Gemini 3.5 Flash and Gemini Omni: Faster, Smarter, and Built for Video
At this year’s Google I/O, Gemini moved from experimental sidekick to core infrastructure. Gemini 3.5 Flash is positioned as Google’s fast, scalable model for everyday tasks, designed to power lightweight interactions across products without sacrificing responsiveness. Alongside it, Google introduced the Gemini Omni video model, used in the overhauled Google Flow creative suite to keep characters and visual style consistent from frame to frame. This means AI-generated scenes no longer feel like stitched-together stills, but closer to coherent short films. For users, these upgrades should translate to quicker answers, smoother AI features in apps, and more reliable visual outputs. For Google, they are the technical backbone of a strategy to embed Gemini into everything from search results to productivity tools, shifting expectations from “chat with a bot” to “let the system quietly handle work in the background.”
Google AI Search Updates: From Links to Conversations
Search received one of the most consequential Google AI search updates in years. The familiar search box is evolving into an “intelligent search box” that supports conversational follow-ups, richer context, and even file or video attachments. AI Overviews and AI Mode are effectively merged into a single, fluid experience: instead of hopping between modes, you refine your query in a chat-like flow. Google is also layering in generated visuals and short explanatory videos directly in the results, using Gemini’s multimodal capabilities. The goal is to keep you inside Google’s results while it synthesizes the web, documents, and media on demand. This could make complex searches—like planning projects or learning new skills—much easier, but it will also intensify debate about what happens to traffic that used to flow to publishers when answers appear inside Google’s own AI-generated summaries.
Gemini Integration in Gmail, Shopping, and Everyday Workflows
Beyond search, Gemini integration in Gmail and other Workspace tools is expanding so users can treat the assistant as a proactive organizer rather than a passive chatbot. Think of Gemini summarizing long email chains, pulling attachments into a brief, and then using those insights when you switch over to Docs or Sheets. In shopping, Google is pushing more agentic AI: you’ll increasingly ask for goals (“help me outfit a home office under these constraints”) instead of typing product keywords, with Gemini comparing options across the web. Video creation also benefits from the Gemini Omni video model, which underpins Google Flow’s filmmaker tools and works alongside image features powered by Nano Banana. Collectively, these changes signal a shift toward task completion: Gemini is meant to follow you from inbox to browser tab to creative canvas, carrying context so you do not have to repeat yourself.
Spark, AI Studio App, and Tools for Developers and Creators
To make Gemini more than a consumer-facing assistant, Google is building a deeper stack of tools. Spark positions Gemini as an agentic layer that can take multi-step actions in apps and on the web, while new agentic coding features in Chrome DevTools and Modern Web Guidance help developers build and debug sites with AI support. WebMCP lets web pages themselves become toolkits for agents, giving them controlled autonomy. On the creative side, Google Flow gains more AI collaborators, such as Flow Agent for script and plot assistance, and Pics for image generation and editing. An AI Studio app from Google ties into this ecosystem, giving developers and advanced users a workspace to prototype agents and multimodal workflows powered by models like Gemini 3.5 Flash. The shared theme is clear: Google wants anyone—coders, marketers, or filmmakers—to productize Gemini-powered experiences quickly.
A More Pervasive Gemini—and the Infrastructure Behind It
All these upgrades hinge on Gemini becoming a quiet, ever-present layer across Google’s ecosystem. The redesigned Gemini app, with its Neural Expressive UI and Gemini Live for fluid voice interactions, hints at how Google envisions everyday use: you talk or type once, and Gemini follows you through tasks, whether that’s searching, drafting, editing media, or orchestrating agents. Under the hood, Google is pairing these ambitions with a massive infrastructure push. A new AI cloud venture with Blackstone, backed by a USD 5 billion (approx. RM23.0 billion) equity investment, will expand access to Google’s Tensor Processing Units as compute-as-a-service for enterprises. That move aims to meet surging demand for AI workloads and challenge existing chip leaders. Taken together, the product and infrastructure announcements show Google betting that making AI both omnipresent and easy to build on will be its competitive edge.