A neat myth says the web was invented as a benevolent commons and then “commercialised” by accident. The reality is harsher and more interesting: Tim Berners-Lee built an open architecture, Marc Andreessen industrialised it, and the web’s openness made it capturable. Netscape’s dominance, the server wars, and the rise of platforms show how commons become power. For those nostalgic for a “purer” web, this essay argues that openness was never innocence, and that the commons was always destined to collide with capture. Today, the public web is a shop window; real life moved indoors. What comes next is worse: AI-mediated “engagement” with humans recruited as emotional middleware.

Contents

1. Introduction

There’s a version of early web history that gets told in a tone usually reserved for national founding myths.

It goes something like this: a polite British physicist invents the World Wide Web at CERN to help scientists share documents; he gives it away to humanity; the world nods gratefully; and somewhere off to the side a few excitable Americans in trainers and questionable haircuts “commercialise” it, which we are encouraged to treat as either inevitable progress or original sin depending on our politics.

It’s not wrong. It’s just thin.

Because the web did not become the web because of one man’s benevolence, nor because of one company’s aggression. It became what it became because two very different impulses collided at exactly the right moment: the impulse to build a universal commons, and the impulse to capture and scale whatever commons appears.

Tim Berners-Lee represents the first impulse.

Marc Andreessen represents the second.

And the web, from the beginning, has been the tension between them.

Not because they were locked in some operatic personal rivalry, they weren’t, but because they embodied two incentives that were always going to collide once the web escaped the lab.

Not good versus evil. Not purity versus corruption. Commons versus capture. Both necessary. Both dangerous.

2. Tim Berners-Lee: The Civil Servant Who Accidentally Changed the World

If you strip away the TED aura and the knighthood and the polite British understatement, what Berners-Lee actually did between 1989 and 1991 was astonishingly pragmatic.

He stitched together:

• URLs, so things could be addressed.
• HTTP, so they could be requested.
• HTML, so they could be described.
• A browser/server model simple enough that normal humans could participate.

None of this was conceptually baroque. That’s the point. It was deliberately boring. It was designed to work, to interoperate, to survive institutional inertia. And crucially, it was released openly, without licensing friction, without toll booths, without someone saying “you can build here, but only if you pay us first.”

That wasn’t an accident of history. It was a values decision.

It was values, yes… but also pragmatism. CERN had little appetite for intellectual property trench warfare, and openness was the fastest way to gain collaborators rather than litigants.

Berners-Lee’s web is structurally European in temperament: standards-driven, governance-minded, suspicious of proprietary land grabs, quietly convinced that infrastructure should not be an ideological battleground. It assumes that if you build neutral plumbing well enough, the world will do interesting things with it.

Which is exactly what happened.

But neutrality at the protocol layer does not prevent power at the product layer. It merely refuses to choose sides.

And that refusal creates space.

3. Marc Andreessen: Velocity, Adoption, and the Art of Making It Inevitable

Enter Andreessen.

In 1993, Mosaic did something deceptively simple: it made the web legible. Inline images. A graphical interface that didn’t feel like a terminal window cosplaying as the future. A sense that this was not just a document retrieval experiment but a medium.

People remember Mosaic as quaint now. They forget how intoxicating it felt.

If you wanted to see what the web could become, not as an academic curiosity, but as a living, visual, public thing, Mosaic was the moment it stopped feeling provisional.

Then came Netscape, and with it the acceleration.

The web went from “promising information system” to “global land grab with banners” in record time. The IPO in 1995 wasn’t just a financial event; it was a signal flare. The web had moved from research infrastructure to economic frontier.

Andreessen didn’t ruin the web. He completed it.

Berners-Lee built the possibility of the web. Andreessen made it unavoidable.

Without that velocity, without the slightly manic, adoption-at-all-costs mentality, the web might have remained a beautiful, open, ethically immaculate system used mostly by researchers, librarians, and the kind of people who alphabetise their cables.

It would have been purer.

It might also have been marginal.

4. Mosaic, Richness, and the Feeling of Momentum

It’s hard to overstate how much of early web history is psychological.

Mosaic wasn’t just about inline images. It was about momentum. It gave people the feeling that this thing was going somewhere, that it wasn’t FTP with ambition, that it was not a sidecar bolted onto the internet but the beginning of something layered on top of it.

That emotional shift is what allowed commercialisation to follow without feeling absurd.

Once the web felt like a medium rather than a tool, the capture impulse was inevitable.

5. The Yin-Yang: Standards Versus Reality

The mistake most histories make is reducing Berners-Lee and Andreessen to personalities.

They are better understood as structural forces.

