Google reshapes Gemini
Google has turned Gemini toward persistent, project‑level workflows by integrating notebook‑style project spaces that let users organise chats, files and sources in one place. (heise.de) The company also gave Gemini the ability to operate apps on Pixel devices and expanded language and home support — moves that make the assistant more agentic but have drawn mixed user reviews on speed and reliability. (androidpolice.com) (androidcentral.com) (techradar.com)
Google is trying to turn Gemini from a chatbot you visit into a workspace you stay inside, with a new “Projects” feature that keeps your chats, files, and source material together instead of scattering them across separate prompts. (heise.de) That shift fixes a basic problem with chat assistants: every new question can feel like starting over, because the model forgets the folder, the draft, and the research trail a human would normally keep on a desk. (heise.de) Google’s answer is a notebook-style container inside Gemini where a user can pin files, collect sources, and keep related conversations in one place, so the assistant works more like a project partner than a one-off answer box. (heise.de) At the same time, Google is pushing Gemini beyond writing and research into action-taking on phones, letting it control apps on Pixel devices instead of only suggesting what the user should tap next. (androidpolice.com) That matters because a useful assistant needs memory and hands at the same time: memory is the saved project context, and hands are the ability to open, send, schedule, or navigate inside other software. (androidpolice.com) Google is also widening Gemini’s reach in the home, expanding support for more languages and regions so the assistant can replace or sit alongside Google Assistant on smart speakers and displays in more markets. (androidcentral.com) The catch is that smarter assistants are judged on waiting time, not just features, and some Google Home users have described Gemini responses as slow enough to break simple routines like turning on lights or answering quick questions. (techradar.com) So Google is making two bets at once: people will want one assistant for research, phones, and the home, and they will tolerate the transition pain while Gemini learns to do jobs that old voice assistants never handled well. (heise.de) (androidpolice.com) (androidcentral.com) If that works, Gemini stops being a better search box and starts looking more like an operating layer that follows a user from a research folder to a phone screen to a living room speaker. (heise.de) (androidpolice.com) (androidcentral.com) If it does not, Google will have added project memory, app controls, and global rollout before solving the older problem that made voice assistants frustrating in the first place: when a machine takes too long to do a small task, people stop asking. (techradar.com)
