Vibe coding is an AI-dependent programming paradigm introduced by Andrej Karpathy in February 2025. It allows programmers to generate software by describing problems in natural language to AI tools, rather than writing code manually. In vibe coding, developers “fully give in to the vibes” and let AI handle most of the coding process, including writing, editing, and debugging code.
Key aspects of vibe coding include:
- Natural language interaction: Programmers describe desired outcomes using plain language or even voice commands.
- AI-generated code: Large language models (LLMs) translate descriptions into functional code.
- Minimal code review: Users often accept AI-generated code without fully understanding it.
- Rapid development: Vibe coding can significantly speed up the creation of prototypes and small-scale applications.
- Accessibility: It lowers the barrier to entry for non-programmers to create software.
Vibe coding is being used for various purposes, including:
- Personal tools and “software for one”
- Rapid prototyping for startups
- Educational purposes to teach programming logic
While vibe coding offers benefits in terms of speed and accessibility, it also raises concerns about code quality, security, and developers’ understanding of their own projects. As of March 2025, about 25% of Y Combinator startups in the Winter 2025 batch had codebases that were 95% AI-generated, indicating a significant shift towards AI-assisted development.
Donate with Cryptocurrency!