What is Vibe Coding?

Vibe coding (also called vibecoding) is a relatively new approach to software development that emerged in early 2025. The term was coined by Andrej Karpathy, a renowned AI researcher and former Tesla AI leader, in February 2025.

What is Vibe Coding?

Vibe coding is an approach to producing software by relying on artificial intelligence, where a person describes a problem in a few sentences as a prompt to a large language model (LLM) tuned for coding. The LLM then generates software based on the description.

The key characteristic that defines vibe coding is the developer's relationship with the code:

When practicing vibe coding, you "fully give in to the vibes, embrace exponentials, and forget that the code even exists." It involves accepting AI-generated code without thorough review - you don't read the code that the AI writes.

As Karpathy described it: "I'm building a project or webapp, but it's not really coding—I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works."

How Does Vibe Coding Work?

At its core, vibe coding is about communicating with AI in natural language to build apps. Instead of writing code, you describe what you want your app to do, and AI tools handle the technical implementation. This frees you up to be more of a manager or director, focusing on the application's outcomes, functionality, and user experience.

For example, instead of writing a complex SQL query, you could simply say: "A social media management dashboard showing analytics, scheduled posts and engagement metrics across platforms." The AI assistant would then generate the code to get the job done.

Benefits and Limitations

Benefits:

Vibe coding changes the whole game of software development by:

• Making software development accessible to non-developers
• Massively speeding up development time
• Being perfect for prototyping
• Boosting developer productivity by letting engineers focus on problem-solving rather than manual coding

Limitations:

Vibe coding has raised concerns about understanding and accountability. Developers may use AI-generated code without fully comprehending its functionality, leading to undetected bugs, errors, or security vulnerabilities.

Karpathy acknowledged that vibe coding has limitations, noting that AI tools are not always able to fix or understand bugs, requiring him to "experiment with unrelated changes until the problems are resolved." He concluded that he found the technique "not too bad for throwaway weekend projects."

Simon Willison, a software developer, cautions: "Vibe coding your way to a production codebase is clearly risky. Most of the work we do as software engineers involves evolving existing systems, where the quality and understandability of the underlying code is crucial."

Current Impact

In March 2025, Y Combinator reported that 25% of startup companies in its Winter 2025 batch had codebases that were 95% AI-generated, reflecting a shift toward AI-assisted development within newer startups.

The goal of vibe coding is to create an AI-powered development environment where AI agents serve as coding assistants making suggestions in real-time, automating tedious processes, and even producing standard codebase structures.

Vibe coding represents a significant shift in how software is created, making programming more accessible to those without traditional coding skills while raising important questions about code quality, security, and the changing role of software developers.