There is a certain irony in the fact that one of the more quietly compelling entrepreneurial stories circulating in tech right now involves someone deliberately walking away from artificial intelligence. The Verge has reported on Craig Campbell, a former Meta engineer and experienced founder, who turned down the gravitational pull of venture capital flowing into AI startups in order to build what the outlet describes as, simply, a website.
To understand why this choice is notable, it helps to understand just how unusual it is. Campbell is not a marginal figure experimenting at the edges of the industry. He sold a previous venture, an e-commerce tool, in 2022, which places him comfortably in the category of founders who could credibly raise money for almost anything with the word "AI" attached to it. The current funding environment has made that easier, not harder. Venture firms that once demanded rigorous product-market fit are now writing checks to founders with little more than a language model integration and a pitch deck. For someone with Campbell's credentials to look at that landscape and conclude that the better bet was a more traditional web property suggests either unusual conviction or unusual perception, and possibly both.
The broader context here is a quiet but growing counter-movement within parts of the tech world that might loosely be called a rediscovery of the basic web. For roughly a decade, the dominant logic of consumer internet held that platforms were everything and standalone websites were legacy infrastructure. Traffic flowed through social feeds, app stores set the terms of distribution, and the idea of building an audience directly through a URL felt quaint. That logic was reinforced by the economics of venture-backed growth, which rewarded network effects and punitive switching costs over the slower compounding of direct reader or user relationships.
What has begun to shift that calculus, at least for some founders and publishers, is a combination of factors that have eroded confidence in platform dependency. Social media referral traffic has declined substantially across much of the media industry as platforms like Facebook and, more recently, others have deprioritized outbound links in favor of keeping users inside their own ecosystems. The rise of AI-generated content has simultaneously flooded many of those platforms and degraded the signal-to-noise ratio for audiences trying to find reliable, human-produced work. And the AI search transition, still unfolding, threatens to further disintermediate websites from the audiences who might otherwise discover them through conventional search. In that environment, the logic of owning a direct relationship with an audience starts to look less nostalgic and more strategically sound.
Campbell's decision, as The Verge frames it, is a bet on what might be called the durable web, the idea that a well-built destination with a clear identity can outlast the platform cycles that have repeatedly redistributed attention over the past two decades. The likely reading is that he identified a specific audience or subject area where aggregated, curated, or originally produced content has genuine scarcity value, and concluded that the subscription or advertising economics could work without the overhead that a venture-backed growth trajectory demands. A bootstrapped or lightly funded web property, if it finds its footing, carries none of the existential pressure to achieve a return multiple that forces companies to scale into markets they are not ready for.
The consequences of this, if Campbell's bet is paying off as The Verge suggests, are worth thinking through in a few directions. For the startup community, it offers a small but visible data point that contrarian positioning against the AI funding wave is not inherently irrational. For the broader media and publishing ecosystem, it reinforces a case that a generation of digital publishers, many of whom collapsed or sold under duress in the late 2010s, perhaps gave up on direct audience ownership too soon. And for the venture industry itself, it raises an uncomfortable question about how many fundable AI companies will, in three years, look like solutions to problems that did not need solving, compared with the more modest and sustainable businesses that founders quietly built while everyone else was chasing the same trend.
What to watch for next is straightforward in outline if uncertain in timing. The real test for any founder making this kind of bet is whether the economics hold as AI-generated content continues to scale and as AI-assisted search changes how audiences discover destinations on the web. Campbell's background in engineering and his previous experience building an e-commerce product suggest he is approaching this as a systems problem, not a romantic one. Whether the audience he is building finds him directly, or whether changing discovery patterns eventually close that door, will determine whether his early results become a durable business or a well-timed but temporary window. The answer to that question will matter well beyond his particular venture.
