Craig Campbell didn’t just step away from the AI rush—he stepped toward something that, at least on paper, looked almost unfashionable. In a moment when founders were being pulled into building models, copilots, and “AI-native” everything, Campbell chose to build a website business. Not a content farm. Not a tool that promises to “use AI to generate pages.” A real web product, built around the idea that the internet still runs on discoverability, structure, and trust—things that don’t disappear just because search is changing.
The story matters now because the ground under online discovery is shifting quickly. For years, the web’s default assumption was simple: you type a query into Google, you get ten blue links, and you click your way to answers. But the last couple of years have introduced a new pattern—one where search results increasingly behave like an interface rather than a directory. Instead of sending users outward, search systems are trying to keep them inside the experience. That shift has created a new anxiety among publishers and startups alike: if users aren’t clicking through, what happens to the businesses that depend on traffic?
Campbell’s bet is essentially that the answer isn’t “stop building websites.” It’s “build websites that can survive the new rules.”
To understand why that’s a meaningful stance, it helps to look at Campbell’s background and the timing of his decision. He’s a former engineer at Meta and an experienced founder. In 2022, he sold his previous venture—an e-commerce tool for businesses using Shopify. That exit came right as the AI boom was accelerating, and Campbell says he felt the pressure from investors to pivot. The pitch was familiar: start something else, move into AI, take advantage of the capital and attention flowing into the space. He describes it as investors breathing down his neck, offering blank-check energy.
But Campbell didn’t treat that pressure as a command. He treated it as noise. The more interesting question is why he believed a website company could still win when so many people were convinced the future belonged to AI-first products.
His reasoning, at least as it’s reflected in how he talks about the project, points to a deeper belief about how the web works. The “old school web” isn’t just nostalgia. It’s a set of fundamentals: clear information architecture, pages that earn their place through usefulness, and distribution that doesn’t rely entirely on one channel. When those fundamentals are strong, the site becomes resilient—even if the path users take to reach it changes.
That resilience is especially relevant as Google’s search behavior evolves. The industry has been watching the “zero-click” trend for a while, but the more recent changes feel less like a gradual shift and more like a redefinition of what search is for. If search increasingly returns synthesized answers, then the traditional SEO playbook—optimize for ranking, then monetize clicks—starts to wobble. Some companies respond by chasing AI-generated content or building tools that try to “outsmart” the new interface. Others respond by doubling down on the idea that the web is still a place where people browse, compare, and verify.
Campbell’s approach leans toward the second camp. He’s not ignoring AI; he’s refusing to let AI become an excuse to abandon the web’s core strengths. In practice, that means building a site that behaves like a reference, not like a stream. It’s the difference between a page designed to rank and a page designed to be used.
There’s also a psychological component to this decision. When the market is loud, founders often feel compelled to match the volume. AI is the loudest thing in tech right now, and it’s easy to interpret that loudness as a signal that everything else is doomed. Campbell’s move suggests a different interpretation: AI is a powerful capability, but it doesn’t replace the need for information that’s organized, accessible, and maintained. Models can summarize, but they don’t automatically create the kind of structured knowledge that a well-built website can offer over time.
That’s where the “old school web” framing becomes more than branding. It’s a strategy for building trust and utility in a world where attention is fragmented and discovery is less predictable.
What makes the story more compelling is that Campbell’s choice wasn’t made in a vacuum. The web’s economics have been under strain for years. Advertising models have shifted. Social platforms have changed how content is surfaced. Search has become more competitive and more opaque. And now, with AI-driven interfaces, the fear is that the web’s traditional distribution engine is becoming less direct.
In that environment, a website business has to do more than exist. It has to earn its visibility. It has to be legible to search systems and valuable to humans. It has to be the kind of resource that people bookmark, share, cite, and return to—because those behaviors reduce dependence on any single moment in the search funnel.
