The King Has Changed His Mind

David Heinemeier Hansson doesn't flip-flop lightly. When DHH—the brains behind Ruby on Rails and 37signals—speaks, the entire Rails world pays attention. So when he recently admitted he'd done a complete 180 on AI coding tools, the ripples spread way beyond the Ruby bubble.

During a podcast chat with Robby Russell, DHH didn't just give AI a polite nod. He called taking it seriously a "professional obligation." Coming from someone who built Rails on ideas like developer craftsmanship and elegant, hand-crafted code? That's not a small shift. That's a philosophical U-turn.

What Vibe Coding Actually Looks Like

You've probably heard "vibe coding" thrown around—it's become shorthand for working with AI assistants to churn out code fast. But DHH's take goes deeper than copying AI outputs into your project.

At 37signals, vibe coding turned into a real internal experiment. Designers and product managers now jump into code right alongside developers. And get this—the sacred six-week Shape Up cycle, something 37signals has protected for years? It's "up for reconsideration." This isn't about replacing developers. It's about opening up the creation process and speeding things up.

The new mantra? "Git reset and try again." When experimentation becomes cheap, AI-generated code that misses the mark isn't failure—it's just fuel for the next attempt.

Basecamp 5: From Theory to Reality

Here's the thing—this shift isn't stuck in theory. It's already shipped. Basecamp 5 launched recently, and it's the first major release where these AI-augmented practices run wild in production at 37signals.

The results tell a new story: faster iterations, cross-functional teams shattering old silos, and a willingness to let AI handle the boring boilerplate so humans can focus on the interesting puzzles.

Rails: Steady Foundation, Fresh Ideas

What makes this shift stand out? Rails itself hasn't needed dramatic changes to get on board. The backend stays so stable that a model file from 2026 fits right next to one from 2013. That stability is actually a superpower in the AI era—AI tools can generate Rails code with high confidence because the conventions are rock solid.

Looking ahead, the Rails roadmap has some neat additions:

• Lexi: A Lexical-based editor coming to ActionText
• Native passkeys: Modern authentication without passwords
• Magic links: Login experiences that just work

These features show Rails isn't just reacting to AI—it keeps evolving as a framework that cares about developer experience and sensible defaults.

What This Means for You

Whether you're running a startup, leading a dev team, or building your first web project, DHH's pivot carries a clear message: the tools matter less than the outcomes. Hand-writing every line of code was never the goal—it was always about building software that actually serves real needs.

AI tools don't kill the craft of software development. They just change where that craft gets applied. The developers and teams who'll thrive are the ones who adapt their workflows, experiment boldly, and remember that code is a means to an end.

At NameOcean, we see this evolution up close. Our customers increasingly build projects using AI-assisted workflows, and the platforms they pick matter. Rails' stability mixed with its openness to new development paradigms makes it a solid choice for teams ready to embrace vibe coding without throwing best practices out the window.

The future of Rails isn't about choosing between craftsmanship and speed. It's about having both—and that's a vibe worth coding to.