The internet was built by the American military to survive a nuclear war. The ARPANET, the network funded by the Defense Department’s Advanced Research Projects Agency that went live in 1969, was designed on a principle of distributed redundancy: instead of routing all communications through a central hub that could be destroyed, it would route packets of data through multiple possible paths, finding its way around damage the way water finds its way around an obstacle. The network could not be decapitated because it had no head.
What it became — within three decades the infrastructure of global commerce, the primary means by which most of the world’s population communicates, and the most powerful surveillance apparatus in human history — was not planned. None of the researchers who built ARPANET imagined they were building the substrate of a new economy. Tim Berners-Lee, the British physicist who invented the World Wide Web in 1989 as a tool for sharing scientific papers among researchers at CERN in Geneva, explicitly designed it as an open, non-proprietary system: the web’s protocols were given freely to the world, and the decision not to patent or monetise them was a choice whose consequences for human civilisation are difficult to overstate, since it meant that the web’s basic infrastructure would be open to anyone. What was built on top of that infrastructure — the commercial layer, the social layer, the surveillance layer — is another story, and it unfolded with a speed and completeness that left every institution designed to govern it years behind the reality it was trying to regulate.
From Academic Network to Mass Medium
The internet of the 1970s and 1980s was a specialist tool: email, file transfer, bulletin board systems, the communities of researchers and enthusiasts who learned to use it and developed its early social norms. The culture of those early networks — collaborative, anti-commercial, oriented toward open sharing of information — left a residue that shaped the rhetoric if not always the reality of the web’s commercial era. The hacker ethic, as the journalist Steven Levy codified it in 1984, held that information wanted to be free, that access to computers should be unlimited and total, and that authority was to be treated with principled suspicion. These values animated the open-source software movement, the Wikipedia project, and the early architecture of the web. They also provided the ideological cover under which a generation of commercial enterprises extracted enormous private value from the open infrastructure the public had subsidised.
The decisive transformation came in 1991, when Berners-Lee released the protocols for the World Wide Web — the system of hyperlinked documents and the HTTP protocol that bound them — as free software. In 1993, Marc Andreessen and Eric Bina at the National Center for Supercomputing Applications released Mosaic, the first graphical web browser, which made the web navigable by anyone with a personal computer rather than by researchers who knew how to use a command line. The number of websites grew from a handful in 1993 to more than 100,000 by 1996 to tens of millions by the end of the decade. The browser wars between Netscape and Microsoft’s Internet Explorer were the first major commercial battle to be fought on the web’s terrain.
The Dot-Com Boom and Its Aftermath
Between 1995 and 2000, the first wave of internet utopianism produced the largest speculative bubble in the history of financial markets to that point. The Nasdaq Composite index rose from approximately 1,000 in 1995 to more than 5,000 at its peak in March 2000, driven by investment in companies whose business models were often no more substantial than a domain name and a PowerPoint presentation. Pets.com, Webvan, Kozmo.com: the companies that became symbols of the boom’s irrationality were real businesses — some of them, like Amazon, survived and thrived — but the valuations placed on them bore no relationship to any plausible near-term earnings. The crash that followed, when it came, wiped out approximately five trillion dollars in market value between March 2000 and October 2002.
What the crash did not do was kill the internet. The underlying infrastructure — the broadband networks, the server farms, the protocols — had been built with real capital and remained in place. The engineers and entrepreneurs who had been drawn to Silicon Valley by the boom remained there, building on the rubble of failed ventures. Google, which had been founded in 1998 on a page-ranking algorithm developed by Larry Page and Sergey Brin as a Stanford research project, survived the crash and emerged from it as the dominant force in internet search. Amazon, which had come close to bankruptcy in 2001, rebuilt itself as not merely a bookseller but a comprehensive retail and logistics platform and, subsequently, a cloud computing infrastructure provider whose AWS division became the technical backbone of a significant proportion of the internet’s commercial layer.
Web 2.0 and the Social Turn
The second phase of internet development — retrospectively labelled Web 2.0, a term whose imprecision did not diminish its usefulness — was distinguished from the first by its orientation toward user-generated content and social interaction. The early web had been primarily a publishing medium: organisations and individuals created content that other people read. Web 2.0 was a participation medium: the content was created by users, and the value of the platform derived from the network of connections between them.
The transformations this produced were rapid and radical. Wikipedia, which launched in 2001, created within a decade the largest encyclopedia in human history, written and maintained entirely by volunteers — a demonstration of what collective intelligence could produce when the right incentives and structures were in place. The blogosphere that emerged in the early 2000s democratised publishing in ways that genuinely expanded the range of voices in public discourse; it also lowered the barriers to spreading misinformation, removed the institutional filters — editors, fact-checkers, legal review — that the previous media system had imposed, and created the infrastructure of the online attention economy that would later be industrialised by social media platforms.
Facebook, founded by Mark Zuckerberg at Harvard in 2004 and opened to the general public in 2006, was the platform that completed the social turn. Its core mechanism — the social graph, the map of connections between users — generated a dataset of human relationships that had never previously existed, and on which it built an advertising business of extraordinary precision: the ability to target messages to individuals based not merely on their demographic characteristics but on their revealed interests, their expressed emotions, their social connections, and their behavioural patterns across the web. Twitter, launched in 2006, created a public conversation infrastructure that was simultaneously more open and more legible than anything that had come before. YouTube, acquired by Google in 2006, made video production and distribution available to anyone with a camera and an internet connection.
