I’ve been building with LLMs a lot lately. One thing has become very clear: there’s a difference between coding and programming. And that difference matters more now than ever.

Coding is typing. Writing syntactically correct instructions in a language a machine understands.

Programming is thinking. It’s understanding how the pieces fit. Knowing what needs to happen before the first line is even written. Choosing the stack. Sketching the system. Deciding what’s worth doing, and what can wait.

LLMs are great at coding. They can type circles around you. But they’re not programmers. They don’t know what you’re building. They don’t know the tradeoffs. They don’t care if your app needs to scale tomorrow or just survive the demo.

You still need a programmer. Someone who can say:

“That shouldn’t be a monolith.”

“We’ll want a separate front end so we can swap APIs later.”

“This should be stored locally, not sent to a server.”

“Let’s split this into phases so we don’t build too much too soon.”

That’s the work. The thinking.

AI might take the keyboard. That’s fine. Just like calculators took arithmetic. But someone still needs to decide what’s worth calculating.

Same with code. LLMs can generate it. But someone still has to guide it. Shape it. Say no to a clever solution that creates a bigger mess later.

Typing isn’t writing. Writing is thinking.

Coding isn’t programming. Programming is thinking.

That part’s not going away.

PS: This blog was published via email using Simple Butler Service.