I started reading Gene Kim and Steve Yege’s new book on vibe coding and found myself asking: Do I aspire to be a developer?
The answer is no. But that’s not the point.
What I do aspire to is solving real problems. And that’s exactly why I’m convinced that experienced developers—those willing to embrace AI-augmented workflows—will be more valuable tomorrow than they are today.
Development as a Skill, Not a Career
The rhetoric around AI and coding tends to swing between two extremes: either these tools will eliminate developers entirely, or they’ll just make developers faster at what they already do. Both miss something important.
Development isn’t a career anymore. It’s a skill. And skills, unlike careers, can be applied to almost anything worth building—whether that’s a product, a system, an internal tool, or a business process.
For those of us who don’t aspire to be “developers” but who need to build things, this is liberating. For experienced developers, it’s an opportunity to become something more valuable: architects of intent and quality.
Why Experience Matters More, Not Less
Here’s the counterintuitive part: vibe coding and LLM-assisted development will increase demand for experienced developers, not decrease it.
The engineers I work with are some of the most capable people I know. They’ve spent years learning not just how to write code, but how to think about systems—tradeoffs, failure modes, maintainability, scale. They have hands-on experience, not just pattern-matched training data.
This matters now more than ever.
When you can generate code quickly—when you’re shipping iterations in days instead of weeks—the bottleneck shifts. It’s no longer “how do we write the code?” It’s “is this code going to hold up? Does this system make sense? What are we missing?”
Those questions require judgment. They require experience. They require the kind of expertise that comes from having actually built things and watched them break.
The real risk with vibe coding isn’t that it generates bad code—it’s that it generates plausible code that fails in subtle, expensive ways. Only experienced developers can reliably catch that.
The Work Gets Better
There’s another thing happening that matters: the work itself is becoming more enjoyable.
When you’re not wrestling with boilerplate, debugging syntax errors, or reinventing wheels, you’re freed up to actually think. You ship ideas faster. You iterate. You collaborate with non-developers who can now articulate what they need without waiting months for implementation.
That’s not a replacement for what developers do—it’s an amplification of it.
The Messy Part
I’m not going to pretend this is a smooth transition. It won’t be. There will be companies that treat AI-assisted development as “developers are now optional.” Some will fail because they don’t have the judgment layer anymore. Others will succeed despite it, for a while.
The friction point isn’t technical—it’s human. People need purpose. When technology changes, purpose changes too. That’s uncomfortable, and it’s been true for every major innovation we’ve had.
The best we can do is what we’ve always done: remember that we’re building for human beings, by human beings. Care for the people who’ve gotten us here. Build thoughtfully. And recognize that those who embrace new tools while preserving judgment and experience aren’t being left behind—they’re being asked to lead.
What’s Next
I’m early in vibe coding. I suspect it won’t be the final form of how we’ll work with AI and code. But I’m convinced that the developers willing to learn it now—to see it as a skill to layer onto their existing expertise, not a replacement for it—will be the ones who shape what comes next.
That’s worth paying attention to.