For Adam Janes, the weird thing about being a software engineer these days is that he doesn’t write much code anymore.
Instead, they spend most of their time drafting specifications in English that describe what the software is supposed to do. AI kicks out the code. He reviews it and improves it.
Janes is not alone.
At Spotify, the company’s top developers haven’t drafted “a single line of code” since December, co-CEO Gustav Söderström said on an earnings call in February. Instead, engineers often oversee the AI as it generates code.
“This is a big change. It’s real and it’s happening rapidly,” Söderstrom said.
The speed at which AI is reshaping software development is being measured in months, not years, with a mixture of excitement and trepidation. Technology company Brock laid off more than 40% of its workforce last week, saying the job cuts reflected AI-driven efficiencies.
For some software developers, further calculations are required. What does it actually mean to be an engineer once the core act of coding (long a source of status and identity) is automated?
Evictions are not the only threat. It’s a complete recalibration of identity.
“A craftsman’s touch”
Janes, an Australia-based fractional CTO, takes pride in the skills he has spent a decade learning how to write “elegant code” that is simple and easy for others to read.
“It really has an artisan feel,” he told Business Insider.
Janes said he is now spending more time creating robust specifications for the AI tools that build the software. That way, you can focus on solving the problem.
“There’s still a lot of engineering know-how and a sense of making things, but it’s moving into a completely different realm,” he said.
For some engineers, it doesn’t feel the same.
When Annie Vera was six years old, her parents got tired of giving her puzzles and bought her a computer with basic code in the manual. She said she was curious, so she started typing the commands.
“I learned at an early age the joy of turning these words into something moving on a screen, albeit in a rudimentary manner,” Vera told Business Insider.
The early thrill of creating something from nothing stayed with her. as A hands-on developer, she enjoyed solving technical puzzles that others couldn’t.
Now that AI systems are becoming more capable, Vera feels conflicted. On the one hand, tools help her act faster, solve more complex problems and improve quality. Additionally, as a technologist, she finds the capabilities of AI amazing.
On the other hand, if she were still in the early stages of her career, Bella said, “I would have been very upset about what was going on.”
Bella, who lives in New Zealand, recently completed her master’s thesis investigating the impact of AI on her profession. There’s a “magical feeling” to seeing an AI spin up large amounts of code at once, she says. Still, Berra says the output doesn’t give you the same feeling of staying up all night fighting a stubborn bug and finally getting it working.
“The satisfaction comes from friction that doesn’t really exist anymore,” she said.
What friction brings
Austin psychologist Mike Brooks said the struggle to overcome challenges is often at the heart of whether people find meaning in their work.
“We evolved to struggle to survive,” he told Business Insider. “We have to try.”
Brooks writes about the loss of purpose engineers feel when AI takes over tasks that once required effort and skill. He said that working hard for something is part of what makes it worthwhile. Scarcity, not abundance, makes something valuable.
“If you haven’t eaten anything in five days and you get a banana, it’s probably the best meal you’ve ever had,” Brooks said.
If AI can almost instantaneously generate the hours of focus that was once required, the psychological rewards could be diminished, he said. For engineers who have built their identities around solving difficult problems, the change can feel unsettling.
Jorge Melegati sees that tension in his research. A software developer and assistant professor at the University of Porto in Portugal, he studies how generative AI reshapes developers’ self-awareness.
Many people start coding because they want to build something, he says. As engineering roles evolve to overseeing AI agents, developers’ jobs could become “much simpler,” but also less satisfying, Melegaty said.
Previous research has shown that aspiring engineers often view the testing role as less prestigious and less difficult than building software from scratch, he said.
“It’s considered an easy job. It’s not that difficult and it’s not that rewarding,” Melegaty said.
But that doesn’t mean the job is no longer in demand. The U.S. government predicts that employment for “software developers, quality assurance analysts, and testers” will increase by 15% from 2024 to 2034. This is much faster than the average for all employment.
shift in focus
As AI capabilities grow, many engineers are still negotiating what their roles will be, including Keenan Block, a Massachusetts software developer who spent years honing his coding skills and learning languages like Java and Ruby on Rails.
“I focused on this one thing, and now it doesn’t matter anymore,” he told Business Insider.
So he had to shift. Block said he enjoys being able to focus on the problems businesses face and how software can help, rather than “arm-wrestling with computers.”
“Now I can read between the lines,” Block said. “When someone says, “I want a faster horse,” you think, “Oh, maybe I want a car.”
Do you have a story to share about how AI is impacting your career? Contact this reporter at: tparadis@businessinsider.com.
