The Joy Is in the Struggle, and AI Is Taking Some of It Away

As AI coding tools move from novelty to daily routine, software engineers are weighing speed against the loss of the hard-won skills and quiet pleasure of debugging.
The Joy Is in the Struggle, and AI Is Taking Some of It Away — Tech Reader
Tech Reader  ·  Software & Craft
Feature  ·  November 20, 2026

The Joy Is in the Struggle, and AI Is Taking Some of It Away

Engineers say AI is making their work faster, but it is also draining some of the satisfaction that drew them to coding in the first place.
As AI coding tools move from novelty to daily routine, software engineers are weighing speed against the loss of the hard-won skills and quiet pleasure of debugging.

When Gemma Rodgers started her computer science degree in 2002, software engineering was considered a safe and lucrative bet. It drew people who liked getting into the weeds of code, chasing down bugs and building programs from nothing. Two decades later, the industry still moves fast, but artificial intelligence has changed the tempo. As CNBC reports, a 2026 Stack Overflow survey found that 83% of software developers use at least one AI tool, and Rodgers says nearly every engineer she knows now relies on them daily.

The Vanishing Struggle

For many developers, the appeal of the job was the grind itself. Rodgers describes the satisfaction of piecing together different parts of a codebase to crack a stubborn bug, sometimes after days or weeks of effort. AI tools can now shortcut much of that process. “Now you can sometimes do it in a few prompts, and it’s great,” she told CNBC, “but it’s a breadcrumb of joy, not a full meal of joy.” Engineers on social media have also described missing the long, uninterrupted flow states they once got from working through code for hours.

Still, the tools have opened the field. As Rodgers notes, people who might not have thrived in software a decade ago now can. The question is what the profession gives up in exchange.

“The sense of joy and fun that I used to experience ... it’s been removed somewhat ... It’s a breadcrumb of joy, not a full meal of joy.” — Gemma Rodgers, software developer

From Writing Code to Judging It

The job itself is shifting from producing code to evaluating it. James Broadhead, chief technology officer of the AI decision-making company Taktile, says engineers who once spent three weeks on a prototype can now generate ten in an afternoon. The bottleneck has moved. “The hardest part of that,” he says, is “having the judgment to decide which of those are any good.”

That shift carries a cost. AI agents produce so much code that engineers often never read all of it, and the codebase changes so fast that the mental map developers used to carry is harder to keep. Engineering has become less like building a house by hand and more like inspecting a fleet of prefabricated ones, each arriving before you have finished the blueprint.

A Widening Gap

Joshua Wöhle, CEO of the London AI training firm Mindstone, sees a growing divide between workers who use AI well and those who do not. The industry once talked about “10X engineers,” who were far more productive than their peers. AI, he says, is stretching that spread toward “100X engineers.” Companies, he argues, will have to decide whether to help the rest catch up or double down on the people already pulling ahead.

Wöhle also points to a harder truth. The people who excelled before were often those who loved writing and working through code, and they are not always the strongest performers now. “That’s the grief that is very, very real,” he says. “The thing you love doing is no longer the thing the company should be paying for, to a degree.”

“That’s the grief that is very, very real.” — Joshua Wöhle, Mindstone

The Case for Supervising the Machine

Not everyone is mourning. Chris Sees, a software engineer with more than 20 years of experience and CEO of the London virtual office company Hoxton Mix, treats AI as a capable but inexperienced junior colleague that needs supervision. In his view, the technology frees developers to spend more time on interesting problems. “We’d be silly to think AI gets everything right,” he says, adding that he takes real satisfaction in spotting where it has gone wrong.

His worry is the next generation. Routine coding and debugging are, as he puts it, “where you learn lessons and build scar tissue.” If AI absorbs those tasks, junior engineers may skip the lessons that make senior engineers good. Sees argues that companies must make AI part of the learning process rather than a shortcut around it.

Rodgers, who works for herself, has found a middle path. She hands the tedious parts of a project to AI and keeps the knotty problems for herself. Her broader point is about friction. Working through difficulty, she says, is part of what makes accomplishment feel earned. “Unless we are able to work that through in so many different aspects of life,” she says, “then you never get the full force of the good feeling on the other side.”

Aaron Rose is the founder and publisher of Tech Reader Magazine, covering system architecture, cloud platforms, and AI policy.