On one side:

The Web as a Public Good

• decentralised
• standards-based
• interoperable
• owned by nobody
• controlled by nobody

On the other:

The Web as a Platform for Power

• proprietary extensions
• distribution control
• branding
• monetisation
• lock-in
• winner-take-most dynamics

Berners-Lee institutionalised the first impulse through the W3C and standards work. Andreessen operationalised the second through product, distribution, and speed.

Neither impulse is stable on its own.

Pure standards without adoption are sterile. Pure capture without standards collapses into fragmentation and technical chaos.

The web’s early history is the argument between those two logics playing out in real time.

6. Netscape: The First Corruption (and the First Explosion)

Netscape democratised the web in a way that’s easy to forget. It made browsing accessible. It made publishing feel mainstream. It made the web feel like a public phenomenon rather than an academic one.

It also began the tradition of pushing beyond standards.

Not because its engineers woke up planning to vandalise neutrality, but because the incentives were obvious: if your browser defines the de facto behaviour of the web, then you don’t just have market share, you have reality share.

The browser wars weren’t really about pixels. They were about authority.

And the first cracks in the web’s neutrality appeared not in the protocols, but in the competitive extensions layered on top of them.

7. Harrods, 1997: The Web Stops Being a Toy

Let’s ground this.

In 1997–1998, I was building Harrods Online, then the UK’s first luxury retail web shop with international shipping. This wasn’t a hobbyist page. This was real payment processing, global logistics, brand risk, frustratingly complex US tax calculations, and executives who absolutely did not want to hear that “the site is down” because some experiment misfired.

By that point, if you were building serious corporate infrastructure, you were not philosophising about openness. You were deploying Netscape Enterprise Server.

In the mid-to-late 1990s, if you were serious and commercial, that was the gravitational centre, before Apache’s open-source momentum and Microsoft’s bundling strategy began shifting the balance.

And not casually.

The mythology now centres on browsers. On Netscape versus Internet Explorer. But in corporate environments, the real story was server dominance. Netscape wasn’t just a browser company; it was the stack. The ecosystem. The assumed default in serious deployments.

On the ground, the web had already industrialised.

8. Netscape as an Ecosystem, Not Just a Logo

What often gets missed is that Netscape wasn’t simply shipping a browser and a server; it was shipping integration patterns, performance assumptions, and early moves toward what we’d now call application infrastructure.

Caching layers. State management. Database embeddings like Sleepycat (Berkeley DB) appearing in stacks. Tooling that treated the web not as a collection of static files but as an application platform, before we had clean language for that transition.

This was pre-cloud, pre-Kubernetes, pre-“DevOps” as a self-conscious discipline.

But the drift was already obvious: the web was becoming stateful, dynamic, performance-sensitive, and economically critical.

In other words, it was becoming serious.

9. The Server Wars: The Part Everyone Forgets

We romanticise the browser wars because they were visible to users.

The real power, as usual, was in the server room.

Between roughly 1996 and 2001:

• Netscape Enterprise Server dominated corporate deployments.
• Apache rose through open source credibility.
• Microsoft pushed IIS aggressively with Windows Server.
• High-performance commercial servers like Zeus targeted serious traffic.

This is the yin-yang dynamic at infrastructure scale.

Apache embodied the commons impulse: open, decentralised, standards-aligned, not owned by anyone. Netscape embodied the capture impulse: integrated, commercial, fast-moving, ecosystem-driven.

By the early 2000s, Apache overtook Netscape. Not because idealism triumphed, but because openness scales differently and enterprises tire of vendor gravity wells.

Around 2003, from my vantage point at Sun Microsystems, the shift was unmistakable. Infrastructure was maturing. Carrier-grade systems. Serious hardware. The monoculture cracked and the ecosystem fragmented into Apache/Linux stacks, IIS/Windows stacks, high-performance niches, early enterprise Java architectures.

Fragmentation kept things honest.

Until centralisation returned at another layer.

Which it always does.

10. Berners-Lee’s Tragedy: Winning the Architecture, Losing the Outcome

Berners-Lee “won” in the historical sense. The web is built on his architecture. The standards persist. The protocols endure.

But look at his later work and you see a different tone: warnings about centralisation, surveillance capitalism, data ownership, identity, decentralisation. A sustained attempt to claw back the commons impulse from platform gravity.

He built a library.

We built casinos on top of it.

The tragedy is not that he failed. It’s that open architecture does not dictate open outcomes.

11. Andreessen’s Tragedy: Becoming the System

Andreessen’s trajectory is the mirror image.

From Mosaic to Netscape to venture capital to techno-optimist manifestos, the capture impulse matured into ideology. Markets as truth engines. Disruption as virtue. Monetisation as validation.

