NerdNews

The US government has announced a $500 billion AI infrastructure project called Stargate, aiming to expand computing capacity for current and future AI projects, including OpenAI's goal of achieving AGI. The project will create 100,000 US jobs and is expected to secure American leadership in AI. The initiative is a collaboration between OpenAI, SoftBank, Oracle, and MGX, with technology partnerships with Arm, Microsoft, and Nvidia.

Google has increased its investment in AI startup Anthropic by $1 billion, bringing its total investment to over $3 billion. Anthropic is a rival to OpenAI and is valued at around $60 billion. The investment is part of Google's efforts to diversify its AI business and take on competitors like Microsoft and Amazon.

Microsoft is reorganizing its entire developer team to focus on AI, with a new division called CoreAI - Platform and Tools. This move aims to develop tools for agentic AI, which is expected to transform the tech industry and the world of work. The new division will be led by Jay Parikh and will focus on building the end-to-end Copilot & AI stack for both first-party and third-party customers.

OpenAI CEO Sam Altman stated that the company is now confident in its ability to build Artificial General Intelligence (AGI), with potential applications in replacing human workers and increasing productivity. However, critics such as Gary Marcus have expressed skepticism, citing the limitations of current language models and the need for further research. AGI is defined as a highly autonomous system that outperforms humans at most economically valuable work, and its development could have significant societal implications.

Google's Gemini is surpassing competitors like ChatGPT, Siri, and Alexa, with its integration on Samsung phones and access to more information and users, making it the most capable virtual assistant on the market.

Adobe Premiere Pro's new AI-powered Media Intelligence tool analyzes thousands of clips in seconds, identifying content and making it searchable by text, potentially saving video editors hours of time. The tool is available in Premiere Pro beta and can recognize people, objects, locations, and more, allowing for quick location of specific clips in large libraries.

Samsung has teased its upcoming AR glasses at the Galaxy Unpacked conference, with the company set to reveal more details later this year. The glasses will feature Google Gemini and AI assistant capabilities, similar to the Galaxy S25 phones, and will include 'agentic' cross-app capabilities.

Google is announcing a new feature that allows users to control their Chromebook with their head and facial expressions, aimed at those with motor impairments. The feature is rolling out to more users with compatible Chromebooks. Additionally, Google is introducing new Chromebooks and classroom-focused ChromeOS features, including Class Tools, which allow teachers to have real-time control of their students' screens.

Nvidia has confirmed a previously leaked prototype of a quad-slot GPU cooler, which influenced the design of the new RTX 5090. The prototype featured a triple-fan cooling system, but the final RTX 5090 uses a modular four-part design with a separate I/O board for a more compact cooler.

Canon has developed a 410-megapixel, 35mm full-frame CMOS sensor, the largest number of pixels ever achieved in a sensor of its size. The sensor has a resolution of 24K, 198 times greater than HD, and 12 times greater than 8K, making it suitable for surveillance, medicine, and industry. It features a redesigned circuitry pattern and a back-illuminated stacked formation, allowing for a readout speed of 3,280 megapixels per second and video at eight frames per second.

Tencent's Hunyuan3D 2.0 is an open-source AI 3D generator that can generate 3D assets from 2D images with improved quality and speed. It uses two AI models: Hunyuan3D-DiT for generating 3D assets and Hunyuan3D-Paint for adding textures. The tool can also animate 3D models and has potential applications in game development, VFX, and ecommerce.

The Samsung Galaxy S25 series features a thinner and lighter design, with a powerful Qualcomm Snapdragon 8 Elite chip, 12 GB of RAM, and a 5,000 mAh battery. The series includes the S25, S25+, and S25 Ultra, with the Ultra model having a 200 MP wide and 50 MP telephoto lens. The phones also come with an 'Integrated AI Platform' that enables cross-app AI capabilities, allowing users to perform tasks such as searching for information and sending messages with a single prompt.

Nvidia Ingest is a microservice that processes document content and extracts metadata into a JSON schema. It supports PDFs, Word, and PowerPoint documents, and can extract structured information from tables, charts, images, and text using optical character recognition. The tool requires a number of supporting services and a minimum of two H100 or A100 GPUs with at least 80GM memory.

Samsung announced the Galaxy S25 series, including the S25, S25+, and S25 Ultra, with improved AI features, a new ProScaler AI tool, and a personalized AI-integrated OS. The phones feature a customized Snapdragon 8 Elite processor, improved cameras, and enhanced battery life. The S25 Ultra has a 6.9-inch display, 12GB of RAM, and a starting price of $1,300. The series also includes new features like Audio Eraser, AI Select, and Now Bar.

Samsung has unveiled its latest Galaxy S25 series, including the S25, S25+, and S25 Ultra, with improved AI capabilities, faster performance, and new camera features. The company also teased the Galaxy S25 Edge, a super slim device. Pre-orders are live now, with official sales starting on February 7.

LinkedIn is being sued for allegedly training its AI models on private InMail messages from its Premium customers. The lawsuit claims that LinkedIn breached its contractual promises by disclosing private messages to third parties for AI training. LinkedIn has denied the allegations, stating that they are 'false claims with no merit'. The case raises concerns about data privacy and the use of personal data for AI training.

Google DeepMind CEO Demis Hassabis expects clinical trials of AI-designed drugs to commence this year, with the potential to speed up development and reduce costs. AI can help predict protein structures and optimize personalized medicine, but it's not a replacement for human scientists. Pharmaceutical companies are interested in AI due to its potential to save time and money, with the average cost of developing a new drug being $2.6 billion.

Ransomware attacks reached a record high in December 2024, with 574 attacks detected by NCC Group. The most active threat actor was FunkSec, responsible for 103 attacks. Asia saw a 58% increase in attacks, and North America remained the most targeted region.

Oracle has released 603 patches, including 318 for its own products and 285 for Linux. A critical patch for the Agile Product Lifecycle Management Framework is highlighted, with a CVSS score of 9.9. Other patches address vulnerabilities in Oracle Communications, Financial Services, and Middleware products.

Two ransomware campaigns are abusing Microsoft Teams to infect organizations and steal data. The campaigns, tracked by Sophos, use a default Teams configuration that allows external users to initiate meetings or chats with internal ones. The attackers use spam emails and fake IT support calls to trick victims into installing malware, which can lead to data theft and ransomware deployment.