Blog · Review Essay · Last reviewed June 25, 2026

The Internet Revolution and the Ideology Inside the Machine

Richard Barbrook and Andy Cameron's The Internet Revolution: From Dot-com Capitalism to Cybernetic Communism is useful because it treats network technology as a political argument disguised as infrastructure. Its two 1990s essays ask why digital culture kept presenting private power, market fatalism, frontier myth, countercultural style, and technical inevitability as if they were the natural meaning of the Net.

The reviewable idea is the Californian ideology: a story in which anti-bureaucratic freedom, entrepreneurial heroism, technical inevitability, and private infrastructure reinforce each other until market design looks like nature. Read in the AI era, the question is not whether the story is attractive. It is what ownership, labor, data, standards, safety, and exit conditions the story makes hard to see.

The Book

The Internet Revolution: From Dot-com Capitalism to Cybernetic Communism was published by the Institute of Network Cultures in Amsterdam in October 2015 as Network Notebook #10. The HvA Research Database lists Richard Barbrook and Andy Cameron as authors, identifies it as a professional book, gives it 51 pages, records print ISBN 9789492302014 and electronic ISBN 9789492302021, and places it in the Network Notebook series. Google Books gives the same title, authors, publisher, year, and length.

The volume is a twentieth-anniversary return to two earlier interventions: "The Californian Ideology," originally published in Mute in 1995 and circulated through nettime, and Barbrook's 1999 "Cyber-Communism." The publisher describes the first as a landmark of early Net criticism and frames the second as a counter-prophecy from the dot-com bubble's peak. The book also includes a new introduction looking back on the "hippie capitalists" who helped shape Silicon Valley's self-image.

That structure makes the book more than a period piece. It preserves an argument from the moment when the web was still being narrated into public meaning. The question was not only what the network could technically do. It was which social story would attach to it: private liberation, public utility, marketplace destiny, military infrastructure, worker craft, common knowledge, or some unstable mixture of all of them.

Current Context

As of June 25, 2026, the book's critique has moved from early web culture into the governance of platforms, cloud infrastructure, foundation models, recommender systems, and AI agents. The old promise that the Net would naturally produce freedom now competes with a more inspectable reality: app stores, search engines, social platforms, model APIs, cloud regions, payment rails, training-data pipelines, and enterprise connectors decide who can build, publish, discover, monetize, automate, and leave.

The current regulatory record does not vindicate any simple anti-market or pro-state answer. It does show that "technology wants" is no longer a serious policy category. The European Commission describes the Digital Services Act as a framework for online services, with stronger duties for very large online platforms and search engines above 45 million monthly EU users, including transparency, systemic-risk assessment, independent audit, data access for vetted researchers, recommender options, and public ad repositories. The Digital Markets Act targets gatekeeper power in core platform services such as search engines, app stores, and messaging services. The EU AI Act adds a risk-based AI layer, with staged application dates, governance for general-purpose AI models, and high-risk duties that increasingly turn model deployment into a recordkeeping problem.

The open-versus-closed AI debate also needs the book's contradiction-aware method. NTIA's 2024 report on dual-use foundation models with widely available weights treats openness as both a source of benefits, including broader participation and reduced concentration, and a source of risks when oversight and accountability mechanisms are weak. That is the mature version of Barbrook's point: commons, markets, and public institutions are not pure categories. The question is which feedback loops, ownership rights, safety controls, and remedies survive once the technology becomes infrastructure.

The Ideology

The book's best-known idea is that Silicon Valley's digital politics fused incompatible traditions and made the fusion feel obvious. Countercultural anti-bureaucracy, McLuhanite media mysticism, entrepreneurial individualism, libertarian economics, frontier history, and a distrust of state planning all became part of one confident story: networked computers would dissolve old institutions and set free creative individuals.

That ideology is not simply a set of opinions held in California. It is a legitimation pattern. First, a technical possibility is described as inevitable. Second, the private institution best placed to exploit it is described as the agent of liberation. Third, objections about labor, ownership, public obligations, extraction, surveillance, or inequality are recoded as fear of the future. Fourth, the resulting market structure is treated as proof that history wanted it that way.

