Peter Yang• Oct 12, 2025• 43:57Interview

Inside Gemini and NotebookLM: How Google is Shipping Non-Stop | Josh Woodward

From Peter Yang

Josh Woodward•VP at Google Labs

Executive Summary

Google Labs operates with a rapid shipping culture, launching new AI products like Flow in under 100 days by using small, autonomous teams of 5-7 people.
User-driven, viral features are a major source of growth and product direction.
The 'custom mini figurine' feature in Gemini (Nano Banana) became so popular it accounted for 30-40% of queries in some regions and strained Google's infrastructure.
The future of Google's AI is a proactive, personal assistant.
A key strategy is integrating Gemini with users' personal data (Gmail, Calendar) and seamlessly connecting different modalities, such as image generation (Nano Banana) and video generation (Veo).
Google's hiring philosophy for innovation teams prioritizes 'tinkerers' who build prototypes in their spare time, learn quickly, and possess a high-energy, optimistic mindset, scaling teams only after a product shows market pull.

12 quotes

Concerns Raised

Managing the immense infrastructure strain caused by wildly successful generative AI features, which can require temporarily capping usage.
The potential for a fragmented user experience as new features are tested on separate URLs before being integrated into core products.

Opportunities Identified

Becoming a proactive, personal life assistant by integrating with user data from Google's ecosystem (Gmail, Calendar).
Unlocking new creative and commercial use cases by seamlessly combining image and video generation (Nano Banana + Veo).
Leveraging the constant stream of new model capabilities from Google DeepMind to rapidly build novel product experiences.
Automating enterprise workflows, such as generating presentation decks from source documents using Notebook LM's 'video overviews' feature.

Key Themes

Rapid Prototyping and Shipping Culture

Google Labs emphasizes speed, moving from idea to public launch in record time (e.g., Flow in 86 days). This is achieved by keeping teams small, prioritizing prototyping over documentation, and getting products into users' hands as quickly as possible to gather real-world feedback.

This highlights a cultural shift within a large corporation to adopt startup-like agility, which is critical for competing in the fast-paced AI landscape.

User-Driven Feature Development

The development process is highly responsive to user behavior. The 'custom mini figurine' feature, for example, was a viral trend started by users in Thailand that Google observed and then integrated as a default prompt, demonstrating a powerful feedback loop between the community and the product team.

This approach ensures product development is aligned with demonstrated user needs and desires, increasing the likelihood of feature adoption and engagement.

Multimodal AI Integration

A core strategy is to create seamless connections between different AI capabilities. The plan to integrate the Nano Banana image model with the Veo video model, and the fact that Gemini was designed to be natively multimodal, shows a focus on building a comprehensive and interconnected creative and productivity suite.

Breaking down the silos between text, image, video, and audio generation unlocks more complex and powerful use cases, moving closer to a single, unified AI assistant.

Strategic Team Building for Innovation

The approach to building 'zero-to-one' product teams is methodical. It involves starting with small teams of 5-7, hiring for traits like a passion for tinkering and rapid learning, and only committing to scaling the team after the product has demonstrated significant user traction (e.g., 10,000 active users).

This provides a practical model for fostering innovation within large organizations by minimizing bureaucratic drag and aligning resources with proven market demand.

Get started free

Topics

AI Product Development Google Labs Google Gemini Generative AI Image Generation Video Generation Nano Banana Veo Model Notebook LM Shipping Culture Team Building User-Driven Innovation Multimodality AI Assistants Product Management

Processed Apr 9, 2026 yt-dlp + mlx-whisper + Gemini