Tech Roundup: Google Gemini 2.0, OpenAI Sora & More

[A recurring feature on the latest in Science & Technology.]

• Google unveils Gemini 2.0 with support for multimodal inputs like images, video and audio, and multimodal output like natively generated images mixed with text and steerable text-to-speech (TTS) multilingual audio; also comes with capabilities to natively call tools like Google Search, code execution as well as third-party user-defined functions.
• Google debuts an updated Project Astra, which can record and summarise videos recorded using an Android app or prototype glasses and answer questions; Jules, an AI coding assistant that can autonomously fix software bugs and prepare code changes as part of a GitHub workflow; Project Mariner, a prototype AI agent built by Google DeepMind that can control Chrome, move the cursor, click buttons and fill out forms; Deep Research, an AI tool to let its Gemini chatbot scour the web and write a detailed report; and Trillium, its sixth-generation AI chip powering Gemini 2.0.
• OpenAI formally launches Sora, its text-to-video generation model, in select countries as it takes on video-synthesis models from competitors such as Google's Veo, Runway's Gen-3 Alpha, Kling, Minimax and Hunyuan Video; lets ChatGPT Pro users create up to 500 videos per month at up to 1080p and ChatGPT Plus users up to 50 videos at up to 720p, but blocks content with nudity and limits the ability to generate videos with people to a limited number of early testers to prevent misuse and address concerns around misappropriation of likeness and deepfakes. (Sora has also raised concerns that OpenAI may have trained its technology on copyrighted materials from YouTube.)
• Apple gets hit with a US$ 1.2 billion class-action lawsuit in the U.S. over its alleged failure to detect and report illegal child pornography on its platform; comes after the company's decision to scrap a controversial plan to scan users' devices for objectionable content over concerns that it could be abused by governments to illegally surveil Apple users for other reasons.
• Bluesky teases Bluesky+, an US$ 8 per month paid subscription offering profile customisations, higher video upload limits, high quality video resolution, custom app icons, bookmarks, post analytics and inline translations, as it says it's not completely ruling out an ad-supported version.
• Google shows off Genie 2, a large-scale foundation world model that's capable of "generating an endless variety of action-controllable, playable 3D environments for training and evaluating embodied agents" based on a single prompt.
• Apple formally releases new versions of iOS, iPadOS, and macOS with Apple Intelligence and OpenAI ChatGPT integration; brings Apple Intelligence outside of the U.S. for the first time to include Australia, Canada, Ireland, New Zealand, South Africa, and the U.K.
• Apple adds Layered Recordings to the Voice Memos app in iOS 18.2, for the iPhone 16 Pro and Pro Max, allowing users to record vocals while listening to an instrumental track out loud without the need for headphones.
• WordPress-owner Automattic faces legal setback after a U.S. court orders the company to stop blocking to stop blocking WP Engine's access to WordPress.org resources and reinstate WP Engine's Advanced Custom Fields (ACF) plugin; WordPress CEO Matt Mullenweg says "I'm sick and disgusted to be legally compelled to provide free labor to an organisation as parasitic and exploitive as WP Engine." (Mullenweg publicly denounced WP Engine in September, calling it a "cancer" to the larger Wordpress open-source project and accusing it of improperly using the WordPress brand.)
• Microsoft says it's starting to roll out a way for users to share files between their iPhone and Windows 11 or Windows 10 PCs via the company's Phone Link app and Link to Windows app.
• Google updates Pixel Screenshots app with integration with Gboard, Circle to Search and Wallet to extract and use data from screenshots; rolls out two updates to Android's unknown tracker alerts feature, letting users pause location updates from their phone for up to 24 hours and pinpoint unfamiliar trackers.
• Google expands Gemini integration in Drive with the ability to summarise documents or provide information about a specific folder; adds automatic inline translation to Google Chat.
• The U.S. government instructs Apple and Google to be prepared to comply with a law that could result in TikTok facing an effective ban in the U.S. next month, after a U.S. appeals court rejects TikTok's request to temporarily halt the law from taking effect. (Although TikTok called the law unconstitutional and said it violates the First Amendment rights of its 170 million users, the court rejected that argument and said in an opinion that the law "is narrowly tailored to protect national security." TikTok has claimed that one month of a U.S. ban would result in U.S. small businesses and social media creators losing $1.3 billion in sales and earnings.)
• Google rolls out a subscription tier and the ability to "engage directly with the AI hosts during an Audio Overview" as part of a refresh of its NotebookLM tool; also adds support for Gemini 2.0 Flash.