Barbrook and Cameron's target is not simply optimism. Their sharper point is that optimism can naturalize power. If the future is technologically determined, then institutional choices disappear. Corporate platforms become evolution. Venture capital becomes liberation's funding mechanism. The entrepreneur becomes a political hero. Public alternatives, labor claims, regulation, and democratic design can be dismissed as backward interference with what the machine already wants.

The Minitel counterexample matters here. Barbrook's introduction contrasts the public-service path of French network access with the more privatized and market-led Anglo-American path. The comparison is not nostalgia for a terminal system. It is a reminder that networks have political economies. Access models, billing systems, ownership, interface design, labor organization, and public obligations are not external to technology. They are part of what the technology becomes.

Cybernetic Communism

The second essay deliberately reverses the dot-com story. Instead of treating the Net as the perfect machine for neoliberal markets, Barbrook points to gift exchange, file sharing, open collaboration, and the weakening of old intellectual-property boundaries. The internet that American capitalism built was also filled with practices that did not behave like ordinary commodities.

That reversal is still useful, especially now that every commons can become training data, every open repository can become product infrastructure, and every user contribution can be reabsorbed by a platform. The book sees the network as contradictory rather than pure. It can enable common production, but it can also centralize extraction. It can support digital artisans, but it can also make their work invisible inside scalable systems.

The word "cybernetic" also matters. The politics of the network is not only ownership. It is feedback: who measures, who adapts, who sees the dashboard, who gets corrected, who becomes data, and who gets to change the rule. The same infrastructure can host mutual aid, surveillance, platform discipline, public knowledge, financial speculation, and automated persuasion depending on which loops are given institutional power.

A useful AI-era commons test asks five questions. Who can contribute? Who captures the value? Who can refuse reuse? Who governs later deployment? Who can see and correct harm after the shared resource is folded into a product? Without those answers, the language of openness can become a supply-chain mask for enclosure.

The AI-Age Reading

Read from the AI era, The Internet Revolution is a book about ideology before deployment. The current AI boom has its own version of the same script: frontier language, founder charisma, civilizational urgency, open-future rhetoric, distrust of democratic delay, and repeated claims that technical progress has already decided the institutional form of tomorrow.

The comparison is concrete. Generative AI firms use public language about creativity, access, research, empowerment, and human flourishing while competing for private control over compute, data, model platforms, payment channels, agent ecosystems, and enterprise workflows. The old web promise that users would become creators now reappears as a promise that everyone will have an assistant, tutor, coder, researcher, therapist, lawyer, or co-worker. In both moments, the emancipatory language can be real and evasive at the same time.

The book also helps explain why AI politics cannot be reduced to "open" versus "closed." Open models, open-source code, public datasets, academic papers, and volunteer labor can all expand agency. They can also become upstream inputs for private enclosure. Closed systems can be abusive, but public infrastructure can fail too if it lacks funding, contestability, labor protections, and democratic accountability. The deeper question is which feedback loops become durable institutions.

AI agents make the ideology more operational. A platform no longer only hosts speech or ranks content; it can mediate tool use, commerce, documents, search, code, workplace records, and public-service forms. The story of empowerment becomes testable at the permission layer: which tools can the agent call, what data can it read, what accounts can it write through, what logs survive, what choices are sponsored or unavailable, and who can reverse an action that looked convenient in the interface?

Governance and Safety

The practical governance lesson is to audit the story before buying the system. When a vendor, platform, or public agency presents a technical product as liberation, modernization, democratization, or inevitability, the review should ask what institutional claim is being made and what evidence would disprove it.

A serious review should leave records, not vibes. For any platform or AI system treated as infrastructure, preserve the system owner, business model, funding or procurement route, data sources, training or reuse terms, labor dependencies, model or ruleset version, user groups, affected rights, safety case, appeal path, incident process, interoperability limits, export path, and shutdown condition. Those facts belong near an AI system inventory, public register, vendor governance file, and platform risk review.

For public institutions, the Minitel lesson is not that every digital service should be state-run. It is that public purpose must be designed into the stack before dependency hardens. Procurement should test whether a service can be audited, ported, explained, exited, and governed by public values. A public alternative, cooperative model, open standard, or publicly governed interface may be the right answer in some domains; in others, the minimum is a contract that prevents lock-in, preserves logs, protects privacy, and gives affected people a usable remedy.

