I called him David. His name was John.
I do this a lot. Not random names swaps, but ones that feel right. I'll call a David "John" or a Michael "Mark". Never something insane like, "Sebastian" instead of "Steve". My mistakes follow a pattern. Why?
They don't even look alike. David and John share zero letters or phonetic resemblance. But they're practically the same. But my brain isn't malfunctioning. My brain, your brain, everyone's brain, organizes details into clusters of meaning. Into what AI researchers, in their most poetic way, call latent space.
We don't process knowledge like computers. We don't recall facts like a dead-eyed accounting machine. We navigate a latent space: a vast shifting landscape of relationships and patterns, where things exist next to each other not because they look alike, but because they mean alike.
John and David: classic, American, biblical male names. Emma and Sophia: Timeless, vaguely aristocratic, possibly prone to sipping tea on balconies. They exist near each other in our minds, which is why, when I reach for one, I sometimes grab its neighbor instead.
Now, machines can think in latent space too. Large Language Models like GPT don’t just store words, they map meaning. They make connections. They can even tell jokes!
Latent space is an incredibly powerful mental model. Not just as a machine learning concept, but as a grounding model for how our minds work.
Chess, Poker and Flautists
High-level chess players frequently excel in poker, despite rather glaring differences between the games.
Dan Harrington was a former U.S. chess master. He didn’t casually dabble, he won the 1995 World Series of Poker (WSOP) and authored a foundational poker strategy book. The transition was seamless, almost natural.
On the surface, this makes zero sense. Chess has no hidden information. Poker bathes in secrecy and bluffing. Chess is cold, deterministic logic, poker a blend of luck and deception.
But in latent space, they’re neighbors.
Both games are about pattern recognition, strategic anticipation, and mental endurance under pressure. Chess players don’t become flautists when they're bored. They drift instinctively toward another game in the same latent neighborhood.
Just as chess masters navigate effortlessly to poker, great entrepreneurs navigate latent space between seemingly unrelated industries and ideas.
Entrepreneurs
A major differentiator for successful entrepreneurs is the ability to connect ideas across different domains. They see connections others don’t. These entrepreneurs intuitively traverse latent space, spotting connections invisible to others.
Steve Jobs didn't just obsess over computers, he explored calligraphy, aesthetics, and intuitive design. These fields aren't obviously related, yet Jobs effortlessly navigated latent space between them. As a result, computers became beautiful and intuitive, not soulless machines.
Great entrepreneurs don’t invent from scratch. Instead, they move laterally and reveal hidden connections across seemingly unrelated fields:
Apple combined calligraphy & computing to bring beautiful typography to PCs
Ford took meatpacking assembly lines and car manufacturing to modernize mass production
Uber combined taxis and GPS on smartphones to bring forth modern ride sharing
Dyson took inspiration from industrial cyclone separators used in sawmills and applied the same principles to vacuum design, eliminating the need for bags and filters
These innovators didn't create from thin air. They navigated latent space better than anyone else, combining seemingly unrelated ideas to yield breakthroughs.
LLMs Join the Party
Humans have always intuitively navigated latent space, but machines traditionally struggled. ChatGPT changed the game. It transformed rigid data into fluid meaning maps, finally giving machines a human-like intuition.
Large Language Models (LLMs) like GPT don’t just store words—they map meaning, recognize patterns, grasp metaphors, and even tell jokes. Machines now explore latent space just as we do, making connections rather than retrieving isolated facts.
This marks a profound shift: what once was uniquely human intuition can now be modeled, mapped, and explored intentionally.
Latent Space as a Foundation for Creativity
There’s an elegance in freely navigating latent space. It manifests itself as wittiness, creativity, and adaptability.
Those who intuitively grasp latent space can make creative leaps that leave others behind. Novelists don't just weave plots—they connect psychology, history, and personal experiences into powerful narratives. Chefs blend flavors from different cultures into inventive dishes. Comedians link storytelling, observation, and surprise to reveal hidden truths.
These aren’t random talents. They're manifestations of our innate ability to navigate latent space. When comedians become novelists or musicians turn into skilled coders, it’s because they've identified latent connections others overlook.
Intentionally Exploration
Most people, however, don’t naturally think this way. We compartmentalize knowledge into rigid categories: chess here, poker there; cooking recipes in separate drawers labeled by cuisine; industries as isolated vertical silos. This rigid categorization limits our creativity.
But imagine intentionally navigating latent space, explicitly moving laterally, rather than vertically. What if exploring conceptual neighborhoods became as deliberate as visiting places on a map?
Today, new tools can help us intentionally and explicitly navigate latent space. They enhance our natural instincts, turning intuition into deliberate practice.
Rather than passively recalling facts, you can now intentionally ask:
“What's conceptually close to what I already know?”
“What ideas haven't yet been linked, but should be?”
“How can I move laterally rather than starting from scratch?”
Understanding latent space explains how creativity, innovation, and learning truly function. It explains why masters of chess easily transition to poker, why entrepreneurs disrupt industries by merging unrelated ideas, and why creative professionals effortlessly blend knowledge from seemingly unrelated fields.
When you consciously navigate latent space, you won’t only notice connections—you’ll create them.
I still occasionally call a David “John”, but now I tell myself it’s a feature, not a bug.