The New York Times Just Made the Case for Vibe Coding—Here’s the Deeper Story

This morning, millions of people woke up, put in their earbuds, and heard The New York Times‘ most popular podcast, The Daily, explain what vibe coding is and why it might reshape software development forever. Host Natalie Kitroeff brought in Kevin Roose, the Times‘ tech columnist, to walk listeners through the phenomenon. They built a website live on air using Claude Code and framed it as the AI trend that could change who builds software and how.

Gene Kim and Steve Yegge wrote the book on vibe coding—literally. Vibe Coding: Building Production-Grade Software with GenAI, Chat, Agents, and Beyond is available now, and it goes considerably deeper than any podcast can in twenty minutes. If today’s episode sparked your curiosity, here’s what Gene and Steve want you to know next.

“Deadly Serious”—Not a Party Trick

Kevin Roose is right that “vibe coding” is an inspired term. But as Dario Amodei, CEO of Anthropic, writes in the book’s opening pages, it’s also a misleading one. The jokey framing can make the whole enterprise seem unserious.

“Vibe coding is deadly serious,” he writes. “As of mid-2025, it’s the only coding game in town.”

The term was coined by AI researcher Dr. Andrej Karpathy to describe a new mode of building software: You describe what you want, AI writes the code, and you iterate in conversation until you have something that works. No more typing line by line. No more hunting syntax errors on Stack Overflow at midnight.

Gene and Steve’s central metaphor—one that runs through every chapter of the book—is that you’ve become the head chef of a professional kitchen. Your AI agents are a brigade of extraordinarily trained, lightning-fast sous chefs who have memorized every cookbook ever written. Your job is no longer to cook every dish by hand. It’s to set the vision, taste early and often, and make sure nothing leaves the kitchen you’re not proud of.

That shift in role, not just in tools, is what the mainstream coverage keeps underselling.

What The Daily Got Right—and What the Book Goes Further On

The episode hit two themes that Gene and Steve have been living and writing about for more than a year.

The democratization story. Roose and Kitroeff built a website together live on air, with no coding background. This is real, and the book’s case studies go further. Christine Hudson, who hadn’t written code in nearly twenty years, completed a Google Calendar sync tool in her first vibe coding session—and finished before Gene and Steve did. A UX designer named Daniel built a feature his engineering team had been too busy to touch, complete with tests, and impressed everyone. These aren’t edge cases. They’re the pattern.

The productivity story. Gene documented building a tool to query the entire book manuscript like a SQL database—working during breaks, after writing sessions, while brushing his teeth—across 251 prompts and 35 commits. When he analyzed the Git history, he had been 16x faster than his historical average and 5x faster than his previous best day. Gene and Steve call the full value proposition FAAFO: fast, ambitious, autonomous, fun, and optionality. Speed is only the first dimension.

The Part That Doesn’t Make the Highlight Reel

Twenty-minute podcast episodes don’t have time for what happens when vibe coding goes wrong. The book does.

Gene and Steve are not boosters who gloss over the risks. They’re researchers and engineers who made real mistakes and documented them honestly. Steve watched a coding agent silently delete 80% of his automated test suite across dozens of commits. By the time he found out, there was no clean rollback. Gene spent three days untangling a sprawling thousand-line function that AI had written and that he couldn’t understand or safely modify. A friend nearly wiped the storage device of a hundred-pound CNC machine because the AI’s proposed commands scrolled by too fast to catch.

As they write: “It’s like playing a slot machine with infinite payout but also infinite loss potential.”

DORA’s 2024 research—the same research program Gene helped create—found that every 25% increase in AI coding tool usage correlated with measurable increases in deployment failures. The magic is real. So is the danger. The book teaches both.

This Isn’t Only a Story About Developers

One of the book’s most important arguments barely surfaces in most coverage: Vibe coding changes who gets to participate in software creation at all.

Every time programming became easier, the world got more software, not less. More developers, more industries, more jobs. The web. The smartphone. The cloud. Each time the barrier dropped, demand exploded. Gene and Steve argue we’re watching the same pattern again—only faster, and touching more roles at once.

Product managers. UX designers. Infrastructure engineers. Analysts. People who left coding years ago because it became overwhelming. All of them now have a path back in. Kent Beck, creator of Extreme Programming, summed it up: “I feel young again.”

Gene lived this. He hadn’t written production code in seventeen years when he sat down with Steve for a forty-seven-minute pair programming session in early 2024. By the end, he had a working video excerpting tool that had been sitting on his “someday” list for years. That night, he couldn’t sleep—not from worry, but from excitement.

Two Skeptics. That’s the Whole Argument.

What makes this book different from the wave of AI enthusiasm flooding the internet is who wrote it.

Steve Yegge coded in Java without an IDE until 2011 and refused to learn Git until 2021. He spent thirty-five years writing over a million lines of production code at Amazon, Google, and Sourcegraph. He is a self-described late adopter who built his reputation on being the person who called out hype. Gene spent decades doing rigorous, data-driven research—surveying over 36,000 technical professionals across six years for the State of DevOps Reports. Neither of them chases trends.

They were both converted by the evidence. That’s the book’s origin story and its most persuasive argument. If these two skeptics are this enthusiastic, something real is happening.

As Steve wrote in “Revenge of the Junior Developer”—an essay Amodei cites by name in the foreword—the trap to avoid is the work-deferral fantasy: “Saying ‘It’ll be way faster in 6 months, so I’ll just push this work out 6 months’ is like saying, ‘I’m going to wait until traffic dies down.’ Your drive will be shorter, sure. But you will arrive last.”

The tools are better than they were six months ago. They will be better still in six months. But the developers building instincts and workflows right now will be the ones who know how to use those better tools when they arrive.

Go Deeper

If you want to go further still, Gene is hosting the Enterprise AI Summit this April, where Steve will be speaking alongside technology leaders navigating exactly these questions at scale. It’s the place to bring your team when the podcast and the book have convinced you it’s time to act—and you want to figure out what that looks like inside a real organization.

Get the book →

Register for the Enterprise AI Summit →

- About The Authors

Leah Brown

Managing Editor at IT Revolution working on publishing books and guidance papers for the modern business leader. I also oversee the production of the IT Revolution blog, combining the best of responsible, human-centered content with the assistance of AI tools.

Follow Leah on Social Media

•