Google DeepMind Launches Interactive AI Magic Pointer Demos: Try the Future of Computing
The project, powered by Gemini, tries to remove the friction associated with AI tools that live in their own window. This in turn forces users to stop what they're doing, switch apps, and explain or re-explain their context. DeepMind's prototype is designed to fix that. The AI follows the cursor across whatever the user is doing, so pointing at something and asking about it, doesn't require a separate, detailed prompt. Hover over a table of statistics, for example, and users can ask for a pie chart, or highlight a recipe and say "double these ingredients," or point at a PDF and request a bullet-point summary to paste directly into an email. The system captures the visual and semantic context around the pointer so the model already knows the context.
Two demos are available right now in Google AI Studio: one for editing an image and one for finding places on a map, both operated by pointing and speaking. Gemini in Chrome is rolling out today as well, with Magic Pointer for Googlebook coming later this year.
Google also announced its Googlebook yesterday. The Googlebook is supposedly a new laptop category built from the ground up for Gemini intelligence, combining what Google considers the best of Android and ChromeOS and is designed to work seamlessly with Android phones. Magic Pointer is considered one of the centerpiece features. Users will be able to wiggle the cursor and have Gemini activate, offering context-specific suggestions based on whatever is on screen. Point at a date in an email to set up a meeting, select a photo of your living room alongside a couch you like, and Gemini can visualize how they look together. Hardware partners confirmed for Googlebooks include Acer, ASUS, Dell, HP, and Lenovo, with devices expected this fall. Intel and Qualcomm have also announced they will be providing chips to power the new Googlebooks. All models will feature a "glowbar" on the exterior, which Google describes as both functional and beautiful, though specifics on what it actually does remain under wraps.
Googlebooks will also support a "Create your Widget" feature, letting users build custom desktop widgets through simple prompts. Gemini can pull from Gmail, Calendar, and the web to build a personalized dashboard, An example might be a family reunion planner that consolidates flights, hotels, restaurant reservations, and a countdown in one spot.
Gemini Intelligence, the broader AI layer announced alongside Googlebook, can access native and third-party apps for food delivery, ridesharing, and travel, among others. Adding items to a cart is another example, but it stops short of placing orders to avoid making financial decisions on a user's behalf, according to the company. That layer of features is expected to roll out to the Galaxy S26 and newer Pixel devices this summer too.
The experimental AI pointer demos are live now at Google AI Studio. Googlebook devices and the full Magic Pointer experience are expected when the hardware launches this fall.
