NerdNews

Google has announced Gemini Robotics, a model designed for real-life robots with articulate hands, and Gemini Robots-ER, a vision-language model with advanced spatial understanding. These AI models will be used to power robots, enabling them to perform tasks and navigate their environment. Google is partnering with companies like Apptronik to build the next generation of humanoid robots.

Google's Gemini 2.0 Flash introduces native multimodal AI image generation, allowing for fast edits, style transfers, and conversational image editing. It integrates multimodal input, reasoning, and natural language understanding to generate images alongside text, enabling developers to create illustrations, refine images, and generate detailed visuals based on world knowledge.

Nvidia's ACE AI game characters will be introduced in Krafton's life simulation game InZOI and Naraka: Bladepoint Mobile, allowing for more realistic and intelligent NPC interactions. The technology uses an on-device Nvidia ACE Small Language Model to govern character thoughts and behavior, enabling them to adapt and respond to their environment. This innovation has the potential to revolutionize the gaming industry by making games richer and more surprising.

Google's Gemini Deep Research, a tool that creates comprehensive reports on complex topics, is now available to all users for free in over 45 languages. The tool is powered by Gemini 2.0 Flash, a chain-of-thought system that enhances research capabilities. It allows users to ask Gemini to create detailed reports, which are more insightful and multi-page. Additionally, Google has introduced an experimental feature called Gemini with personalization, which tailors responses based on users' past searches and activities.

The iPhone 17 series is expected to be announced in September, with a new ultra-thin iPhone 17 Air, iPhone 17, and iPhone 17 Pro. The iPhone 17 Air may have a single camera lens and a basic A19 chip. iOS 19 is also expected to have a major redesign and potentially more AI-powered features.

House Democrats sent letters to 24 federal agencies asking for assurances that Elon Musk's DOGE team is not feeding sensitive government data into unapproved AI systems, citing concerns over cybersecurity and data privacy. The lawmakers are seeking explanations on the legal authority used to access government data, what data has been collected and analyzed, and what privacy protections are in place.

Microsoft is testing an AI-powered text summarization feature in Notepad, allowing users to condense content. The feature, called 'Summarize', is available to Windows 11 Insiders in the Canary and Dev Channels. It can be accessed by selecting text, right-clicking, and choosing 'Summarize', or using the Ctrl + M keyboard shortcut. The feature requires a Microsoft account and uses AI credits linked to Microsoft 365 subscriptions.

A study by Columbia Journalism Review found that AI search engines, including ChatGPT Search and Grok 3, provide incorrect information over 60% of the time, with some models confidently delivering incorrect responses more often than their free counterparts. The study tested eight AI-driven search tools and discovered issues with citations, publisher control, and URL fabrication.

Microsoft is testing AI-powered summaries in Notepad for Windows 11. The feature allows users to summarize text by highlighting it, right-clicking, and selecting 'Summarize'. The Snipping Tool is also getting an update with a 'draw & hold' feature to draw straight lines. These updates are rolling out to Windows Insiders in the Canary and Dev channels.

Google's Gemini can now use your search history to provide more personalized responses. The feature is powered by the Gemini 2.0 Flash Thinking Experimental model and is optional. It will reference your search results if its AI model finds it 'helpful'. Gemini's personalization feature will also connect to other apps like YouTube and Google Photos to provide more insights.

DeepMind's Gemini Robotics AI model enables robots to perform complex tasks like folding origami and closing Ziploc bags, with improved generality, interactivity, and dexterity. The model is being used to build the next generation of humanoid robots and is available to trusted testers.

Google's Gemini Robotics AI model enables robots to understand and interact with the physical world, performing tasks like origami folding and packing snacks into bags. The model demonstrates strong generalization, allowing it to perform novel tasks without specific training. Google has partnered with Apptronik to build the next generation of humanoid robots using Gemini 2.0.

Google DeepMind has announced new AI models called Gemini Robotics, designed to enable real-world machines to interact with objects, navigate environments, and more. The models were trained to generalize behavior across different robotics hardware and can connect items robots can 'see' with actions they might take. A slimmed-down model, Gemini Robotics-ER, has been released for researchers to train their own models, along with a benchmark called Asimov for gauging risks with AI-powered robots.

Ditto, a company that provides edge data synchronization, has raised $82 million in a Series B round of funding. The company's platform allows for peer-to-peer data synchronization between edge devices and the cloud, reducing latency and optimizing bandwidth requirements. Ditto's solution is particularly useful for industries that require real-time data processing, such as airlines and autonomous vehicles.

OpenAI has submitted proposals to the White House Office of Science and Technology, including ensuring access to data for AI training, adopting measures to drive growth, and maintaining a three-tiered AI diffusion rule framework. The company also wants the US government to prevent foreign countries from imposing their copyright rules on American AI firms.

Travis Kalanick, Uber's former CEO, expressed regret over the company's decision to abandon its autonomous driving program, stating that having an autonomous ride-sharing product would be beneficial now. Uber sold its self-driving unit to Aurora in 2020, and Waymo's self-driving cars are now being used in various markets, including a partnership with Uber in Austin.

Nvidia's RTX Remix allows modders to upgrade classic games with ray tracing, AI-enhanced textures, and more. A Half-Life 2 RTX demo will be available on Steam, showcasing the tech's capabilities. RTX Remix has exited beta and now includes DLSS 4 with multi-frame generation and RTX neural rendering tech.

Microsoft is developing Copilot for Gaming, an AI assistant that will help gamers improve at their favorite titles by offering tips and guides. The assistant will be available on the Xbox mobile app and will have the ability to see the game screen to provide real-time guidance. However, details on its functionality are limited, and it's unclear what features will be available at launch.

Microsoft has updated Azure Elastic SAN with auto scale, snapshot support, and CRC protection. The auto scale feature simplifies management by automatically increasing capacity when storage runs low. Snapshot support allows users to take instant backups of their workloads, and CRC protection helps maintain data integrity.

Apple is planning to introduce a live translation feature to its AirPods, allowing for real-time conversations to be translated. The feature, tied to iOS 19, will enable users to translate conversations from one language to another, with the translated audio played through the AirPods. This feature is similar to one already available on Google's Pixel Buds.