Campbell’s bet appears to be that the “old school” approach—pages that are genuinely useful, built with care, and structured for clarity—still creates compounding returns. Even if fewer people arrive via classic click-through, the site can still become a destination. And destinations matter because they create their own distribution: links, mentions, repeat visits, and word-of-mouth among the communities that care about the topic.
This is where the story takes on a unique angle. Many discussions about “the old web” focus on aesthetics or nostalgia: static pages, simpler design, fewer scripts. Campbell’s version is more functional. It’s about the mechanics of how information travels. In a world where AI can answer questions, the sites that remain valuable are often the ones that provide context, depth, and verification—things that are hard to synthesize without access to reliable sources and without the ability to maintain accuracy over time.
A good website can do that work continuously. It can update. It can correct. It can document decisions. It can show evidence. It can link out to primary material. It can build a body of knowledge rather than a one-time answer. Those are advantages that don’t vanish when search changes; they just become harder to ignore.
There’s also a subtle but important point about how founders think about risk. AI startups often carry a different kind of uncertainty: performance depends on model quality, data access, and the pace of iteration. Website businesses carry uncertainty too, but it’s more operational. Can you build something that stays useful? Can you keep it updated? Can you attract the right audience? Can you make it discoverable in a shifting environment?
Campbell seems to have accepted that operational risk rather than the model risk. That’s not a small choice. It implies confidence that the team can execute on fundamentals and that the market will still reward clarity and usefulness.
And execution is where the “paying off” part comes in. The Verge piece frames Campbell’s progress as evidence that the bet is working. While the excerpt doesn’t list every metric, the underlying message is clear: the website business is gaining traction, and Campbell is leaning into the old-school approach rather than abandoning it for the next AI wave.
That decision—staying with the fundamentals when the market is offering easier narratives—is often what separates durable companies from short-lived ones. AI trends can create a sense of urgency that pushes founders into building features instead of building value. Websites, by contrast, force you to confront the question: what does the user actually need, and how will they find it?
If you’re building a site that functions like a reference, you’re implicitly designing for long-term use. That changes how you measure success. You’re not only tracking immediate traffic spikes; you’re tracking whether the site becomes a resource people rely on. Over time, that can translate into stronger organic visibility, more inbound links, and better performance even as search interfaces evolve.
Another reason this story resonates right now is that the “Google Zero event horizon” isn’t just a phrase—it’s a real strategic challenge. When users get answers directly in search, the incentive to click decreases. But the incentive to verify, explore, and go deeper doesn’t disappear. People still want sources. They still want comparisons. They still want the “how” behind the “what.” They still want to understand tradeoffs.
Websites that provide those layers can still win, even if the click-through rate changes. The trick is to design for the moments when users need more than a summary. That might mean building pages that are structured for scanning, that include clear sections, that provide actionable steps, that cite evidence, and that connect related topics in a way that feels like browsing a library rather than scrolling a feed.
Campbell’s “old school web” stance suggests he’s building for those moments.
There’s also a distribution lesson here that founders sometimes learn the hard way: relying on a single channel is fragile. Search is one channel. Social is another. Direct traffic is another. Partnerships are another. Email newsletters are another. Communities are another. When search changes, the companies that survive are often the ones that have multiple ways to reach users and multiple reasons users return.
A website business can build those multiple paths if it’s intentional. It can create content that earns links. It can build tools that people share. It can develop a brand that shows up in conversations. It can create internal navigation that encourages exploration. It can also build a product layer—something beyond static pages—that gives users a reason to come back.
The Verge excerpt hints at this broader philosophy: Campbell walked away from investor pressure to chase AI and instead focused on building a website that can stand on fundamentals. That’s not just a technical decision; it’s a go-to-market decision. It’s choosing a path where the product itself is the distribution engine.
It’s worth noting that “old school web” doesn’t mean “no innovation.” The modern web is still full of sophisticated engineering. The difference is where the sophistication goes. Instead of using AI as a shortcut to generate content at scale, the sophistication goes into making the site genuinely useful and maintainable. It goes into