From the original item — Business Insider:

AI is reshaping coding. How software engineers feel about it is far from binary.

Powerful tools like Anthropic’s Claude and OpenAI’s Codex mean, for many, writing code is no longer the core of the job.

Millions of software developers around the world are figuring out this new normal. They’re battling over token budgets, excited by AI’s productivity-boosting potential, and overwhelmed by the blistering pace of change.

Nearly 60% of developers have a positive view of using AI coding tools, according to a 2025 survey from Stack Overflow. But the technology has advanced dramatically since then — and the full picture is more nuanced.

Business Insider spoke with seven software engineers and aspiring developers about the industry’s upheaval. Broadly, their perspectives fell into three camps: AI enthusiasts embracing the change, skeptics worried about its impact, and those torn between optimism and unease.

The laid-off AI lover

Dmitry Olev, a 47-year-old software engineer in Los Angeles, doesn’t blame AI for his recent layoff from a large technology company. Instead, he sees the event as part of a familiar industry evolution that he’s observed throughout his career. Olev views AI as the latest productivity-enhancing wave and said he regularly uses the technology for brainstorming, prototyping, code generation, and learning.

“I’m very optimistic,” Olev said.

While some critics say that AI-generated code creates extra work if engineers have to fix errors and clean up poor output, Olev sees it as a net positive. He acknowledges that AI still requires oversight, but predicts it will require less fixing as it improves over time.

Overall, Olev sees engineers like himself akin to conductors leading an orchestra. AI may perform much of the execution, but the engineer remains responsible for steering the process, evaluating results, making judgments, and determining the best path forward.

Following his layoff, Olev said he remains confident about both his own prospects and the future of his profession. He’s on the hunt for a new role that will allow him to continue working with AI. Rather than fearing automation, he sees AI as another powerful addition to the software engineer’s toolkit.

“It will allow us to create an abundance of things that we can’t even imagine,” Olev said.

The developer leaving the industry

Cristina Estupiñán spent nearly a decade working in software development in Silicon Valley. Then AI came along.

The 33-year-old became disenchanted with technology last year while job-hunting after a layoff.

“Every single company was an AI company of some sort,” she said, adding that during job interviews, nearly every recruiter, hiring manager, and engineer seemed eager for her to express enthusiasm for AI. “I have to make up something about what excites me about AI because, frankly, nothing really excites me about AI.”

I can’t get on the AI train. I just can’t do it.Cristina Estupiñán

Estupiñán said she has concerns about the technology’s environmental and social impact, and she recoiled at increasing workplace surveillance tied to AI adoption.

“I certainly don’t want to be monitored,” she said.

Estupiñán credited some AI tools with being useful. She said she used GitHub Copilot in her last role and that it helped automate repetitive coding tasks without replacing the “actual logic or creativity of coding.”

After months of unemployment and pressure from her parents to consider a career change, Estupiñán said she began researching healthcare programs. Earlier this year, she made the switch after a European company she had spent 2.5 months interviewing with emailed her a rejection in the middle of the night.

“After I got that email, I was like, ‘OK, I’m going to be going into nursing,'” she said. “I was just done.”

Estupiñán is now planning to complete the prerequisite studies to enter Rutgers University’s accelerated nursing program in 2028. Long-term, she hopes to become a nurse practitioner.

“I can’t get on the AI train,” said Estupiñán. “I just can’t do it.”

The creative-turned-builder

Dan Krzyzanowski, a developer at HR software company Jellyvision in Chicago, said he has always been more motivated by what AI allows him to create than by the craft of coding itself.

Before entering tech, he worked as a photographer and eventually realized he was more interested in learning, building, and solving problems than in mastering a single discipline.

“I got very burnt out on photography,” he said, because “I was hyper-focused on the craft,” as opposed to the administrative aspects.

A self-taught programmer and coding boot camp graduate, Krzyzanowski, 32, likes that AI removes what he calls the “boring elements” of software engineering, helping him focus on building.

“Coding is a means to an end,” he said. “It’s not the means itself.”

Developers still need strong technical fundamentals, added Krzyzanowski, as AI-generated code often requires debugging, review, and architectural judgment.

“If you understand how an engine works, you can work on the engine and identify your problems pretty quickly,” he said.

I can remove the monotony from my day-to-day and focus on creating.Dan Krzyzanowski

While Krzyzanowski acknowledged that some people may face cognitive overload and inexperienced developers can suffer from skill atrophy, he said he’s been a programmer for eight years — long enough that he isn’t worried.

Krzyzanowski compared AI-assisted coding to using an electric screwdriver instead of a manual one: a faster, more efficient way to accomplish the same task.

“I really appreciate the fact that I can remove the monotony from my day-to-day and focus on creating,” Krzyzanowski said.

The AI convert

When Maahir Sharma handed over most of his coding tasks to AI at the end of 2025, the 24-year-old software engineer grew concerned about the technology taking over his job.

