You can now explore AI-generated versions of actual locations with Google’s Project Genie.

      Google is advancing its experimental AI world-building initiative into remarkably realistic realms. The company has announced that Project Genie can now utilize real-world images from Google Street View to create interactive virtual environments, merging genuine locations with creatively styled AI-generated visuals.

      Essentially, Genie represents what Google terms a “world model” — an AI system capable of constructing explorative digital environments where AI agents, robots, or users can interact seamlessly. Previously, these environments were largely artificial. However, with the latest update, Genie can reference actual places sourced straight from Street View imagery, giving a sneak peek into the future of simulation.

      Google invites you to explore AI fantasies.

      This feature operates within Project Genie, a prototype from Google’s experimental Labs. Users can now choose locations in the United States via a Maps pin and subsequently transform these areas into stylized AI-driven worlds. For instance, the Golden Gate Bridge could be reimagined as an underwater adventure zone teeming with marine life, while the Fort Worth Stockyards can be re-envisioned as a vintage black-and-white scene from the 1920s, replete with saloons and classic cars.

      Moreover, users can personalize the characters navigating these environments, whether by becoming a comic book hero, an animal, or even a claymation creature. The intriguing aspect here goes beyond mere creativity; it involves how Google is training AI systems to comprehend real-world geography and settings through simulation.

      Google's broader AI ambitions are becoming evident.

      Google has indicated that Genie has already been employed internally for AI research, and even by Waymo to model realistic driving scenarios. Grounding these simulations in Street View data could significantly enhance the practicality of future AI training.

      Project Genie is now being globally rolled out to qualified Google AI Ultra subscribers, though Google still classifies the platform as experimental, likely explaining the company's cautious approach in promoting it extensively.

      Google aims to transform your TV remote with Gemini and pointer controls.

      Google is making a significant push into the living room, this time focusing not just on what you watch but also on how you engage with your TV. At Google I/O 2026, the company unveiled a new set of updates for Google TV and Android TV developers, all revolving around a single concept: TVs are no longer just static screens in your living space. With over 300 million monthly active devices across Google TV and Android TV, Google clearly views television as its next key AI arena, with Gemini at the heart of this strategy.

      The company claims that Gemini is already assisting users in discovering content through intuitive voice interactions. However, Google now seeks to make the experience feel more interactive and conversational, akin to searching the web — but from your couch. Rather than just presenting static results, Gemini on Google TV can now reply with a mix of visuals, videos, and text snippets to fulfill queries. So if someone inquires about a thriller featuring a strong female lead or a documentary on space exploration, Gemini retrieves relevant recommendations directly from streaming services and their metadata.

      At present, nearly every significant AI chatbot follows a similar model: attract users with a surprisingly effective free tier, then gradually encourage them toward a subscription as they become more dependent on it. In truth, for many users, the free versions are already quite satisfactory. You can pose questions, create images, summarize documents, and even brainstorm ideas without hitting a paywall repeatedly. That is why a newly discovered change within Google’s Gemini app is notably intriguing.

      A user on X has shared a screenshot suggesting that Google might be testing stricter usage monitoring and potential weekly limitations within Gemini. The screenshot reveals a new section explaining, “Plan limits determine how much you can use Gemini over time.” This implies Google could be preparing a more robust system that tracks how often free users engage with Gemini, especially when utilizing more intensive AI models.

      Scientists have recently set a wireless speed record that could influence the future of 6G.

      Researchers have achieved a wireless connection speed of 112Gbps in the 560GHz band, demonstrating capabilities that current mobile networks cannot match. A team from Tokushima University utilized soliton microcombs to generate a more stable terahertz signal for anticipated 6G systems.

      The immediate benefit isn’t a faster mobile device. Instead, it lies in the underlying infrastructure that transfers data between network sites, where backhaul capacity determines whether future 6G speeds remain tangible or become hindered by saturated network pathways. Consequently, this is a significant breakthrough in 6G speed to monitor, even if consumers won’t see it reflected on specification sheets anytime soon.