DevOps Pioneer: Vibe Coding 100x Bigger Than DevOps Revolution
Gene Kim has written six books that helped define how modern software gets built and deployed. His seventh book took half the time to write, used 70 million AI tokens in the process and earned praise from his editor as “the cleanest handoff” manuscript they had ever received.
The man who gained fame from DevOps and transformed how enterprises think about software delivery is now making an even bolder claim: AI coding represents a transformation “10 to 100 times bigger than DevOps” — and he’s betting his reputation on the potential of “vibe coding.” Indeed, Kim described writing the book as “the most fun I’ve had in my entire career.”
The Book That Writes Itself
Kim’s latest book, “Vibe Coding”, co-authored with Steve Yegge, a software engineer at Sourcegraph (and formerly at Amazon and Google), isn’t just about AI transformation — it’s a product of it. The authors used 70 million AI tokens during the writing process, making it simultaneously a manual for AI-assisted development and proof of concept for their methodology. The book will be released on October 21.
“We used it [AI] like Seinfeld said, comedy is a game of tonnage,” Kim told The New Stack. “For every joke that makes it on stage, it’s like 100 jokes that didn’t. We used AI to generate five ways to tell a story, the age of inclusions, right?” The result? Their editor called it “the cleanest handoff of any book that we’ve organized — closest manuscript ready, shortest editing times.”
As the title proclaims, the book centers on vibe coding — a term popularized by Dr. Andrej Karpathy that describes turning your brain off and letting AI code for you. But this is not casual experimentation. Kim calls their work “the professionalization of vibe coding,” moving from novelty to systematic methodology.
From DevOps Revolutionary To AI Convert
Kim’s journey to AI advocacy was not immediate. Despite his illustrious career studying high-performing organizations, he admits he “never wrote production code” until recently. That changed when he met Yegge, an engineer famous for his technical rants.
“I met him last year, and man, we just had so much in common,” Kim recalled. “He’s a professional developer who wrote over a million lines of production code by hand and probably writes very few lines of code by hand now.”
The breakthrough moment came when Kim wanted to solve a personal problem: turning piles of screenshots from video calls into searchable clips. Working with Yegge, they built a functioning tool in 47 minutes. “I’ve been taking these screenshots for 15 years, and I’ve never done anything with them, maybe one or two times,” Kim says. “That changed my life.”
The FAAFO Revolution
In their book, Kim and Yegge outline five “superpowers” that vibe coding provides, captured in their FAAFO framework:
- Fast: “We routinely ship features in minutes that used to require weeks or months.”
- Ambitious: “Giant leaps in your decades-long aspirations and goals can be achieved in a weekend.”
- Autonomous: “One developer with five agents feels like a whole team, and you have access to information that used to require going to other people.”
- Fun: “The slog of typing in code by hand disappears, and instead, you unleash your ability to create things like never before.”
- Optionality: “Parallel experiments cost pennies, so you never have to stick to the first idea that you try.”
The book argues that we are approaching a world where “all you have to do is explain what you want, and your words become working software almost instantly. When something’s not right, you don’t spend hours debugging — you just describe what needs to change.”
Enterprise Reality Check
This isn’t just Silicon Valley hype. Kim points to real enterprise adoption with measurable results. The book cites the example of Fernando Cornago, global SVP of digital and e-commerce at Adidas, which has 700 developers using GitHub Copilot daily. Cornago told Kim that Adidas ran a comprehensive generative AI (GenAI) pilot program that delivered a 2x increase in “happy time” — hours spent on actual coding, testing and skill development rather than meetings and administrative tasks.
However, the benefits were not universal. Teams with loose coupling and deployment independence saw dramatic improvements, while those trapped in legacy, tightly coupled systems with infrequent deployments saw minimal gains. “When you’re under those conditions, vibe coding is not going to help you,” Kim said, “because the bottleneck is so far removed from coding.”