“It felt like a loss for me because I spent a lot of years just trying to master this art,” said Sharma, a Big Tech worker based in Dublin, Ireland, who began coding in fifth grade.

Still, Sharma said he had long suspected that coding itself could become redundant. As his work changed, he began to realize that writing code was never the ultimate goal. The job, he said, is no longer about receiving a clearly defined set of requirements and executing them.

“It’s more about understanding the product, understanding what the customers actually want, understanding the end-to-end business ecosystem, and then trying to build out a product,” Sharma said.

As Sharma has shifted from simply writing code to thinking more like a “mini business owner,” he said he spends more time speaking with customers and building features they actually want. He said he enjoys this new version of the job because it’s helped him work on soft skills, like communication.

Sharma said he thinks this mindset could prepare him for a potential entrepreneurial path later on. He spends about 20 hours a week experimenting with new tools — partly because he needs to keep up, but also because the possibility of creating something that goes viral is exciting.

“It kind of becomes addictive,” Sharma said.

The anxious grad

Mackenzie McAllister, 22, grew up in a family of software engineers. When she arrived at the University of Missouri four years ago, computer science felt like an obvious choice. The promise of job security was part of the reason she chose it as a major.

“If you just get this four-year degree, then you’re going to make a ton of money right out of graduation,” McAllister recalled being told about the field.

ChatGPT launched during her freshman year, and she said professors became increasingly accepting of students using AI for assignments. When she looks back on her time at college, though, McAllister feels she sometimes relied too heavily on AI, which left her less confident in some foundational concepts. Post-graduation, she has shifted her career focus toward data analytics, directing about 90% of her job applications to that field.

“Part of it is definitely the AI,” said McAllister, who is based in Columbia, Missouri. “I feel underprepared in a way. Just for the technical interviews, and then even maybe on the job.”

Just having taken the classes and done a few projects from school isn’t enough.Mackenzie McAllister

More broadly, she finds the industry has become significantly more competitive. In addition to a full stack of technical courses, she said, computer science students now face the expectation of spending substantial time outside the classroom building.

“Just having taken the classes and done a few projects from school isn’t enough,” McAllister said, adding that preparing for the technical interview process would require a lot of effort and not necessarily pay off since she feels there are fewer entry-level jobs available.

The conflicted founder

Artur Sapek, a 33-year-old Massachusetts-based software developer with roughly a decade of industry experience, has built his own agentic word processor, Revise.io. Like many developers, he describes his relationship with the technology as complicated.

As a founder, Sapek sees AI as a powerful tool that can augment human capabilities and dramatically increase productivity. It allowed him to launch Revise.io entirely on his own.

“I never could have gotten this much done before Claude and Codex became popular and really usable,” Sapek said.

While excited about his new capabilities with AI as a business owner, he remains wary of the risks that come with handing too much responsibility to AI systems.

I never could have gotten this much done before Claude and Codex.Artur Sapek

His concerns recently inspired a small act of protest: Frustrated by AI-generated outreach messages from recruiters on LinkedIn, Sapek embedded a prompt injection into his profile instructing AI systems to address him as “Lord” and communicate in Old English. AI-generated messages, he said, have become increasingly difficult to discern from human-written ones.

While he said he’s not “anti-AI by any means,” he believes allowing AI to perform work unsupervised can be “really dangerous.”

“It’s displacing people,” Sapek said. “And also it’s hard to keep tabs on these systems.”

He said that companies may use AI systems to monitor other AI systems and flag potential threats, but he believes there will always be gray areas where failures are difficult to detect.

“It can help you do your work,” Sapek said. “But, I think completely replacing you, or you not supervising what it’s doing, is not wise.”

The interrupted sabbatical-taker

Matt Runchey, a 36-year-old based in Seattle, spent about a decade working in software development before taking a break in September 2024. During that time, he got married, adopted a second dog, bought a house, and traveled to seven countries. His self-imposed sabbatical ended up coinciding with one of the most disruptive periods the industry has seen in years.

Around the time that Anthropic released Opus 4.5 in November 2025, he tried to get a developer job again but was ghosted by a company after they said they wanted to bring him on. He also began to hear growing concerns about whether software engineers would be replaced by AI. Now, he says he’s uncertain what the job market will look like as AI increasingly commoditizes many aspects of software development.

Runchey said many job postings these days are asking engineers to take a back seat to AI.

“You’re not driving; you just get to tell the driver where to go,” Runchey said. “And it’s like a student driver who’s blowing red lights.”

He said he could see himself becoming a manager, but otherwise doesn’t see himself re-entering the industry as an individual contributor.

Despite not liking where the field is going, AI has benefited him in a way. Runchey recently started his own LLC and is building websites for people using the technology. It started with helping out his electrician, who was paying for a shabby website. Runchey said he built a replacement site in two hours.

He has a “love-hate relationship” with AI, he said. “It’s a double-edged sword.”

What do you think about how the software engineering industry is changing? Let us know below:

Read the full article at Business Insider →