Some of the Best Users of AI Don't Know How to Code. They Use English.

Jensen Huang put it plainly at London Tech Week last year: “There’s a new programming language. This new programming language is called ‘human.’” He’s said versions of that line a couple of different times since. You don’t need C++ or Python to get a computer to do what you want anymore. You need to be able to say clearly what you want.

I didn’t think much of it the first time I heard it. Then I started noticing it happening around me.

The pattern I keep seeing

I’ve been meeting more people lately with zero coding background who are building real things with AI. Not toy projects. Working tools, working products. What struck me wasn’t that they were using AI well. It was how they were using it.

They don’t assume the AI understands them. Because they’ve never written a line of code, they don’t have the confidence that comes from knowing the syntax will just work if you get it right. So they take the time. They explain what they’re building in plain language, then ask the AI to explain it back simpler, until it clicks for both of them.

Before they open the chat window, they ask themselves what they’re actually trying to build. Then they hand that to the AI and get it to stress test the thinking. Poke holes in the assumptions. Point out what they missed.

By the time they start implementing, they’ve got a sharper version of the problem than most developers get from a full requirements document. The AI didn’t hand them the answer. It helped them ask a better question.

English was always the hard part

Huang’s line can make English sound easier than code. It isn’t. You still have to describe the problem precisely. Syntax used to hide weak thinking because you could write technically correct code that solved the wrong problem. Plain English exposes gaps sooner.

Testers already have this skill

Manual testers already practise this skill. You explain bugs clearly enough for a developer who never saw the issue to reproduce and fix it. That takes precise language and a clear account of what happened.

Those habits transfer well to working with AI.