I didn’t expect a 19-minute video about paper coffee filters to offer the clearest case against the strip-mining era of AI, but here we are.
Yes, I watched it. A deep dive comparing the taste of pour-over filters. For most people, that sounds oddly specific. For people like me who roast their own beans, it’s a normal Saturday morning. But here’s the thing: that video has over 450 thousand views at the time of this writing. Not because it magically converted the world into coffee obsessives, but because it tapped into a massive, passionate community that already cares deeply about this stuff.
The coffee world is filled with people who debate grind size the way others debate baseball stats—and James Hoffmann speaks their language fluently. That’s what makes it powerful: it’s not random virality; it’s the precision of a creator who knows exactly who he’s talking to and why it matters. And that, I think, is the lesson for anyone trying to build something that endures in the age of AI.
Because we’ve entered what Jason Calacanis calls the “strip-mining era of AI.”
Large language models are vacuuming up the open web—every article, post, and how-to guide—distilling them into clean, context-free answers that rarely credit or compensate the people who did the work. Users get instant information. Platforms get engagement. But the small publishers, independent creators, and niche experts who produce the knowledge? They get buried under the very data they created.
Big publishers have armor. The New York Times, News Corp, Axel Springer—these giants can negotiate licensing deals or rely on subscription models. But small creators don’t have that leverage. They survive on reach, community, and the slow compounding of trust. When traffic disappears, so does oxygen.
Which brings us back to Hoffmann’s 19-minute video. It’s a masterclass in the one advantage AI can’t replicate: specificity.
Nobody asked for a video comparing pour-over filters. There was no keyword demand, no trending topic. Yet it exploded. Why? Because it went so deep into a topic that even the most advanced model couldn’t fake it.
That’s the paradox of creativity in the AI age: the only content that thrives is either massive or microscopic. If you’re not Netflix, be James Hoffmann.
At one end of the spectrum are massive entertainment ecosystems with scale as their moat—Disney, Marvel, Taylor Swift. At the other end are creators who turn narrow passions into cult followings. The middle—generic content, recycled commentary, SEO-driven filler—is where AI will feast first.
Specificity is the new scalability.
It’s what makes a creator unreplicable. Not just the topic itself, but the way they approach it: the cadence, curiosity, and conviction that signal real expertise. Hoffmann’s video didn’t just compare filters; it unpacked the science of flow resistance, paper texture, and roast behavior. It felt like watching a scientist in a t-shirt. And that’s precisely why it worked.
In the age of generative content, knowledge isn’t a commodity, it’s a craft. The difference between “good” and “great” used to be efficiency: who could publish more, faster. Now it’s intimacy: who can make the audience feel that what they’re watching couldn’t have been made by anyone else. That’s the moat.
The internet was built on accessibility; AI is forcing a return to authenticity.
It reminds me of the early days of the coffee industry itself. Once upon a time, all coffee was just “coffee.” Then came the wave of specialty roasters who obsessed over origin, altitude, and grind size. They didn’t win because they scaled faster, they won because they cared deeper.
That’s the opportunity for creators now: go deeper, not wider. Be the person who knows something so well that even AI has to cite you. The future belongs to those who make their knowledge a specialty product, not a commodity.
So yes, I’ll keep watching long videos about paper filters. Not only because I’m a coffee snob, but because it’s a glimpse into what craftsmanship looks like in a world of instant answers.
When everyone else is optimizing for speed, optimize for resonance.
Be the person AI can’t predict, the source it can’t paraphrase, the voice it can’t flatten into “neutral.” In the strip-mining era of AI, the moat for creators isn’t data, it’s depth. And like good coffee, depth takes time.