This as-told-to essay is based on a conversation with Keith Ballinger, Google’s vice president and general manager for Developer & Experiences who is based in Seattle. His identity and employment have been verified by Business Insider. The following has been edited for length and clarity.
I’ve been a software engineer most of my life.
I joined Microsoft in the late ’90s, working on C# and .NET in those first few years, and then did a bunch of startups that were focused on developers. Later on, I came back to Microsoft through a startup that they acquired, Xamarin. My boss and I ran the deal that acquired GitHub, and then I was the senior vice president of engineering at GitHub.
I came to Google about three and a half years ago. I’m the vice president and general manager for Developer & Experiences so I focus on AI assistance for developers across the software development life cycle.
If you thought your job as a software engineer was about writing code in a specific language and framework, the world is definitely changing in front of you. I’ve always loved to code, but that’s not really what a software engineer’s job is. It’s to apply engineering techniques and use technology to solve problems. It’s about taking something messy and making it crisp.
That’s a skill that people can pick up, but now it’s more important than ever, and certainly more important than memorizing how an API works.
I love the era we’re in right now. I’m building faster and better than ever because I can now offload complex workflows to AI, making this the most productive and exciting period in my engineering history. As a developer it really feels like the only limits right now are my imagination and time.
If you’re a software engineer, there’s never been a greater time to be alive.
How AI has changed the creative process
I do normal things in my job, like various meetings, reviews, and paperwork. But I try to peel off time to ideate with Gemini. Sometimes that ends up being during evenings or weekends, and sometimes it’s during working hours.
I’ll talk to Gemini Enterprise about an idea I’ve had or a problem I’ve seen, and it becomes a sounding board. Developers talk about the “rubber duck technique,” where you talk to a rubber duck about your problem, and then get to the solution without having to ask anyone. When I talk to AI, it’s not necessarily about having a problem, but it’s about needing a place to write down my thoughts and get a reaction.
That generally leads to an idea, and I’ll write a document about it that may include a spec for how it works or details around its architecture. Then I go to Gemini CLI, which is the terminal where developers work and interact with their tooling. I’ll give Gemini CLI and the newly created Conductor extension what I’ve written, and also have it interrogate me, so that it gets an idea of what I’m trying to build.
The Gemini CLI Conductor extension helps me break down the problem, and then I let it go off on its own, and it does the tasks in order or in parallel. It lets me know once it’s done writing the piece of software or component of the UI, and then I’ll review the code or just the output itself, and play with it. If I don’t have blocking feedback, I let it keep going and it moves on to the next task. I steer it as we go along, and then at some point I have something done, and I’ll start using it.
As I use it, I’ll come up with changes and I’ll go back to the Gemini CLI to make those. Throughout this process, I’m partnering with AI to act as a product manager, but I’m also working as an engineer on the most important thing engineers can do: Technical architecture and how to break the problems down.
I experiment with tools 20+ hours a week
As an executive, the role I take with writing code is a bit different than what an individual contributor does. AI helps me dive into ideas for the future by prototyping very quickly. If those ideas make sense, helps me advise the team on solving similar problems in the future.
My team builds agentic developer tools like Gemini CLI, and I’m a firm believer that leaders need to use the software that their teams create. I send an email every Friday to my team that includes what I’ve worked on. That’s how I keep myself accountable — making sure I’m always talking about these things.
I don’t have children or a lot of obligations outside of work — and I love coding. So I spend a significant amount of time outside normal hours experimenting with tools. It’s usually around 20+ hours a week.
It’s not because I have to. It’s because I love it so much.
It’s better to embrace AI than to ignore it. As you get more senior, your ability to write is one of the most important skills, and you have to spend a lot of time doing it to get better. The better you are, the more you can leverage AI to increase your productivity and make what you’ve written come to life.
