Google keeps transforming Gemini desktop into superapp

Google's ongoing development of Gemini desktop points to a broader strategy to position its AI as a central operating system for users, moving beyond a simple chatbot. The introduction of Ask and Assign modes, particularly Assign's ability to operate within specific folders, suggests Google is building an AI agent with deep local integration. This could offer significant productivity gains for Asian professionals who rely on organized digital workspaces. The potential for remote control features via connections to other Gemini desktop instances also bears watching, as it could redefine how distributed teams collaborate on AI-driven tasks. The integration with Obsidian, a popular knowledge management tool, indicates Google's interest in leveraging existing user workflows or developing similar internal capabilities. While the current build is limited to Google apps, the Customize tab for skills and plugins suggests a future open ecosystem. The mention of Nano Banana 2.5 Flash and Gemini 4 in code traces confirms Google's commitment to continuous model upgrades, which will be critical for Gemini's competitiveness against rivals like Claude Desktop in the Asian enterprise market.