Josh Woodward on Google's AI Revolution
When Jewel and I began coaching Josh Woodward through the Google School for Leaders’ Hub program, he was not yet the VP in charge of the Gemini AI app. What we remember most, aside from his scintillating intelligence, was his infectious enthusiasm and laughter.
Josh embraced our jazz model with open arms and enjoyed the musical examples we shared, from big bands led by Roy Hargrove (“September in the Rain”) and Wynton Marsalis and the Jazz at Lincoln Center Orchestra (“2/3’s Adventure”) to smaller ensembles led by Betty Carter (“Tight”) and George Benson with Herbie Hancock, Ron Carter, Johnny Pacheco, and Billy Cobham on Charlie Parker’s “Billie Bounce.”
Considering our previous posts on AI here and here, and with the very successful launch of Gemini 3.0 on November 18th, we appreciate Josh taking the time for this interview.
Greg Thomas: When we worked together two years ago through Google's School for Leaders, you were VP of Google Labs. Since then, you've added Gemini to your leadership portfolio. Congrats! The headcount of your organization has therefore grown exponentially. How have you handled that transition?
Josh Woodward: It's been fun and full-on. In my first week, I prioritized learning, and I'm still doing that. I set aside dedicated time to learn from the team, from partners, from people that love our app, and people who stopped using our app. These conversations keep me grounded in the biggest opportunities and challenges.
Greg: At Google Labs, you begin with small teams of five to seven people, which can be nimble like an excellent jazz band. Why do you start so small, and how do you measure success?
Josh: We forget that every big product started small, with a few motivated people and a dream. For a large organization like Google, with a portfolio of successful products and businesses, doing small things can be hard. That's one reason we brought back Labs, so we could have a place that's tuned to spot and build small things of high potential, and then nurture them. Success for Labs can take many forms, but the two most common paths are when one of our ideas graduates into an existing Google product or when it becomes a standalone product that's so valuable that people will pay for it, such as NotebookLM, Flow, or AI Studio.
Greg: I hear you. In jazz, one of the most exciting practices is uptempo swing, which demonstrates virtuosity at high velocity. One of the core cultural objectives of Google Labs is velocity, with sprints that move from an idea to a working product in users' hands within 50 to 100 days. Why such speed?
Josh: When you're doing early-stage work, the biggest risk is that you're making something no one wants. The longer you wait to find that out, the worse off you're going to be. So we prioritize getting outside the building, going straight to people, and co-creating with them in the first 50 to 100 days. This type of action always produces information, and we often see our favorite ideas off the mark early on. The prototypes will be misunderstood; features will land with a thud; core assumptions will prove wrong. That's okay. It's the unavoidable cost of doing new things, but the best recipe for success is to iterate and learn as fast as you can.
Greg: When you took our leadership workstyle assessment, your results placed you in the Integrator zone as a generalist. In the age of AI, do you think that generalists who are entrepreneurial will have a leg up in the marketplace?
Josh: It seems like generalists might have a comeback of sorts. So far, it seems like one of the biggest effects of AI is that it helps lower the barriers for anyone to learn, build, and create. So if you're a generalist with an idea and some elbow grease, it's never been easier to make something new and give it a shot.
Greg: You're on record saying that soon, the AI text interface of written prompts will be replaced by more contextual visual and voice interactions. How will this change the way users interact with LLMs?
Josh: When you're using these models every day with your camera on or by talking to them, you realize that text input will still exist, but its prominence will diminish over time. For many people, typing is slower. For many global languages, typing is harder. And for a lot of people in the world, visuals or voice are a more accessible, universal language.
Greg: In your keynote at Google I/O in May, you said that a core goal is to make Gemini the most personal, proactive, and powerful AI assistant. Tell us why.
Josh: For a lot of us, we're trying to live in the future with this technology. We see potential everywhere. We envision a world where people will have access to these superpowers across multiple devices and can interact with it multiple times per day - sometimes for seconds, other times for extended periods. We're motivated to bring this vision to life because we think it will increase productivity and creativity 10- to 100-fold in society. We also think it can have profound personal implications, by helping you understand your goals and work on them in the background to empower you.
Greg: The last part of your answer points to the agentic potential of LLMs. Before my next question, allow me to congratulate you on the successful launch of Gemini 3!
Josh: Thanks, Greg!
Greg: Tell us more about how Gemini and other products in the Google ecosystem are striving to enable AI agents to serve humans.
Josh: Gemini 3 comes with our most reliable, multi-step tool use capabilities to date. It does a nice job following instructions and combines features like browsing and coding, along with pausing for your input before proceeding at critical steps. Our goal is to make complex tasks significantly easier and faster to complete, all while keeping you informed and in charge.
Greg: A few questions ago, you pointed out how using visuals and voice with AI will become more ubiquitous. Does Gemini 3 advance this direction?
Josh: Yes. Gemini 3 can create interactive widgets and websites on the fly, which teachers love for adaptive learning in particular. We think these types of "generative interfaces" are the start of something new and exciting because the replies should become more personalized, visual, and interactive. On voice, we've also made major improvements recently. Gemini Live is now smarter, more expressive, and has adjustable speeds, which was one of my favorite updates.
Greg: Thanks for your time. Happy Holidays, Josh.
Josh: Same to you and Jewel, Greg.