NerdNews

Google has updated its AI Mode with a new feature called Personal Intelligence, which allows the AI to scan users' Gmail and Google Photos to provide more personalized and accurate responses. This feature is currently available to Google AI Pro and AI Ultra subscribers and can be enabled or disabled at any time.

Google's Gemini AI now has a 'personal intelligence' feature, allowing it to connect to Gmail, Photos, Search, and YouTube to provide more useful answers. The feature is optional, starts with paid users, and will roll out over the coming weeks.

Apple announced a multi-year partnership with Google to use its Gemini language models in the next-gen Siri, choosing it over OpenAI's ChatGPT. The new Siri is expected to arrive in an update to iOS 26, iPadOS 26, and macOS 26 Tahoe later this year.

Neurophos, a photonics startup, has raised $110M to develop tiny optical processors for AI inferencing, claiming its technology can outperform traditional silicon GPUs in terms of speed and energy efficiency. The company's metasurface modulator enables thousands of units to be fitted on a chip, resulting in faster calculations and lower power consumption.

Spotify introduces Prompted Playlists, an AI-powered feature that allows users to create playlists by describing what they want to hear. The feature is available to Premium subscribers in the US and Canada, and uses natural language processing to understand user requests and generate personalized playlists.

Google's updated Veo model can generate vertical videos from reference images with 4K upscaling, supporting mobile-first video and higher-resolution outputs, and is available in YouTube creator tools and the Gemini app.

Adobe After Effects has received its biggest update in years, featuring new tools such as native 3D design, improved vector workflows, and motion typography. The update also includes AI-powered Object Masking, improved shape masks, and Firefly Boards integration. Additionally, users can now access and apply Substance 3D materials to native meshes and imported models.

Inferact, the startup behind the open-source project vLLM, has raised $150 million in seed funding at an $800 million valuation. The company aims to commercialize vLLM, a technology that makes AI tools run faster and more affordably, and is co-led by Andreessen Horowitz and Lightspeed Venture Partners.

Humans& is building a new foundation model architecture designed for social intelligence, focusing on coordination and collaboration. The startup aims to create a 'central nervous system' for the human-plus-AI economy, enabling more effective teamwork and communication between humans and AI tools.

The article discusses the increasing presence of robots in daily life, with the author's building having multiple robots for cleaning and maintenance. The cost of hardware for robots has decreased, and AI advancements are expected to make robots more capable. However, the development of humanoid robots is still in its early stages, with companies like Tesla working on prototypes. The article also touches on the geopolitical implications of the robot revolution, with China potentially leading the way.

BioticsAI, the winner of TechCrunch Disrupt's Battlefield competition in 2023, has received FDA clearance for its AI-powered fetal ultrasound product. The technology uses computer vision AI to support fetal ultrasound quality assessment and anatomical completeness. The company plans to scale its product across health systems nationwide and add more features for fetal medicine and reproductive health.

LiveKit, a developer of infrastructure software for real-time AI voice and video applications, has announced a $100 million funding round at a $1 billion valuation. The company powers OpenAI's ChatGPT voice mode and has other notable customers such as Salesforce and Tesla.

A former Google trio is developing an interactive AI-powered learning app called Sparkli, designed for kids aged 5-12. The app uses generative AI to create immersive learning experiences, including audio, video, images, quizzes, and games. Sparkli aims to teach kids about modern concepts like skills design, financial literacy, and entrepreneurship, and has raised $5 million in pre-seed funding.

Ukraine is working with Palantir to develop an AI-powered system for training interceptor drones. The system, called Dataroom, will utilize real-world battlefield data to improve the autonomy of the drones. Palantir's software will help curate and normalize the data for training autonomous drone algorithms.

Cursor used AI agents to write a browser with 3 million lines of code, but it has an 88% job failure rate and is considered a demonstration of AI's ability to produce large, low-quality codebases.

An authentication bypass vulnerability in SmarterTools' SmarterMail email server allows unauthenticated attackers to reset admin passwords, gaining full privileges. The issue is now being actively exploited in the wild, with hackers reverse-engineering a patch released on January 15. Users are advised to upgrade to the latest version, Build 9511, to address the flaw.

Hackers are targeting Fortinet FortiGate devices in automated attacks, creating rogue accounts and stealing firewall configuration data. The attacks exploit an unknown vulnerability in the devices' single sign-on (SSO) feature. To secure their firewalls, admins can temporarily turn off the vulnerable FortiCloud login feature.

Apple is working on a wearable device that utilizes AI models, expected to be released as early as 2027. The device, similar in size to an AirTag, will feature a speaker, three microphones, and two cameras, with a single physical button and magnetic inductive wireless charging.

Confer is an open-source AI assistant that provides strong assurances of user data privacy, using end-to-end encryption and a trusted execution environment to prevent data from being read or tampered with by server administrators or hackers. It allows users to store conversations on Confer servers with confidence that they can't be read by anyone other than themselves.

Airbus and Keysight are collaborating on the Airbus UpNext SpaceRAN demonstrator to explore 5G non-terrestrial networks (NTNs) using advanced software-defined satellite technology. The partnership aims to tackle technical challenges such as seamless integration between terrestrial and non-terrestrial networks, reliable mobility management, and maintaining high-performance connectivity across dynamic satellite orbits.