Even off the clock, Danny Hamam feels like he’s falling behind.
The software engineer, who lives in New York City, said every new AI tool release can trigger a fresh wave of anxiety.
“The first thought that I get isn’t that, ‘Oh, this is so exciting. Another AI tool dropped.’ It’s, ‘I’m behind. I have to learn this ASAP,'” Hamam said. “So you start freaking out.”
Hamam, like many tech workers, is confronting an ever-growing wave of AI tools that kick-started with the launch of ChatGPT in late 2022. Since then, Anthropic, Google, and OpenAI have been in an arms race to outdo each other with the latest frontier model.
Janice Chung for BI
The monthly cadence of major model releases has roughly quadrupled since 2023, said Peter Assentorp, a coder and designer based in Denmark who built a database to keep tabs on the expanding AI horsepower.
“The pace got high enough that I was losing track of what was newest and best, and I build and code with these models every day,” he said.
The number of major AI releases climbed from 18 in 2023 to 69 in 2025, according to Assentorp’s tally. By mid-2026, major AI companies had released another 30 models.
“The Great Coding Reset” is a multi-week series exploring how AI has sparked an existential transformation in software engineering:
The rapid pace of AI releases can make it feel futile to master any particular tool, said Jack Boudreau, CEO and cofounder of Habits, a fintech company focused on financial planning.
“It’s almost not worth it for you to become a subject matter expert, because wait one more week, and they’re going to simplify it for you,” he said.
AI advances have meant the potential for greater productivity and excitement, but also mean keeping up with a blistering pace of change, software engineers told Business Insider. For some, the information firehose has resulted in workplace paralysis and a growing sense that the tools they feel pressured to master are gaining the upper hand.
“I believe it’s only the beginning,” Sacha Greif, a developer in Japan, said of AI’s impact on coding. He predicts the AI giants will eventually make many stand-alone software products, such as project-management tools, unnecessary, leaving less room for the companies that make them. “I see a hollowing out of the industry.”
Coding is AI’s first big breakthrough into the world of work, putting the tens of millions of software developers around the world at the forefront of this transformation — and giving other white-collar workers a preview of the existential reckoning coming for everyone else.
Losing agency to the agents
The shadow of that future is already visible to those fearful that AI could replace their jobs. Greif runs the tech research publisher Devographics, which recently conducted a developer survey with roughly 7,000 respondents. In it, more than four in 10 said that AI tools threaten their job security.
Those fears relate in part to how quickly the work itself is changing. As AI takes on more coding tasks, developers are increasingly spending their time guiding and managing AI systems — and now, agents can do the prompting, too.
Janice Chung for BI
“We’re making the machine that makes the machines now,” said Annie Vella, a developer in New Zealand.
So far, though, those changes haven’t translated into a collapse in hiring. Software development job postings have recently been inching higher. Yet it’s not just the fear of replacement — the advances are forcing developers to rethink their relationship with the craft itself. Each wave of new models is becoming better at writing code, causing some to wrestle with what it means when tools can perform — quickly, if not perfectly — skills they spent years mastering.
Some developers also worry that AI could increasingly dictate how they work, undermining their skills and turning them into “service drones” for the technology, said Cary Cooper, a professor of organizational psychology and health at the University of Manchester’s business school.
Cal Newport, author and Georgetown University computer science professor, said “the deep work required to write code from scratch creates deeper satisfactions in the long term” than managing bots.
“Endlessly waiting for models to spit out code — what many now call ‘botsitting’ — is boring,” Newport said.
At the same time, workers are facing added pressure from employers to go all in on AI. Companies are using dashboards to track employees’ AI use, monitor how many tokens workers consume, and incorporate AI adoption into performance reviews.
Herminia Ibarra, a professor of organizational behavior at London Business School, said some of the pressure comes from organizations and managers overestimating how quickly engineers can adopt AI and what the technology can and cannot do — then judging workers against those expectations.
“The engineers are stuck because they are being asked to deliver innovation in business-as-usual mode,” she said.
Ben Eubanks, an analyst and researcher who studies HR technology and workforce trends, said the anxiety among some software engineers is so deep that he is hearing some are considering career pivots into sales or support roles.
‘How much more can I optimize?’
Not everyone is exhausted. Developer Rafa Rafael said AI means he spends less time troubleshooting problems or researching fixes and more time understanding requirements and thinking through features.
“I feel more involved in the overall product and not just the code itself,” said Rafael, who lives in the Philippines.
As a developer, Angga Pratama now largely oversees workflows and manages multiple AI tools simultaneously rather than writing code — a change that he said has increased the mental intensity of the job.
“The faster things become, the more the pressure shifts from ‘Can I finish this?’ to ‘How much more can I optimize?'” said Pratama, who is based in Indonesia.
For some, conversations about the technology are difficult to escape. At work, online, and in everyday life, AI chatter can feel constant.
Akshyae Singh, who studied cognitive science with a specialization in computing and AI at UCLA, has tried to channel some of his feelings of uncertainty into action through his San Francisco startup, Frame, which backs creators who make videos about AI and its impact on society.
Janice Chung for BI
Singh said he expects humans won’t be able to keep pace with AI systems. “It’s not biologically possible,” he said.
Though Rafael now enjoys riding the AI wave, it took some adjustment. He initially found AI more tiresome than traditional coding as he spent extra hours experimenting with prompts. Rafael said he now avoids coding after work, choosing instead to spend time with his family, watch TV, or relax. He has also stopped trying to keep up with every new AI release.
“There is always something new, so I only look into it when I think it could actually help with my work,” Rafael said.
The challenge is as much organizational as it is individual. Many engineers are grappling with the double anxiety of keeping up with AI while also worrying about what it could mean for their careers, said Kathy Gersch, CEO of the change-management firm Kotter.
To help “quiet some of the noise,” she said, companies should encourage workers to share what they’re learning with one another. That would help workers feel like they’re “moving with the tide versus being hit by the tide,” she said.
For all the anxiety AI can generate, Hamam, the software engineer in New York, said the upside is seeing how much he’s learned and accomplished in the past few years.
“You’re building things that you thought you wouldn’t have built if it wasn’t for the added pressure,” he said.
