Vibe coding - is it even real coding?
Published on: | by Uditha Atukorala
If you’ve been following tech in February, 2025 you have probably heard of “vibe coding”, a term introduced by Andrej Karpathy1 and very likely made popular by Y Combinator (YC).
YC titled their March 5th, 2025 Lightcone podcast episode “Vibe Coding Is The Future”2 and claimed 25% of the W25 batch have 95% of their code generated by AI.
“There’s a new kind of coding I call “vibe coding”, where you fully give in to the vibes, embrace exponentials, and forget that the code even exists.” - Andrej Karpathy (AI researcher)
“This isn’t a fad. This isn’t going away. This is the dominant way to code. And if you are not doing it, you might just be left behind.” - Garry Tan (YC CEO)
Large language models (LLMs) can generate code so there’s no denying “vibe coding” is real. But is “vibe coding” real coding?
I think so, let me explain.
Fundamentally a programmer’s task is to write code to run on computers to produce some output. This could be as low level as writing assembly to control a dot-matrix display using a micro controller, or as high level as using a markup language (like HTML or LaTeX) to display formatted text on a computer screen.
If the desired output is the same, does it really matter how a programmer writes code?
A programmer may use a code editor (like vi) or an IDE (like VSCode) to write code, but the output of the code is the same.
A programmer may reference the documentation, ask a colleague or use the internet to find information, but the output of the code is the same.
If a programmer use an LLM to assist and vide code, it’s still real coding since the output of the code is the same. If vibe coding helps programmers to offload tedious and repetitive tasks to LLMs, and improve their productivity, that’s even better!
What if a programmer can only vibe code (i.e. they can’t code on their own)? Arguably, they shouldn’t be called programmers. They are users of LLMs.