Sergey Brin, Google's co-founder, has returned to the company on an almost daily basis since 2023 to contribute to the development of AI products, particularly focusing on the Gemini AI model. His return signifies Google's intensified focus on AI, as it competes with OpenAI's ChatGPT. Brin's work involves delving into technical details and assisting in training the latest Gemini models, leveraging his scientific interests and allowing DeepMind CEO Demis Hassabis to manage the broader operations.

Brin also reflected on the failure of Google Glass, attributing it to a lack of understanding of consumer electronic supply chains and manufacturing complexities. However, he is optimistic about Google's new XR (Extended Reality) glasses, believing that advancements in AI will enhance their capabilities. Google's I/O 2025 keynote showcased significant AI upgrades across its product ecosystem, emphasizing a strategic shift towards embedding generative AI. Key announcements included updates to Gemini, AI Mode in Search, new image and video models, Beam for 3D video calls, and smart glasses on Android XR.

The Gemini app is receiving a substantial upgrade, with Gemini Live (built on Project Astra) now supporting camera input and screen sharing for real-time conversations. Gemini is being integrated across Google's apps, enabling users to interact with Gmail and Calendar through natural language. A new Agent Mode will allow Gemini to handle multi-step web tasks, while Deep Research mode supports file and image uploads. AI Mode in Search is also undergoing a major overhaul, offering more visual and interactive results, personalized through Gmail and search history integration. The feature is currently rolling out in the United States, with plans for global expansion.

Google unveiled Imagen 4 for text-to-image generation and Veo 3 for video, offering higher visual fidelity, improved text rendering, smoother motion, and AI-generated soundtracks. Beam, Google’s new 3D video conferencing system, aims to create lifelike remote meetings through a six-camera array and a custom light field display. Android XR, an extended reality operating system, supports mixed reality on headsets and glasses, with partners like Warby Parker and Gentle Monster developing XR-enabled products.

Google's strategy involves embedding Gemini into daily workflows such as Search, Chrome, Gmail, and Meet. The company is also monetizing AI through an “AI Ultra” subscription, bundling high-end Gemini models, early feature access, and higher usage caps. Google is aiming to lower switching costs for creators and developers through upgraded generative tools and agentic APIs, positioning itself as the “pro-am” creative suite alternative. Projects like Astra and Mariner are generating high-value, real-world task traces to improve AI capabilities.

The new “AI Mode” in Google Search allows users to switch to a Gemini-generated answer pane, making chat-style search the default. Google is betting that distribution plus incremental agentic capabilities will outweigh raw model horsepower. The company is also lowering the talent and capital floor in markets like video production and browser automation by bundling tools like Veo 3 and Mariner. Google's efforts at I/O 2025 reflect a strategy to lock in user attention by embedding Gemini across various platforms, monetize AI directly, and prime the next platform (XR) before competitors can dominate.