1945 report to the president of the United States, Bush called for an expansion of government support for science and pressed for the creation of the National Science Foundation.²⁰ Bush was simultaneously writing the memex essay and designing the institutional architecture of American postwar science. He understood that the problem of knowledge management was not solved by building better filing cabinets — institutional or mechanical. It was solved by changing the logic of how knowledge was organized relative to the decisions it was meant to support.

What Bush’s second contribution tells us about institutional design

Bush’s parallel output in 1945 — the memex essay and the Science, The Endless Frontier report — reveals something important about his thinking that is often missed when he is discussed purely as a technology visionary. In Science, The Endless Frontier, Bush maintained that basic research was “the pacemaker of technological progress.”²¹ The institutional implication was that you could not optimize your way to breakthrough insight. You had to create the conditions — funding structures, institutional relationships, time horizons — within which insight could emerge associatively, the way the memex was supposed to work.

For founders, this is a governance argument as much as a knowledge-management argument. The companies that make the best decisions at scale are not the ones with the most data. They are the ones that have built institutional structures — meeting cadences, decision frameworks, knowledge-sharing norms — that allow associative reasoning to happen at the organizational level. Vannevar Bush established the U.S. military-university research partnership that later invented the ARPANET, and wrote the first visionary description of the potential use for information technology, inspiring many of the Internet’s creators.²² The partnership was the product. The ARPANET was the downstream consequence.

Three things founders can do this week

Bush’s argument translates into three concrete practices that require no new software and no budget.

Build trails, not folders. Every significant decision your company makes should leave an associative trail: the question, the evidence considered, the reasoning applied, the outcome, and the link to subsequent decisions that depended on it. This is not documentation for compliance. It is a memex for your organization. When the same question recurs — and it will — your team navigates the trail rather than starting from scratch.

Distinguish signal from feed. Audit every information source your leadership team consumes in a given week. Classify each as either decision-relevant (it bears on a specific question your business must answer) or feed (it is interesting but untethered to a specific decision). Reduce the feed aggressively. This is not anti-curiosity; it is pro-judgment. Bush’s concern was not that scientists were reading too much. It was that the structure of what they read was not organized to support the decisions they needed to make.

Make your trails shareable. The memex Bush described was personal — “a device in which an individual stores all his books, records, and communications.” But the most powerful application of associative trail-building is organizational. When your head of product and your head of growth are navigating the same trail of evidence about customer behavior, they make better joint decisions. When they are each consuming their own separate feeds, they make decisions that are locally coherent and organizationally incoherent.

What this means

The forward view: Bush’s problem is getting harder, not easier

The volume of information available to any founder today would have been unimaginable to Bush in 1945. The structural problem he identified — that human cognitive architecture is associative while most information systems are hierarchical or chronological — has not been solved by the internet. In some respects, the internet has made it worse by making the feed the default mode of information consumption for billions of people.

The tools that most closely approximate Bush’s memex — bi-directional linking tools, personal knowledge graphs, AI systems that can traverse a founder’s prior reasoning and surface relevant connections — are becoming more capable rapidly. But the tools are not the constraint. The constraint is the mental model. Founders who understand that their job is to build decision instruments, not consume information feeds, will use those tools well. Founders who do not will use them to build faster feeds.

Bush wrote “As We May Think” at the moment of maximum information anxiety in modern history — the end of a war that had generated more scientific knowledge, more quickly, than any prior period. He did not propose that scientists read less. He proposed that they navigate better. Eighty years later, that is still the right answer. The question is whether you are building a memex or a feed. Business Growth Accelerator (a FounderWise brand) works with founders on exactly this kind of structural clarity — the decision architecture that sits beneath strategy, not just the strategy itself.