Lenny's Podcast: Google Labs Playbook for Building Like a Lab Team

SPEAKER_1 (0:01)

When AI moves this fast, companies cannot afford to stand still. Today we're breaking down Lenny Rachitsky's conversation with Josh Woodward, who leads Google Labs, the Gemini app, and AI Studio. His main argument? Every company now needs to think and operate like a lab team.

And Josh's reasoning is pretty compelling. The frontier is shifting so quickly that the winners will be the teams that experiment, learn, and ship before everyone else. But here's what caught my attention—he says the best ideas rarely come from formal workshops or design sprints.

Right, they come from curiosity, hallway conversations, weekend tinkering, and people who keep pulling on something that catches their eye.

Josh has this great line: 'You cannot microwave good ideas.' So leaders need to create space where unusual ideas can grow, where small groups of obsessed people can build something real. That frontier mindset also shapes how he evaluates products. Get this—the earliest signal of product-market fit isn't a dashboard, it's a human reaction. He says, 'I am always looking at people's eyes—that is the metric.' If someone leans in, lights up, and wants to know more, that matters more than retention charts in the beginning.

And when an idea's not working? He says the team usually knows before the leader does. Now, let's talk about what Google Labs is actually doing. Josh describes a spectrum from zero-to-one experiments to products that scale to millions of users. The whole organization follows this mantra: 'users first, Google second, and our product third.' There's also this concept of getting to what's 'almost possible,' then building quickly once the technology crosses that threshold. Speaking of the future, Josh argues that consumer tech isn't dead at all. AI is opening up new opportunities in entertainment, messaging, and products that help with time, money, and real-world connection. He mentioned Google is actively working on more personal and proactive experiences, including better use of connected data.

But here's one of his sharpest takes—benchmarks are overhyped. People in the real world don't care about ELO scores or model charts nearly as much as the industry does. What's underhyped then? The role of principles and values, because products quietly communicate what a company believes. Josh also describes some emerging skills becoming more valuable: 'unlearning rate,' 'explosive endurance,' and people who can collaborate across humans and agents. Teams are getting smaller, faster, and more cross-functional, with builders increasingly blending PM, design, and code. Though he's careful not to erase specialization entirely.

The episode closes with a warning and a playbook. Internal labs often fizzle out after a few years unless they can bridge experimentation into the larger company. To avoid that, Josh says leaders need independence, clear goals, and enough freedom to 'create a space where weird things can grow.' That's the blueprint right there.