It was a summer holiday, sometime in the late 1990s. I was a young kid, around 8 or so. My dad had taught me the rules of chess before and we played from time to time. But this time was different: As we were about to start he stood up and walked away from the board, sitting down in a chair across the room and facing the other way.

Whenever I moved a piece I was to simply say the move "pawn from E2 to E4" out loud. My dad would then reply back with his move, which I could re-enact on the board to actually see what was going on. We were playing blindfolded chess. I remember telling him I was moving a piece from X to Y, while the piece wasn't on X. Without hesitating he told me exactly where the piece was. I lost the game pretty quickly, but I was fascinated.

Strong players generally don’t picture a perfectly detailed board like a photograph. They track a structured set of relationships:

• Where the important pieces are
• Which squares are attacked or defended
• Open files, diagonals and pawn structures
• Tactical relationships such as pins, forks and loose pieces
• What changed after each move

Interestingly, blindfold players often say their concentration is directed more toward relationships and legal possibilities than toward literally “seeing” the board. Their chess knowledge reconstructs much of the position automatically. The exceptional skill is therefore a combination of spatial working memory, pattern recognition, disciplined calculation and attention control—not simply photographic memory.

- Claude Fable 5

I never got deep into chess, instead I turned to computers and programming.

With the rapid rise of LLMs in programming over the last year and a bit, everything is changing drastically for software engineers. Everyone is trying to figure out how to effectively do things in this new world. Opinions are all over the place. What is clear to everyone is that if you use AI in a way where you spend little time saying what you want, and no time reading what it coded up - you are vibecoding and the resulting software is not going to last very long (if it works at all). So how much coding should you let the AI do, and how much should you control and read (and deeply understand)?

In limited ways this shares resembles to higher level programming languages. Before we had those the programmer was required to have a deeper understanding of the actual hardware and how a CPU actually executes instructions (things like managing memory, pointers, threads). We can forget about all these things if we program in Java or Python, but the list of things you do need to understand is still fixed (it's just a shorter and simpler list). The difference now is that when you use AI there is no real requirement to know anything anymore. I obviously think that if a seasoned programmer sits behind Claude Code, the quality output will likely be higher than if a non technical person does it (though that's hard to measure). The seasoned programmer can still get tired or bored, stop reading everything and just let the AI take control. And humans are inherently lazy, it's hard to stay sharp and focussed when it feels like you don't have to because the AI is at the wheel anyway.

Thus in many ways programming with AI is the opposite of blindfold chess: you don't have to pay attention every turn, you don't have to remember what the important pieces are, the details of the tactical relationships (such as code interfaces and APIs). But I'm here to argue that the skills that define a strong blindfold chess player are the same as those of a programmer who can thrive behind a Claude Code terminal whilst not reading nor writing any code. And many really good programmers had already acquired these skills prior to LLMs. For example being able to have a deep mental model (in your head) of how code weaves together, and how it's always an expression of a more abstract idea, or goal. And being able to simplify or improve upon some existing implementation by simply thinking about it deeply. It's been harder for me to build such a mental model of software I didn't actually physically write (or read very much of), and it's still hard for me to get the right balance between talking talking in detail versus at a high level at different times.

As these LLMs and their tools get better and better, those who can really tame these beasts and stay in control can move faster and smarter than ever before. Not limited to the speed at which one writes/reads the underlying code. Did you know the very best blindfolded chess players can play many chess games at once?

The problem will be that unlike chess where it's easy to rank someone's skill, programming effectively this way will become impossible to measure (as the AI will solve any quantifiable metric you can throw at it).