The iPhone, released by Apple in June 2007, was the final piece of the architecture: it put the internet in everyone’s pocket, connected people to the network continuously rather than in discrete sessions, and created the mobile attention economy in which the social media platforms would invest most heavily. By 2020, the average American adult spent more than three hours per day looking at their smartphone, a figure that represented a transformation of daily life whose scale had no precedent in peacetime.
Surveillance Capitalism
The economist Shoshana Zuboff, in her 2019 book The Age of Surveillance Capitalism, provided the most comprehensive theoretical account of what the internet economy had become: a system in which human experience is the raw material from which behavioural data is extracted, processed, and sold as predictions of future behaviour to advertisers and others who want to influence human action. The logic of the system, Zuboff argued, was not merely to know what people had done but to predict and modify what they would do — to turn the accumulated data of human life into an instrument of behavioural modification at scale.
The data collection was extraordinary in its comprehensiveness. Google knew what people searched for — which is to say, what they wanted, feared, wondered about, and desired in the most direct and unmediated terms. Facebook knew who people were connected to, what they said to each other, what content made them react with joy or anger. Smartphones tracked people’s physical location continuously. The ecosystem of third-party trackers embedded in almost every website created a system in which every online action was monitored, recorded, and processed by entities the user had no relationship with and could not identify.
The political consequences of this system began to become visible in the mid-2010s. The filter bubble — the tendency of algorithmically curated feeds to show people content that confirmed their existing beliefs rather than challenged them — had been identified by the internet activist Eli Pariser as early as 2011. What the 2016 American presidential election and the Brexit referendum demonstrated was that the social media infrastructure could be weaponised for political purposes at scale: micro-targeted advertising, coordinated disinformation campaigns, foreign state actors operating through fake accounts to amplify division. The Cambridge Analytica scandal, which broke in 2018, revealed that data harvested from Facebook had been used to build psychological profiles of tens of millions of American voters and target them with politically calibrated messages without their knowledge or consent.
The Promise and the Problem
The internet had genuinely transformed human life in ways that contained real advances. The democratisation of information — the ability of anyone with an internet connection to access the accumulated knowledge of human civilisation, to learn anything, to find communities of people who shared their interests or their struggles — was real, and its benefits were unequally distributed but genuinely significant. The Arab Spring of 2010–12 demonstrated that social media could enable political mobilisation under authoritarian regimes in ways that previous communication technologies could not; it also demonstrated that such mobilisation could be reversed by authoritarian governments who learned to use the same tools for surveillance and repression.
The knowledge economy that the internet enabled generated enormous wealth concentrated in a small number of companies and a small number of places. The five largest American technology companies — Apple, Google, Amazon, Microsoft, and Facebook — by 2020 had market capitalisations that exceeded the GDP of most countries. The workers in their Californian headquarters were among the most highly compensated in history; the workers in their warehouses, delivering packages and processing returns, were among the most surveilled and productivity-measured of any workforce since Frederick Winslow Taylor.
The regulatory frameworks designed to govern the internet lagged far behind its development, and not by accident: the technology industry had invested heavily in the political argument that regulation would kill innovation, and governments across the world had largely accepted it. The General Data Protection Regulation, introduced by the European Union in 2018, was the most serious attempt to impose meaningful privacy protections on the data economy; it produced compliance costs and cookie consent popups but did not fundamentally alter the economic logic of surveillance capitalism.
What the Network Did
The internet’s most consequential effects were not the ones most readily visible. The shift of human attention from broadcast media — which delivered the same content to large audiences simultaneously — to personalised feeds organised by algorithmic recommendation engines reorganised both the economics of media and the epistemological commons on which democratic politics depends. When people receive different information, curated to confirm different prior beliefs, organised by platforms whose financial incentives push toward engagement rather than accuracy, the shared factual basis on which political argument is possible erodes. The public sphere that Jürgen Habermas theorised in the 1960s — the sphere of rational-critical discourse through which citizens engage with each other on matters of common concern — had been built on the infrastructure of mass media. The social media era did not destroy that sphere but fragmented it in ways whose consequences for democratic governance are still being worked out.
The network also reorganised economic life in ways whose full implications were unclear by the early 2020s. The platform economy — in which internet platforms mediated between workers and customers at scale, extracting rent from the transaction while classifying the workers as independent contractors to avoid the obligations of employment — represented a genuine restructuring of labour relations. The gig economy that Uber, Deliveroo, and their equivalents created was enormously efficient at matching supply with demand and genuinely useful to its customers; it was also a system in which the productivity gains of digital organisation were captured by platform owners rather than shared with the workers who provided the actual labour.
The internet was the most significant new infrastructure of the late twentieth century, more transformative in its social effects than any technology since the printing press and in some respects more rapidly transformative than that. It democratised knowledge and concentrated wealth. It connected people across every previous barrier of geography and enabled new communities of interest and solidarity. It also created new mechanisms of manipulation, new vectors for the spread of hatred and disinformation, and new forms of surveillance that states and corporations could deploy against the populations they served or governed. These contradictions were not accidents or failures of implementation. They were the structural features of a technology whose possibilities were genuinely open, and whose actual development was shaped — as all technologies are shaped — by the specific economic, political, and social conditions of the world into which it arrived.
Leave a Reply