The Grief When AI Writes Most of the Code

Gergely Orosz, author of The Pragmatic Engineer, is publicly processing something many developers are feeling privately: the loss of coding as a craft identity.

The Core Insight

“I’m coming to terms with the high probability that AI will write most of my code which I ship to prod, going forward.”

This isn’t a prediction—it’s a confession. AI already writes code faster than Orosz can type it, with similar quality for familiar languages and better quality for unfamiliar ones. The competitive advantage of being “good at coding” is evaporating.

The psychological weight is significant: “It took a lot of effort to get good at coding.” Years of learning C in university, getting lost in complex codebases on first jobs, studying books and blogs, absorbing wisdom from senior developers. Decades of deliberate practice to develop a skill that was both valuable and validatable—you could prove your worth by shipping code that worked.

Now? The skill still matters, but the act of typing code—being “locked in” balancing multiple ideas, getting into flow state, compiling and seeing it work—that’s increasingly something you supervise rather than do.

Why This Matters

This is grief. Real grief, for something valuable being taken away suddenly.

The Kübler-Ross stages are all there in developer communities: denial (“AI can’t handle complex systems”), anger (“they’re coming for our jobs”), bargaining (“AI is just a tool”), depression (“my craft is dying”), and eventually acceptance (“I need to adapt”).

Orosz is at acceptance, but he’s not pretending it doesn’t hurt:

“Some of my best memories of building software are about coding… It’s been a love-hate relationship, to be fair, based on the amount of focus needed to write complex code.”

The love-hate is key. Coding is hard. That’s why it felt valuable. Take away the hard part, and what remains?

Key Takeaways

• The shift is real: AI already writes most production code for some developers
• Speed + quality: AI is faster and equally good (or better for unfamiliar stacks)
• Identity crisis: Coding skill was validatable identity; what replaces it?
• Flow state migration: The “zone” may shift to higher-level problem decomposition
• It’s okay to grieve: Something is being lost, even if something new is gained

Looking Ahead

Orosz wonders if the satisfaction will migrate: “Perhaps with AI agents, being ‘in the zone’ will shift to thinking about higher-level problems, while instructing more complex code to be written?”

Maybe. But the original satisfaction came from the combination of mental challenge AND physical creation—thinking through complexity while your fingers expressed it in code. Instructing agents is a different activity entirely.

Some developers will thrive in the new paradigm. They’ll become “code architects” or “AI wranglers” or whatever label emerges. Others will mourn what coding used to be—the craft of translation between human intent and machine execution.

Both responses are valid. The industry is changing faster than our emotional processing can keep up. Acknowledging the grief is healthier than pretending the transformation is pure upside.

Based on analysis of “The grief when AI writes most of the code” by Gergely Orosz, The Pragmatic Engineer