For AI safety, the book pushes beyond model benchmarks. The relevant risk is not only bad output. It is ideological capture: an institution starts treating a product roadmap as political destiny, adopts a vendor's category system as reality, lets open collaboration become uncredited extraction, or delegates decisions to platforms whose incentives cannot be inspected. Safety therefore includes contestability, labor rights, data provenance, privacy, source visibility, incident reporting, exit rights, and the capacity to keep alternatives alive.

NIST's AI Risk Management Framework is useful here because it turns broad claims into governance functions: govern, map, measure, and manage. Read through Barbrook and Cameron, those functions should be aimed at the whole political economy of the system, not only the model artifact. The question is not "is this future exciting?" It is "what control loop are we installing, who can see it, who benefits from it, and who can change it?"

Where the Book Needs Care

The pamphlet form gives the book force, but also limits. It is polemical, compressed, and sometimes too eager to make Silicon Valley's contradictions resolve into one named ideology. Real technical cultures are messier: workers disagree with executives, open-source communities overlap with companies, public institutions can be extractive, and users often adopt tools for practical reasons that do not match the ideology sold around them.

The "cyber-communism" argument also needs updating after cloud platforms, app stores, social networks, content moderation markets, data brokers, creator economies, and foundation models. The internet did not simply abolish scarcity. It moved scarcity into attention, compute, distribution, identity, moderation, payments, trust, and legal control. Common production survived, but so did enclosure at higher layers of the stack.

The public-service comparison also needs friction. Public infrastructure can be slow, exclusionary, surveillant, underfunded, or captured. Markets can sometimes broaden access, and open communities can create real public goods. The book's value is not a complete institutional recipe. Its value is the refusal to let any one recipe pretend to be the automatic will of the network.

Those limits do not weaken the book's value. They show how to use it. Treat it as a diagnostic instrument, not as a finished map. When a technology company presents its business model as the natural path of history, ask which political choices are being hidden inside the story.

What This Changes

The recurring pattern is the conversion of technical possibility into social inevitability. A network appears. A class of interpreters explains what it means. Their explanation becomes product language, investment thesis, policy assumption, workplace demand, and everyday common sense. The interface arrives as a fact, but the fact has already been narrated.

The Internet Revolution is valuable because it catches that narration early. It shows how belief forms around infrastructure: not by argument alone, but through magazines, conferences, demos, entrepreneurs, interfaces, funding, jargon, and the lived pleasure of using powerful new tools. Once the story works, people can mistake a political economy for a technological destiny.

The AI lesson is plain. Do not let models tell society what institutions must become. Do not let assistants make dependency look like empowerment by default. Do not let open collaboration become uncredited extraction. Do not let governance arrive only after the infrastructure has trained everyone to accept its terms. The future is not hidden inside the machine. It is made through the rules, ownership, labor, access, memory, and feedback loops built around it.

Source Discipline

This review separates book facts, regulatory facts, and interpretation. Book metadata comes from the Institute of Network Cultures, the HvA Research Database, the INC PDF, Mute, and Google Books. Current governance context comes from official European Commission, NIST, and NTIA materials. Claims about ideology, platform enclosure, AI agents, commons extraction, and site relevance are interpretive arguments built from those sources, not claims that the book predicted every present system.

Regulatory sources need jurisdiction and date discipline. The DSA, DMA, and AI Act cover different services, actors, thresholds, and timelines. A transparency duty is not the same as a competition remedy; a gatekeeper designation is not the same as a safety finding; an AI Act implementation page is not proof that a deployment is safe. The useful habit is to name the instrument, date, actor, system, evidence record, and remedy path.

This page makes no claim that any AI system is conscious, divine, or AGI. It treats AI systems as infrastructures made of data, models, interfaces, cloud dependencies, permissions, workers, markets, policies, and institutions.

Sources

Book links are paid affiliate links. As an Amazon Associate I earn from qualifying purchases.

Buy on Amazon Browse Books

Return to Blog · Return to Books