In this episode of Lenny’s Podcast, host Lenny Rachitsky is joined by Josh Woodward, Head of Google Labs, the Gemini app, and AI Studio, who has spent over 16 years at Google experimenting with and scaling new AI products. Lenny’s core thesis is that every company must now think and operate like a labs team. Because AI technology and capabilities are changing so rapidly, organizations need a dedicated team operating at the frontier—trying new models, running experiments, and uncovering opportunities before competitors or startups do.
### Finding Great Ideas and Product-Market Fit
Woodward agrees with this premise, emphasizing that good ideas rarely come from conventional brainstorm meetings or design sprints. Instead, they emerge organically from people hacking away on weekends or taking side projects. Woodward shares a useful tactical habit: maintaining a list of things that are "almost possible" and watching for the exact moment a technological shift or phase change makes them a reality.
When evaluating early prototypes, Woodward relies on a simple human metric rather than data dashboards: looking at people’s eyes to see if they light up and lean in. Ideas worth pursuing solve a painful user problem so effectively that people are willing to pay for it. Conversely, projects should be killed when passion runs out and the team knows it isn't working. Woodward stresses the importance of fostering a culture where the team feels safe enough to call out failing ideas before leadership does.
### The Labs and Gemini Structure
Woodward oversees both Google Labs (zero-to-one experiments) and the Gemini app (serving over a billion users globally). Despite the massive difference in scale, both share a zero-to-one heartbeat and a guiding mantra: users first, Google second, and products third. Woodward notes that Google Labs categorizes its projects into stages—zero-to-one labs, one-to-10 labs, and 10-to-100 labs—each optimized for very different metrics and challenges.
### Over/Under-Hyped Trends and Valuable Skills
* **Under-hyped:** The discussion around product principles and values. Woodward believes products encode the principles of the people who build them, and the industry needs more conversations about the kind of future they actually want to create.
* **Over-hyped:** Model benchmarks and ELO ratings. Most people in the world do not care about benchmarks; what matters is whether a product delivers real value.
In terms of trending skills, Woodward looks for "unlearning rate"—how fast someone can learn something and then walk away from it based on new data. He also values "explosive endurance" (the ability to go hard while maintaining long-term sustainability) and the capacity to build trust and collaborate seamlessly with both AI agents and humans. While job roles are blurring, Woodward warns against the overhyped notion that everyone is now just a generic "builder," noting that professional specialties and deep domain expertise remain vital.
### Setting Up an Internal Labs Team
For companies looking to set up their own labs team, Woodward offers three key tips:
1. **Independence:** Create a space where weird things can grow without being nestled under an existing business unit.
2. **Clear alignment:** Ensure the lab's goal is to make users and the core company successful rather than just serving its own glory.
3. **Good trouble:** Use the lab not just for technology exploration, but to challenge and pilot new internal processes, such as job ladders and ways of working.
The conversation concludes with lighthearted lightning-round questions, including Google Labs’ unique internal tradition of awarding miniature Etsy rakes to teammates who successfully optimize and harvest TPUs.