More Developers, Not Fewer
One of the book’s most counterintuitive arguments challenges developer replacement fears. Kim predicts vibe coding will create demand for more developers, not fewer, drawing an analogy to digital photography. Kim said that when costs dropped dramatically (cell phones), photo taking exploded from millions to trillions annually.
“Junior developers will not become redundant. Far from it,” the book states. “Their role is evolving. Instead of primarily executing leaf tasks, they might become the ‘station leads’ of the kitchen, who help integrate contributions from non-engineers across the company.”
The authors describe the trend where people outside traditional engineering roles — UX designers, product managers, infrastructure ops — use AI to contribute directly to codebases. “A junior engineer, like a junior doctor, is still highly trained and can be super valuable in helping this new generation of budding ‘field medics’ contribute directly to the code,” the book says.
Shadow AI Reality
The book documents how “vibe coding is starting to happen everywhere in the organization, where people are waiting for developers or engineers who aren’t getting prioritized. In the past, these people were either stuck, had to use outside vendors or had to escalate up the hierarchy. Now, they can create the software themselves — building prototypes, fixing issues and maybe building features.”
This “shadow AI” trend represents a fundamental shift in organizational dynamics, democratizing software creation beyond traditional engineering teams.
Leadership in the Age of AI
The book also explores how managing AI agents resembles traditional team leadership. “Steve was dumbfounded: he had no choice but to be a team lead again, despite having explicitly stepped down from leadership and thinking he was a solo dev,” the book reads. “With vibe coding, you have all new team-related concerns, and they’re not quite the same as human teams.”
Industry Validation
Moreover, the book carries heavyweight endorsement from Dario Amodei, CEO of Anthropic (makers of Claude): “Take it from someone who employs many of the best coders in the world: The ‘vibe coding’ way of working is here to stay. If you’re going to be doing any coding at all — if you’re going to use that comparative advantage — you need to get involved with vibe coding today.”
Dr. Erik Meijer, a legendary computer scientist and the designer behind Facebook’s Hack language and lots more technology, provides a key validation for the book: “We are likely the last generation of developers who will write code by hand. … But let’s have fun doing it!”
Getting Started
For skeptics and newcomers alike, the book offers practical advice.
“Start small and start today,” the book states. “Hand your AI assistant a self-contained task, watch it stumble, correct it, and tighten the loop until the stumbling stops. Then double the scope. By the tenth iteration, you’ll notice the conversation feels less like tooling and more like leading. That’s the moment you’ve become the head chef.”
The Bottom Line
Whether Kim’s boldest predictions prove accurate, the enterprise adoption data suggests something significant is happening. For a DevOps pioneer who spent decades studying technology transformations, his conviction that this represents the biggest shift he’s witnessed — captured in a book that itself demonstrates the power of AI-assisted creation — deserves attention.
Kim’s and Yegge’s “Vibe Coding” is a serious effort that represents a vast amount of work accomplished in a short period of time. It’s chock-full of valuable information for vibe coders, particularly in the enterprise.
As the book puts it: “Vibe coding can change your life, like it changed ours. Mastering vibe coding enables you to take on ambitious projects, work faster and more autonomously, and, perhaps most importantly, rediscover the joy of building software on your own terms.”
Vibe Coding With Gene Kim
Beyond just talking about the book, Kim and I had a fascinating and quite fun time discussing technology, even joining up to vibe code a little app I’ve been wanting to write to keep track of offensive accomplishments of my beloved Baltimore Ravens. With the NFL season about to start, I was hoping to have an app ready for Week 1.
Kim selected Vercel’s v0 as his vibe coding tool of choice (primarily because he was having trouble logging onto Lovable), and we just told it — via voice interface — who we wanted to track, what parameters, etc., and bip, bap, boom! We had the makings of a nice little usable app. He took it through a couple of iterations to iron out quirks and add stuff, but it worked like a charm.
So, it seems to me, the question is not whether AI will change software development — it’s whether developers will adapt quickly enough to stay relevant in a world where mere conversations create code.
