The Poet Without Verses: When AI Changes the Craft of Coding

A Poet with No Verse Saturday, Feb 21, 2026 5 minutes

It's been a while since my last post. Since then, I left my consulting gig and joined an early-stage startup. I wore more hats than I expected, got to lead and mentor people, and along the way entered my 30s (midlife-crisis / existential-dread era). Just when I felt like I had found my footing and mapped out my trajectory, AI showed up and reshaped the landscape.

Now nearly a decade into writing code, it feels like the right time to pause and reflect on what I've been feeling lately.

At my first job, fresh out of college, I was lucky to start with a seriously strong dev team. In there I quickly realised writing production code is very different from the showcase projects I did or the DSA problems I had to grind to crack the interview. I had to understand the business model and also internalise the software modelled around this. I'd get assigned a task and start from one tiny corner of the codebase, trying to make sense of how it all worked.

And as I read the code, I found patterns and abstractions of how the business was modelled in it. I couldn't help but feel awe at some parts of it. How did the author come up with this idea? What should I learn to be able to write code as such? The codebase - it moved me.

So I asked the author of it one day those questions and he said - "Code should be like poetry" along with other advice advocating to be the student of the game. He never did give any further explanation for that quote nor did I ask for the source of it. The quote in itself was a beautiful abstraction.

But in my head that idea bloomed such that, coders are like modern day technological poets. Where one could string in beautiful verses that can only be read and appreciated by a select few who had spent the time and effort to learn this craft. Unlike the poems in the traditional sense, here they can - not only move their readers but also - the electrons. It bound machines to your will! How cool is that I get to do this as a day job and be paid for?!

In my professional career, I was yet again fortunate to work with and be mentored by some really smart and inspiring folks. And I have a couple of quotes from each one of them that I keep in my head and still try to align myself with. But this one was my favourite. Mainly because I'd like to think I'm a creative person. Like I always have a need to do some sort of self expression as fix. If I don't, it bottles up and then I have to create something, like this post I'm writing for instance.

Coding for me is a form of self expression. And as I moved through my career, I worked with different codebases. And in each of them no matter the conventions and rules we keep for it - once you settle in, you start recognizing patterns. You can almost tell who wrote what just by moving through certain sections. The way someone names things. The abstractions they lean toward. The simplicity or cleverness they prefer. In the process you end up learning stuff, form your own opinions of it and be inspired.

One can't help but acknowledge the author's craft. Those feelings that get invoked in me while I read such code is what I strive to impart in others when I write code. "For those who come after."

With the advent of AI, I fundamentally feel the shift in the way we write code. I really do feel a lot more productive than I was earlier. I'm able to explore ideas quicker and ship faster. But I no longer get the sort of fulfilment I did before I started incorporating AI into my daily workflow.

Earlier, wrestling with a problem meant sitting with it and let a solution emerge as I kept thinking about it. Now I get solutions instantly and I find out what fits best and stick it in. It doesn't feel authentic and I'm skewed toward what solutions the AI provides. Even when I write code on the back of my mind it's like - would anyone even read this? It's going to be consumed and modified by some LLM.

I also empathise with the next generation of programmers working their way up. The business always expects you to ship faster, especially in startups where the pace is the biggest advantage. All the reading and internalising of code is delegated to the AI. Would they get inspired the way I was back in the day? Where would they get inspiration from?

From a business standpoint, I don't think any of this is a problem. Businesses have always leaned toward what's measurable - commits, output, velocity. If it can be quantified, it can be assessed. The soul of the craft only gets attention when money isn't the constraint. And money, whether we like it or not, is what keeps the machine running.

But when you look at the big shifts in our industry, the technologies that truly changed things - they were almost always driven by people who cared deeply. People who were obsessed. Passion has always been part of the story.

I'm not saying AI is bad because I'm a big time advocate for it at my current workplace. Often I'm amazed what I can do leveraging it. Maybe the uneasiness I feel is because my creative outlet is withering away. Maybe the real question now for people like me is: If code is poetry, how good is a poem that's never read? That never moves anyone? One which was never written by you.

Because these days, by that analogy, if I were a poet, I'd be one with no verses to write.

---

This reflection captures something profound about the current moment in software development. The author's journey from seeing code as poetry to feeling like a poet without verses mirrors a broader cultural shift that's happening beneath the surface of our industry.

The comparison to poetry is particularly apt. Traditional poetry requires craft, revision, and a deep understanding of form and meaning. A poem that's never read, never moves anyone, and was never truly written by the person claiming authorship - that's not poetry at all. It's just words on a page.

What makes this essay powerful is how it articulates the tension between productivity and craft. The author acknowledges the undeniable benefits of AI - faster shipping, quicker exploration, increased productivity. Yet there's an emptiness that comes with it. The wrestling with problems, the struggle to find elegant solutions, the joy of discovery - these are being replaced by instant answers and copy-paste solutions.

The concern for the next generation is particularly poignant. Where will new programmers find inspiration if they're never forced to sit with difficult problems? If they never read and internalize great codebases? If everything is abstracted away by AI assistants? The craft risks becoming a cargo cult where people follow patterns without understanding why they work.

There's also an interesting meta-point here about this very essay. The author mentions needing creative self-expression as a fix - if they don't create something, it bottles up. This post itself is an act of that self-expression, wrestling with these ideas in a way that feels authentic and thoughtful. Would they have written this if AI had generated the initial draft? Would it feel as meaningful?

The business perspective is pragmatic but perhaps short-sighted. Yes, businesses optimize for what's measurable. But the technologies that truly changed our industry - from operating systems to web browsers to programming languages - came from people who cared deeply about the craft. The passion, the obsession, the willingness to sit with hard problems for years - these aren't just nice-to-haves. They're often what separates transformative innovation from incremental improvement.

What's striking is that the author isn't anti-AI. They're an advocate for it at their workplace. This isn't a rejection of the technology but a mourning of what's being lost in the transition. It's the feeling of becoming less of a creator and more of an editor, less of an author and more of a curator.

The final question - "If code is poetry, how good is a poem that's never read? That never moves anyone? One which was never written by you" - is devastating in its simplicity. It suggests that perhaps the most valuable thing we're losing isn't productivity or even quality of code, but the very essence of what made programming meaningful in the first place.

We're left with a paradox: the tools that make us more productive might also be making our work less meaningful. The question isn't whether AI is good or bad, but whether we can find ways to preserve the craft, the inspiration, and the self-expression even as we embrace the productivity gains. Because if we can't, then we really will be poets with no verses to write.