From Coder to Conductor: The Evolving Role of Developers in the Age of AI

We’re in a season of acceleration.

I recently joined Mitch Ashley, VP of Research at Futurum, and Erik Bethke, CTO of The Futurum Group, for a DevOps Unbound panel on how AI is reshaping what it means to be a developer. Not in some future state—right now.

If you’ve followed my Zero to Builder series, you know I’ve been relearning software development, this time with AI as my co-pilot. What once took hundreds of lines of GW-BASIC and late nights is now possible in minutes using LLMs like GPT-4 or tools like Cursor. But this transformation isn’t about speed. It’s about redefinition.

In fact, I recently published a deeper look at this shift over on DevOps.com: “Scaling Vibe-Coding in Enterprise IT: A CTO’s Guide to Navigating Architectural Complexity, Product Management, and Governance.” That piece breaks down what happens when hobbyist-level experiments start to become enterprise-critical systems—and how CTOs can’t afford to be caught flat-footed.

The Developer Role Is Changing—Even If Your Title Isn’t

The role of the developer is shifting—from implementer to orchestrator. Erik framed it well during the panel: “The scarcest resource in the world is knowing what you want to build.”

AI can help write code. But it still can’t decide why it matters.

That means today’s developers—whether they sit in a product team or on a business analyst desk—need to evolve into systems thinkers. Personally, I don’t write a line of code anymore until I’ve run my design through a series of LLMs: one interviews me, another critiques the architecture, a third suggests how to test. It’s not overkill—it’s velocity with guardrails.

Chaos, Vibes, and Peer Programming: Why Words Matter

We spent part of the panel unpacking “vibe coding,” a term made popular by Andrej Karpathy. But Erik made a key point: not all AI-assisted dev is the same.

He offered a sharper breakdown:

• Chaos Programming: No prior experience. No structure. Just pure experimentation with AI.
• Vibe Coding: Prototyping for speed. More informed, but still loose.
• Peer Programming: A genuine back-and-forth with AI, grounded in solid engineering principles.

Each mode has a place. But they require different guardrails, different expectations. And for leaders? Your job isn’t to outlaw vibe coding—it’s to cultivate the context in which it produces value without inviting risk.

This Isn’t Just About Devs—It’s About Process

As I said in my DevOps.com research piece, AI is accelerating developer workflows faster than enterprise governance can keep up. And that’s where the real danger lies.

It’s not that AI writes bad code. It’s that organizations are shipping AI-assisted code into production while still testing it like it’s 2015. The mismatch isn’t technical—it’s process and policy.

So What Now?

If you’re a senior engineer, your job isn’t going away. But it is evolving.

You’re now expected to do more than code—you’re expected to shape outcomes. To design for maintainability. To ask better questions before you ever write a test.

And if you’re earlier in your career? The path to mid-level is changing too. The old “learn by doing” model is becoming “learn by prompting.” The key isn’t knowing how to write the boilerplate—it’s knowing how to steer the system.

What’s Next: StarCraft, Not VS Code

Erik painted a powerful picture: imagine your dev interface looking more like StarCraft than Visual Studio. Dozens of intelligent agents, each handling parts of the stack. You, the conductor, orchestrating the action, directing strategy, optimizing systems.

That’s not sci-fi. That’s the trajectory we’re on.

We’re no longer just developers. We’re orchestrators.

Beyond Orchestration: The Real Why

But there’s a deeper shift underneath all of this—one that goes beyond engineering practices.

At the end of the panel, Erik offered a perspective that’s stayed with me. He said:

“I hope AI does eliminate work… I hope all the mundane, hazardous, repetitive, uncreative work on the planet gets done by AI. What I hope it does is free up humans to build, and dream, and play.”

That’s the “why” that gets lost in most enterprise AI discussions. We’re not just trying to increase productivity or reduce backlog. We’re aiming for a world where humans get to spend more of their lives doing meaningful, creative work—whether that’s coding, mentoring, designing, or imagining entirely new things.

AI will raise the baseline. It’ll automate the tasks we never liked doing. And it’ll challenge us to level up—to move from doing the work to deciding what work is worth doing.

Final Thought

The real unlock isn’t the AI.

It’s the human on the other side of the prompt—someone who knows what outcome they want to create and can bring systems, teams, and technology together to deliver it.

The conductor’s role isn’t just to lead the orchestra. It’s to choose the music.

And for the first time in a long time, we might all have the freedom to compose.