Google has launched a feature called Skills in the Chrome browser that lets users save AI prompts and run them again with a single click. The feature is available through the Gemini sidebar in Chrome on desktop. Without Skills, repeating an AI task required users to type the same prompt each time they visited a …
Google has released a native Gemini app for macOS. The app is available for free to all Gemini users on macOS 15 and later and can be downloaded at gemini.google/mac. Unlike using Gemini in a browser, the desktop app is always accessible via a keyboard shortcut: Option + Space. This opens a pill-shaped prompt bar …
YouTube has introduced a new feature that allows users to create a photorealistic AI avatar of themselves for use in Shorts. Abner Li reports for 9to5Google that the tool captures both a user’s appearance and voice to generate personalised video clips. To set up the avatar, users record a live selfie and read a few …
Google is adding a notebooks feature to the Gemini app, giving users a dedicated space to organize chats, files, and documents around specific topics or projects. The company describes notebooks as personal knowledge bases that will eventually work across multiple Google products. Users can create a notebook from the side panel in the Gemini app. …
Google’s AI Overviews are accurate about 91 percent of the time. This sounds good at first, but Tripp Mickle and colleagues report for The New York Times that this still means the search engine delivers tens of millions of incorrect answers every hour. The New York Times commissioned AI startup Oumi to test Google’s system …
Google has expanded its video creation tool, Google Vids, with several new AI-powered features. The updates cover music generation, customizable avatars, and direct publishing to YouTube. Users can now generate original music tracks directly inside Google Vids. Powered by Google’s Lyria 3 and Lyria 3 Pro models, the tool lets users type a text prompt …
Google has launched Gemma 4, a new family of open-weight AI models that the company describes as its most capable to date. The models are built on the same research and technology as Google DeepMind’s proprietary Gemini 3 system and are released under an Apache 2.0 open-source license, which allows developers to use and modify …
Google has added Veo 3.1 Lite to its Gemini API, positioning it as the most affordable model in the Veo 3.1 family. The company says the model targets developers building high-volume video applications. According to Google, Veo 3.1 Lite costs less than 50% of Veo 3.1 Fast while offering the same generation speed. Pricing on …
Google Research has released TurboQuant, a new compression algorithm designed to reduce the memory demands of large language models. The company says it can shrink a model’s key-value cache by at least six times and speed up a core processing step called attention computation by up to eight times, all without retraining the model or …
Google has expanded its AI music generation capabilities with the launch of Lyria 3 Pro, a model that creates tracks up to three minutes long. Myriam Hamed Torres writes for Google DeepMind that the model understands musical structure, allowing users to prompt for specific elements such as intros, verses, choruses, and bridges. Lyria 3 Pro …