As AI coding tools rapidly evolve, Google’s project manager for developer tools, Ryan Salva, has a unique vantage point on how they’re transforming software development. Salva, who previously worked at GitHub and Microsoft, now oversees tools like Gemini CLI and Gemini Code Assist—and he believes they’re redefining the way developers work.

In a recent interview, Salva discussed Google’s latest research on AI adoption in software engineering and how developers are using AI coding tools to streamline workflows and rethink their roles.

AI Adoption Accelerates with Reasoning Models

According to Google’s survey, the median date when developers started using AI coding tools was April 2024, coinciding with the release of Claude 3 and Gemini 2.5.

“This was really the dawn of the reasoning or thinking models,” Salva explained. “Tool-calling became a game-changer—allowing models to compile code, run unit tests, and even self-correct.”

How AI Coding Tools Fit into Developer Workflows

For Salva, most coding today involves command line tools like Gemini CLI, mixed with platforms such as VS Code, Cursor, Zed, and Windsurf. His professional workflow often begins with a simple GitHub issue, which Gemini CLI turns into a 100-line requirements document in Markdown.

“Seventy to eighty percent of my work is just using natural language in the terminal,” Salva said. “Gemini writes the code, troubleshoots issues, and even updates the requirements doc as it progresses—each step creating commits and pull requests automatically.”

The Future Role of Developers

When asked about the future of programming, Salva predicted a shift from writing raw code to architecting solutions.

“Developers will focus more on breaking down complex problems and defining outcomes,” he said. “The coding itself may increasingly be handled by AI, while humans guide the bigger picture.”

Over time, Salva expects IDE usage to shrink as natural language interfaces and AI-driven coding become the standard.

AI Coding Tools Are Changing Everything

With Google investing heavily in Gemini AI for development workflows, Salva believes the future developer experience will look very different: less time wrestling with syntax, more time solving real-world problems.

“We’re heading toward a world where coding feels more like collaborating with a partner than writing every line yourself,” he said.