We are Entering the Era of Vibe Engineering

Code is just a part of the larger software cycle. However, it is the code that creates a lot of friction in the team. Ask most developers, and they’ll tell you that code reviews, meant to ensure quality, can sometimes feel like heated debates rather than helpful feedback sessions. 

Aravind Putrevu, director of developer GTM at CodeRabbit, believes that AI is about to soften the edges of this process. “The way I see it, AI brings impartiality to code reviews,” he said. “It doesn’t have an ego. It’s not there to nitpick or win an argument. It’s there to help.”

Putrevu pointed out that developers are often introverted by nature. Apart from team standups or collaborative debugging, the one place where they’re forced into meaningful conversations is during code reviews. “That’s where opinions clash, egos get bruised, and things can sometimes get ugly,” he said. “You’re defending your code while someone else pokes holes in it.”

AI tools like CodeRabbit are trying to change that dynamic by stepping in as the “first line of defence”. They catch the obvious mistakes, enforce style guidelines, and suggest improvements before a human reviewer takes over. “We’re not removing the human from the loop,” he clarified. “The developer still makes the final call, but AI can take care of the silly stuff.”

Click here to check out CodeRabbit on VSCode.

This new workflow—dubbed ‘vibe checking’—mirrors the vibe shift in coding. Earlier, developers wrote everything by hand. Now, with tools like GitHub Copilot, Cursor, and Lovable, AI helps generate significant chunks of code. The review process is just catching up.

The Rise (and Risks) of Vibe Coding

Another big trend Putrevu noted is the rise of “vibe coding” tools—applications like Lovable or Replit that let non-coders build apps using natural language prompts. 

On the surface, this seems like a revolution. Guritfaq Singh, co-founder at CodeRabbit, offered a word of caution. “It’s like learning to drive but not knowing that petrol is flammable,” he said. “You can go fast, but you’re missing the fundamentals. You don’t understand version control, code maintenance, or why code behaves a certain way.”

He recounted a recent interaction with a popular educator experimenting with an AI tool. “He didn’t know what YAML was. And this person was teaching others how to use these tools!” Singh laughed. It’s not his fault, really—it just shows that we need better guidance and more senior developers mentoring these new users.

And that’s where CodeRabbit comes in. “It acts like a senior engineer sitting next to you,” he says. “It guides you, reviews your pull request, and helps you avoid costly mistakes.” Compared to other tools, he said that CodeRabbit generates roughly 5 billion tokens per day. This is mainly for reviewing.

Scaling AI and the Growing Role of Developers

There’s no doubt AI is generating a lot of code. Cursor reportedly outputs close to a billion lines of code per day. Google recently said over 50% of the code in its repositories is now AI-generated, and YC Startups are generating 90% of their code with AI.

Alarm bells went off when Anthropic CEO Dario Amodei predicted that in less than six months, AI would handle 90% of coding. This is similar to what Sridhar Vembu, founder of Zoho, thinks. He recently said that 90% of what programmers write today is ‘boilerplate’.

But that doesn’t mean we need fewer engineers. Quite the opposite. There’s a whole new generation of developers coming in—many from non-traditional backgrounds. The demand is only going to grow.

However, there’s a catch. Developers who resist learning these tools could be left behind. “It’s like going from a hand-pulled rickshaw to an engine-powered one,” he said. “You’re still the driver, but if you don’t learn how to use the machine, you might become obsolete.”

On social media, there’s been chatter about AI tools replacing entire developer teams. Posts have gone viral claiming that a combination of Cursor and Claude could function as a $20/month senior engineer.

Putrevu laughed this off, saying that you can’t replace the experience and judgment of a seasoned dev with a few prompts. What you can do, though, is give every junior dev a powerful companion. 

In essence, AI tools today are not the new team—they’re the new teammate.

In the past, developers leaned heavily on Stack Overflow and IDEs for instant code assistance—whether debugging a cryptic error, understanding a library, or finding the right syntax. Putrevu says, “These tools helped individual productivity, not team collaboration.” 

While Stack Overflow served as a lifeline for lone developers, it rarely addressed issues like code consistency, review quality, or contextual feedback within a team setting. Similarly, IDEs were optimised for writing code, not understanding how that code fits into a broader system. 

This is where tools like CodeRabbit step in, not to replace Stack Overflow or IDEs, but to evolve the development experience into a more collaborative, review-driven process—embedding intelligent context and feedback right where the developer works.

Quality Still Matters

Despite all the promises, there are risks. Too much automation can result in what Putrevu calls “comment overload.” He jokes about a meme where a 10-line PR gets 500 comments, while a 200-line PR gets a casual ‘LGTM’. That’s the danger—when AI-generated code is accepted blindly.

Good code still needs good judgment. That’s why the human reviewer isn’t going away any time soon.

As for the future of vibe reviewing? “It’s not about replacing conversations,” he says. “It’s about making those conversations more meaningful. AI doesn’t shut down the vibe. It sets the tone.”

It remains clear that while the world is vibing to code, there is still no such thing as vibe engineering. While vibe coding can accelerate prototyping, it doesn’t replace software engineering. 

Real engineering involves long-term system design, reliability, scalability, and maintainability—concerns that current AI-generated code cannot handle. So, despite the hype, there’s no “vibe engineering”. It’s still just engineering, with or without the code typing.

The post We are Entering the Era of Vibe Engineering appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source