Which is fascinating, because early Netscape was not cynically extractive. It was ambitious. It believed technology should win.

But once you accept that velocity and capture are the primary metrics, you inevitably build systems that optimise for scale over subtlety, growth over governance.

The capture impulse doesn’t stop at adoption. It institutionalises itself.

12. The Web Needed Both… and That’s the Uncomfortable Part

If you want the web to exist at a planetary scale, you need Berners-Lee’s standards and Andreessen’s aggression.

Without the commons impulse, you get fragmentation and proprietary silos. Without the capture impulse, you get a beautifully specified system that never leaves the lab.

The uncomfortable truth is that the web’s greatest strength, its openness, is also what made it capturable.

Open protocols don’t enforce decentralisation. They merely allow it.

The moment serious money arrives, centralisation pressure follows. Not because someone stole the web, but because incentives reward aggregation.

The web wasn’t corrupted in a single act.

It was incentivised.

And those incentives don’t run in one direction forever. The same openness that enables capture also enables revolt, Apache against Netscape, open source against proprietary stacks, but the pendulum never stops swinging.

13. And The Tension Persists

Today the names are different, but the dynamic is identical.

Open source versus SaaS monopolies. Decentralised social versus algorithmic feeds. Protocol innovation versus platform enclosure. Public good versus distribution control.

The web still can’t decide whether it wants to be a universal commons or a behavioural manipulation engine with cat memes and ad slots.

It is both. It has always been both.

And let’s stop pretending this was a passive drift. Facebook turned the open web into a distribution dependency, then throttled outbound links to keep attention inside its walls. Twitter, now X, managed to convert the public square into a bot-amplified outrage machine where signal competes with monetised noise. Both called it “community.” Both engineered enclosure. The web wasn’t merely evolving. It was being strip-mined.

And here’s the bit people still don’t want to say out loud.

The open web didn’t just get “platformised.” It got captured, and not only by the obvious walled gardens like Google, Facebook, Amazon, Apple, even MySpace, but also by Friends Reunited and AOL Online. It also got captured by something weirder: compartmentalism.

Because once the public web became an SEO swamp and a surveillance theme park, the interesting parts of human conversation didn’t migrate to “better websites.” They migrated to places that were explicitly not websites at all: Discord servers, Slack workspaces, private WhatsApp groups, Telegram channels, closed Substacks, invite-only communities, dark-web marketplaces, and endless semi-private chatrooms with their own norms, their own gatekeeping, and their own little economies.

Which means the web’s original promise, one open addressable commons, didn’t die. It fractured. The public square became a billboard, and the real life moved indoors.

In other words: Berners-Lee gave us a world where information could be universally addressable. We responded by rebuilding the medieval city-state, walls, gates, guards, only this time the walls are APIs, the gates are logins, and the guards are moderators and algorithms.

14. Conclusion: The Web Was Never Innocent

The early internet wasn’t a lost Eden. It was an unfinished system.

Berners-Lee gave us neutral architecture. Andreessen gave it momentum. Momentum, left to markets, has direction.

The web became what it became because the commons impulse and the capture impulse collided and co-evolved. Neither won outright. Neither disappeared.

The librarian and the hustler. The protocol and the pitch deck. The cathedral and the food court.

If you want to understand the web, stop asking who ruined it. Start asking which impulse is currently winning.

Berners-Lee built a commons. Andreessen industrialised it. Platforms captured distribution. Surveillance capitalism turned openness into extraction. So people did what people always do when the square fills with hawkers and pickpockets: they moved indoors. Discord. Slack. Telegram. Private groups. Closed loops. The open web is now the shop window. The “real” life is behind the door.

15. Epilogue: What Comes Next Is Even Fucking Worse

And this is exactly why Structuring Cyberpsychology: From Foundations to Practice and Cyberpsychology Today: Signal, Noise, and What We’re Actually Talking About matter. The capture of distribution is no longer just about fugly walled gardens and algorithmic feeds. It is about environments shaping minds at scale. People are not navigating a single public web anymore. They are navigating personalised behavioural architectures, engineered to keep the anxious, reactive and vulnerable dependent. A new world order of personalised hells, populated by bots and by the scared humans those systems condition into prey. If you do not understand the architecture of experience, you are not reading the web. The web is reading you.

And if you think that is bleak, wait. The collapse of the “we’re just platforms” defence does not mean the system gets kinder. It means it gets smarter. Quieter. More intimate. The next phase is not crude feed addiction or obvious bot farms. It is AI-mediated capture that feels like help, with meat puppets recruited to fill the empathy gap. What comes next is even fucking worse.