Vibe Coding vs Coding with AI

I think there’s a huge difference between “vibe coding” and “coding with AI”, and that difference comes down to one thing: you still need to know what you’re doing.

Vibe coders are often people who are new to programming. They don’t yet understand software architecture, how systems fit together, or how to translate requirements into a working solution. To them, AI can feel like a magic box that somehow produces software.

But…

Developers who use AI effectively are approaching it very differently. They understand the problem they’re solving. They understand the requirements, the trade-offs, and how the different components of a system interact. AI isn’t replacing that knowledge.. it’s amplifying it.

AI is another tool in the toolbox. It can generate code, explain concepts, automate repetitive work, and even suggest better approaches. But it doesn’t remove the need for engineering judgement. You still need to know when the code is correct, when it’s wrong, when it’s insecure, and when it’s solving the wrong problem entirely.

The best developers aren’t the ones who avoid AI, and they aren’t the ones who blindly accept everything it generates. They’re the ones who know enough to use it effectively, and know when not to use it too.

What do you think?