What You Need to Know About Vibe Coding

Author: Zeb Hermann, Keith Messick

Original: Read the full article

Vibe coding is rapidly changing how software gets built. This article draws from Vercel's State of Vibe Coding Report, highlighting the key takeaways.

In February 2025, Andrej Karpathy coined the term vibe coding: a new way of programming with AI—"surrendering to the vibes, embracing exponential speedups, and even forgetting that the code itself exists."

Within months, vibe coding had noticeably changed how both developers and non-developers work. Over 90% of US developers now use AI coding tools, with adoption rising rapidly across other roles as well. The notion that "English is the fastest-growing programming language" is no longer just a joke.

From No-Code to Vibe Code

Fifteen years ago, no-code tools let people build applications by dragging and dropping components. This marked a major wave of "what you see is what you get" (WYSIWYG) software tools: users edited content in a visual interface while seeing the result in real time.

Today, vibe coding tools like v0, Lovable, and Replit push this path further, into a "what you say is what you get" era: you describe what you want, and the system generates the application, website, or workflow. The barrier continues to drop, and with it, team collaboration patterns, project division of labor, and the boundaries of what individuals can accomplish on their own are all shifting.

Core Advantages of Vibe Coding

Increasingly, major companies are openly embracing vibe coding. Amazon launched its own tool, Kiro, and over 30% of Google's new code is now generated with AI involvement. But this shift isn't limited to big tech.

Teams of all sizes can now complete in days what previously took months. For example, in March 2025, Garry Tan noted that a 10-person AI-powered team can now produce what previously required 100 engineers. Vibe coding is compressing the minimum unit of productivity from "team" to "individual."

Faster iteration, lower cost, and a shorter distance between idea and execution are becoming increasingly real. The true barrier is converging on whether you can articulate your requirements clearly.

Vibe Coding Is for Everyone

Vibe coding is redefining "who can program" and "what programming means." Traditional developers rely on programming languages, code comprehension, and long training cycles. In a vibe coding context, communication skills, a basic understanding of technical principles, and a clear enough problem statement are often sufficient to get started.

In fact, many vibe coding tools' primary users aren't professional developers. About 63% of people exploring these tools are non-developers, using products like v0 and Cursor to simplify their workflows or build custom solutions for specific problems.

This increased accessibility creates new opportunities, but also new requirements: as more users can directly build software, platforms must embed more safety and guardrail mechanisms by default, rather than assuming every user has a complete engineering background.

Building Guardrails for Vibe Coding

As adoption grows, security capabilities need to be built directly into the tools themselves. Since not every user has complete security training, platforms must assume more default protective responsibility.

In other words, vibe coding tools need to handle at the product level the security concerns that users may not explicitly raise but that are genuinely important. This is especially critical given that more and more people are entrusting data, trust, and even business processes to these tools.

For Vercel, this means providing automatic intervention capabilities: real-time vulnerability scanning and blocking, emergency stops when necessary, and catching high-risk issues at the deployment stage.

The most successful vibe coding platforms will likely be those that build security capabilities directly into the interface and default workflows.

The Future of Vibe Coding

Interest in vibe coding will only continue to rise.

Enterprises can certainly embrace vibe coding to ship faster with leaner teams; they can also continue using more constrained engineering processes for critical systems, reserving AI for lower-risk components. What truly matters isn't picking a side, but whether you've started preparing for this change.

Whatever you think of the name, this shift is already happening. To learn more, check out the full State of Vibe Coding Report.