How Google’s Dev Tools Manager Makes AI Coding Work

Google’s senior director of product management for developer tools, Ryan Salva, is at the forefront of how AI is transforming programming. With experience at GitHub and Microsoft, Salva now oversees tools like Gemini CLI and Gemini Code Assist, helping developers adopt “agentic programming,” where AI actively contributes to code development.

In a recent interview, Salva discussed a new survey highlighting developer adoption of AI tools. Surprisingly, the median start date for developers using AI tools was April 2024, coinciding with the release of Claude 3 and Gemini 2.5 — early milestones in reasoning-focused models with improved tool-calling capabilities. These tools allow AI to leverage external information, run tests, and self-correct, making coding more efficient.

On a personal level, Salva primarily uses AI for hobby projects. His workflow is terminal-based, combining Gemini CLI, Claude Code, Codex, and multiple IDEs like Zed, VS Code, Cursor, and Windsurf. AI assists him in writing specification documents, translating under-specified GitHub issues into detailed Markdown requirements, and then generating the corresponding code. AI even tracks progress, creating commits and pull requests automatically. Salva estimates 70–80% of his coding work involves guiding AI in natural language, while IDEs mainly serve as reading platforms.

Looking ahead, Salva predicts a shift away from traditional IDEs. Developers may spend less time writing raw code and more time designing solutions, defining requirements, and managing workflows. The role of a developer could evolve to resemble an architect — focusing on breaking down complex problems and overseeing AI-assisted implementation.

With tools like Gemini CLI, Google is enabling a future where AI handles much of the repetitive coding, letting developers concentrate on higher-level problem-solving. For those embracing AI-powered development, the terminal may become the new workspace, and the IDE a reference tool rather than the hub of creation.

FAQ Section

Q1: What is Gemini CLI?
A: Gemini CLI is Google’s AI-powered coding tool that helps developers write, test, and troubleshoot code using natural language commands, integrating seamlessly with existing IDEs.

Q2: How do AI tools like Gemini CLI change software development?
A: They allow developers to focus on high-level design and requirements, while AI handles repetitive coding, testing, and documentation, effectively turning developers into architects of software solutions.

Q3: When did developers start widely using AI coding tools?
A: According to research, the median adoption date was April 2024, coinciding with the release of Claude 3 and Gemini 2.5.

Q4: Can AI replace traditional IDEs?
A: While AI tools reduce the time spent coding manually, IDEs will still serve as platforms for reviewing and debugging code. Over time, the focus may shift from writing code to defining requirements and workflows.

Q5: What programming workflow does Ryan Salva follow using AI?
A: Salva primarily works in the terminal with natural language commands, guiding Gemini CLI to generate code and update requirements. IDEs are mostly used to read and review the generated code.

Topics: AI, AI coding, Gemini CLI, Google, developer tools