AI is automating many aspects of knowledge work, but some might end up becoming more important than before.
Naval Ravikant — the entrepreneur, investor, and philosopher best known as the co-founder of AngelList — has long been a source of contrarian but prescient thinking on technology and human potential. In a recent conversation, he turned his attention to an idea that cuts against the grain of most AI discourse: that as artificial intelligence makes creation cheaper and more abundant, the rarest and most valuable skill won’t be the ability to generate — it will be the ability to judge. And at the centre of that judgment, he argues, is design.
“Design is always a moat,” Ravikant says. “Good design requires good taste, requires a good eye. AI can generate anything you want, but that means the world is going to be full of average. And then eventually people will get tired of average. So they’ll say, I want extreme. So then AI will fill the world full of extreme, right?”
The end state he’s describing is something close to total creative saturation — every permutation of every idea, every aesthetic, every product configuration, eventually existing somewhere. To illustrate this, Ravikant reaches for a literary reference:
“There’s going to be every possible permutation of everything imaginable out there. Think of it like Borges’s The Library of Babel — great story, by the way. It’s like every permutation of every book is in this library, including the refutation of every book, the translation of every book, the opposite meaning of the book, et cetera.”
In Borges’s famous short story, an infinite library contains every possible arrangement of letters — which means it contains all truth, but also all nonsense, with no way to distinguish between them. Ravikant is suggesting AI-generated content is heading somewhere similar: not a shortage of material, but an overwhelming excess of it, with no inherent signal as to what is good.
In that world, he argues, the bottleneck shifts entirely: “So at some point you realise that the core task becomes editing, it becomes selection, it becomes curation. So developing your judgment, your taste is really important. And so that means having a good eye for products, not getting caught up emotionally, disliking something just because somebody else did it.”
Ravikant goes further, pointing to the habits of mind that build that kind of taste: “Always trying to learn, always having that intellectual curiosity to figure out what makes something great. And then if you’re hardworking, you can take that taste and you can literally translate it into creating something great.”
As AI tools flood the market with generated content, interfaces, and products, the companies that stand out are the ones with strong, coherent design vision at their core. User interface design has long been underrated as a competitive advantage; in an AI-abundant world, it may become the primary one. As one analysis put it, the best technology doesn’t always win — products with thoughtfully designed experiences consistently outperform technically superior competitors.
The point extends beyond visual design. Taste, in Ravikant’s sense, is about judgment across every dimension of a product — what to build, what to cut, what to say and what to leave unsaid. It’s a faculty that requires cultivation, curiosity, and a certain independence of mind. Renowned AI researcher Fei-Fei Li has made a parallel argument about creativity more broadly — that AI, for all its impressive capabilities, lacks the kind of abstract, paradigm-shifting insight that produces truly original breakthroughs. The ability to see what makes something genuinely great, and to bring that into existence, remains a distinctly human capacity.
What Ravikant is pointing to, ultimately, is a rebalancing. For decades, competitive advantage in tech accrued to those who could build — engineers, coders, people who could translate ideas into functional software. AI is rapidly democratising that. What it cannot democratise is the eye that knows which thing is worth building in the first place, and what it should feel like when it’s done. In a world where the Library of Babel is being printed in real time, the person who can find the right book — and recognise it as such — holds the moat.