Every day, in newspaper offices around the world, editors sit down to an improbable task.
The day has yielded thousands of news events: births, deaths, political slugfests, tales from conflict and war, scientific findings, boardroom decisions, small acts of bravery, large betrayals, tales of bureaucratic action and inertia. The editor’s job is to reduce this to perhaps 100 stories that will appear across the daily newspaper.
One question drives it all: what, of all that has happened, most deserves to be on these pages?
That question has always belonged to editors. Now it belongs to everyone: the musician choosing between 50 AI-generated chord progressions, the brand team deciding which of a dozen copy variants goes live, the filmmaker selecting from multiple AI-generated scene treatments, the researcher determining which of the machine’s myriad hypotheses is worth following up.
Every creative and cognitive role, in every field, now involves this upstream act. Not of creating from scratch, but of choosing from tsunamis of abundance.
***
The Argentinian writer Jorge Luis Borges first imagined this in 1941.
He wrote of a Library of Babel, an infinite trove containing every possible book: every permutation of every letter in every sequence, producing every text that could ever exist, and every text that never could.
The catch is ruinous. Every answer is present, but no answer can be found. Every truth exists within the library, but no truth is locatable.
The librarians spend their lives wandering, descending into madness. Some begin to believe in the existence of a single book that holds the key. Others attempt to compile the catalogue of all catalogues; find the perfect arrangement. None of it works.
Infinite possibility has destroyed the idea of relevance. The Borges library is the horror of abundance, made absolute.
What Borges imagined as metaphysics has parallels in what generative AI is manifesting.
Artificial intelligence tools dramatically accelerate the ideation phase, producing in seconds more directions than a team once generated in days. And the question the librarians faced, of how to find what matters, is now the central creative and commercial challenge of our era.
***
Gabriel Garcia Marquez once said that a writer is known more by what he does not publish than by what he does.
This will be the scarce resource, in the age of AI: discretion, elevated to a creative act. Creative ideas will be infectiously abundant.
The decisive human contribution will be the “No, not good enough”; the evaluation, and the ability to wrest more from the machine and elevate its output.
This is where the human element endures. Taste and discernment are formed over years of exposure, failure, revision, and the quiet accumulation of reasons for preferring one thing over another (reasons that are often entirely valid, but almost impossible to express or fully explain).
Design and curation are the two disciplines that give this capacity practical form.
Design reaches far deeper than the surface. Think of it as an iceberg: what an audience or customer encounters, a coherent pattern, a navigable user interface, a curated set of clear choices, is the visible tip. The design intelligence that produced it is the mass below the waterline.
Curation operates at a different level: not how things are experienced, but what deserves to be seen at all.
In the age of AI, we need more curation and better curation simultaneously. More, because of the sheer volume. Better, because human curators are under mounting pressure from algorithmic recommendation systems that select not for quality but for engagement.
Algorithms have no embarrassment. Platforms that curate through engagement optimisation select relentlessly based on what people will unthinkingly watch rather than what they would thank you for having shown them. Hence the growing prevalence of AI slop.
Culture, over time, narrows toward whatever has already performed well.
The accountable curator is the corrective. Not because they’re always right, but because human curators have skin in the game. A critic who champions a terrible book or movie is diminished by it. A publisher who releases a fraudulent memoir answers for it publicly, and over time. A label that staked its reputation on an artist carries the result.
The machines already have a reputation for creating slop. A lot of humans do not.
***
Each of us will have a role to play too, in choosing wisely, furthering scepticism and aiding curation.
A Pew Research survey conducted last year found that 66% of adults under 30 would like a painting less if they found out AI made it. The market has long recognised and rewarded this distinction of human provenance.
By every rational measure of function, Apple Watch does more than any Audemars Piguet. Yet the Royal Oak, assembled by hand, carrying the accumulated meaning of craft and human presence unavailable to any algorithm, thrives and far outstrips it (in value, and in prestige).
The dominant narrative is that AI is becoming a creator. The more consequential shift is that humans are becoming editors.
As generative systems take over execution — writing drafts, composing music, generating visuals — the human role can move upstream from production to judgment.
The act of creation, then, becomes less about making something from nothing and more about deciding what should exist at all.
***
The Borges librarians went mad because they had infinite options and no principle of refusal: no way to discard the gibberish, dismiss the false, or tell slop from meaning. Taste, design, curation, command: the principle of refusal is what turns that library into something viable.
As Picasso put it: Art is the elimination of the unnecessary.
In the age of AI, that is both an aesthetic instruction and a survival strategy.
(Kashyap Kompella is a tech industry analyst and author of three books on AI)
