Google developers discover AI elevates judgment over JavaScript

When you enter a prominent tech company's office, you might anticipate seeing countless software developers and engineers fixated on their screens, exhausted from endless coding.

However, what’s the scene today? AI now handles the majority of programming tasks. 

Senior engineers at Spotify haven't engaged in code writing since December, as revealed by co-CEO Gustav Söderström during last month's earnings call. Reports suggest Anthropic employs AI to produce 70-90% of its code. 

In October, Google's leadership noted that AI systems are generating half of all code.

The number is "substantially higher now," mentioned Ryan J. Salva, senior director of product management at Google.

"We're witnessing a profound transformation in the industry's approach to building software," Salva shared with Business Insider. 

Dora's 2025 study, a Google Cloud research initiative, surveyed 5,000 tech professionals globally and discovered that by September, 90% of those in software development were utilising AI at their jobs, a 14% rise from the previous year.

So what are software developers at the leading national tech companies focused on, with AI becoming a staple in coding?

According to Julian Togelius, a computer science and engineering professor at New York University, developers and engineers are shifting from programming and syntax towards design and supervision.

 They're embracing roles where judgment is prioritized over JavaScript, effectively redefining the role of a software development professional. This evolution comes with its own challenges and needs careful management.

Reflecting on the past five years, Salva recalls that a developer's expertise was once centered around programming languages such as Python or JavaScript. 

Their routine mainly involved working in a code editor and composing "if-then" statements.

This shift sees certain individuals thriving, particularly those experienced in managing people, according to Togelius. 

By overseeing numerous AI coding agents, they're using similar skills required for team management, such as rapidly switching contexts, or composing high-level guidelines and instructions for agents.

Coordinating multiple agents can give a developer a sense of immense capability, Togelius added. 

In fact, Dora's 2025 study found that 80% of software development professionals believe AI has heightened their productivity.

However, frequently checking in on various agents can also cause fatigue.

Salva alerted that excessive technological shifts in the industry can present human-related change management challenges. 

As focus transitions from coding, teams are now concentrating on adapting to evolving engineering practices. "We must ensure there's time and opportunities for engineers to adapt to the new approaches," he stated.