I recently read Existential Dread and the End of Programming by David Whitney. It didn’t read like hype but read like a quiet lament and I felt that.

There’s this strange mix at the moment — as Whitney says – excitement and exhaustion. On one hand, the tools are incredible. On the other, the ground is shifting under a role many of us have spent decades building our identity around.

Maybe This Is the Revolution

Years ago I remember hearing that software engineering was overdue a revolution. For all the cloud, DevOps, containers and distributed systems, the core loop hasn’t changed much: Write code, compile, test, deploy, maintain….repeat.. But maybe this is the real shift where we don’t just have better tools — but a different level of abstraction entirely. And if that’s true, then the role changes too.

Programming and Composing

I’m also a composer. (shameless plug incoming….) I’ve written for orchestras, ensembles, solo instruments. Spotify. And honestly, composing and programming feel very similar to me. In both cases you’re writing structured instructions to be executed by something else — musicians or machines — to produce an outcome. There’s deep technical craft in both. But when I compose, although the process of creating a well written score requires skill, experience and craft – ending with something quite beautify, the notation isn’t why I did it. I love the music.

The technical layer is in service of something expressive. So maybe that’s the real question for us: Have we fallen in love with the notation — the syntax, the cleverness — rather than the outcome?

If AI starts handling more of the notation, maybe what’s left is what mattered all along: shaping systems that actually deliver value.

“Product Engineer” Isn’t New

David talks an alternate name for the future full stack engineer – somethinh like a “product engineer”. Part of me thinks: isn’t that what we were always supposed to be? Understanding users. Owning outcomes. Connecting architecture to business value. Maybe AI just exposes the gap between people who write code and people who build systems that matter.

The Maintenance Reality

Greenfield is fun. Maintenance is expensive. I agree with Whitney on that. We’re being told agents will diagnose, refactor and maintain systems autonomously. But at enterprise scale? In regulated environments? With messy, long-lived systems? I don’t think that’s proven.

I suspect it won’t be big enterprises that demonstrate the future. It’ll be new, agent-native companies built from day one assuming AI as a core operating primitive. They’ll show us what actually works because there might be nothing to compare it to.

Building in the Middle of It

I’m building a startup product right now, and we’re leaning into this thinking. Can we architect our system as interacting agentic components? Can those agents work directly with our customers? Can we create a new level of value on both sides?

It’s exciting. And yes, AI is helping us move faster.

But I’m also tired. As Whitney says “I’ve never been more productive. I’ve never been more exhausted.” I hear you dude.

The pace is relentless. Even with multiple agents running, the cognitive load doesn’t disappear — it fragments. The constant steering, reviewing, validating… it’s a lot.

Yet the narrative being sold is pure acceleration. Sometimes it feels like a PR campaign aimed at management by a car salesman just because the test drive went well. We haven’t done 200,000 miles yet.

Halfway Through My Career

I’m about halfway through my career and I probably assumed the second half would look broadly like the first — just with better tooling. This was probably naive. What I’m most concerned about isn’t job loss. It’s skill erosion.

If we keep abstracting upwards, who understands what’s happening underneath? When genuinely new classes of failure appear — and they will — someone will need to go under the hood like a mechanic under a car, not just a robot? There are always unknown unknowns when practice shifts this dramatically and my worry isn’t that AI can’t reason it’s that we might forget how to.

The Music Parallel

Recorded music didn’t make musicians obsolete.

Production software didn’t eliminate composers.

Streaming didn’t remove the need for taste.

AI music hasn’t replaced human intention.

New roles emerged. The human element didn’t disappear — it shifted and I suspect software will be similar. Typing code may become less central but creating useful, valuable, effective systems for real people? That won’t go away.

Hope

Despite the existential wobble, I’m optimistic.

In Vibe Coding, Gene Kim and Steve Yegge describe FAAFO — Fast, Ambitious, Autonomous, Fun, Optionality. That feels like the right mindset.

Experiment. Move quickly. Keep options open. Don’t cling too tightly to old forms. Maybe this isn’t the end of programming. Maybe it’s just another abstraction layer forming. And like every abstraction before it, it will empower us — and hide complexity at the same time.

The real question is whether we stay intentional as it happens.