NerdNewsJanuary 31, 2025 |
News & Trends
Microsoft adds DeepSeek R1 to Azure AI Foundry and GitHub
Microsoft has added the open-source DeepSeek R1 large language model to its Azure AI Foundry and GitHub platforms. This move comes after DeepSeek R1 was found to be efficiently trained and capable of reasoning before answering queries, potentially disrupting the US tech industry. Microsoft plans to bring Distilled DeepSeek R1 models to Copilot+ PCs, with the first release available soon. The company has also undergone rigorous testing and evaluation to mitigate potential risks associated with the model.
DeepSeek stirs intrigue and doubt across the tech world
DeepSeek, a Chinese AI company, has made claims of building its latest R1 model at a fraction of the cost of rivals, stirring intrigue and doubt across the tech world. The company's claims have been met with skepticism, with some questioning the validity of its open-source credentials and the potential for gaming the market. Despite this, the company's arrival has sent shockwaves through the industry, with Nvidia's stock value plummeting by $600 billion.
Nvidia's latest GPU driver lets you activate DLSS 4 in games and apps
Nvidia has released a new GPU driver that allows users to activate DLSS 4 in games and apps, even if they don't currently support it. The driver also includes a new override feature, improved image quality, and reduced VRAM usage. Additionally, Nvidia has updated its RTX Video Super Resolution feature to upscale old blurry YouTube videos and introduced a new Smooth Motion feature for RTX 50-series owners.
Google blocked 2.36 million risky Android apps from Play Store in 2024
Google blocked 2.36 million Android app submissions to the Play Store in 2024 due to policy violations. 158,000 developer accounts were banned for attempting to publish harmful apps. AI-assisted human reviews helped block 92% of violating cases.
Microsoft's Surface Pro and Laptop for Business Get Intel's New AI PC Chips
Microsoft's Surface Pro and Laptop for Business are getting Intel's new Core Ultra 200V AI PC processors, offering improved AI performance, faster multitasking, and better battery life. The new chips feature a larger 48 TOPS NPU, allowing for more complex AI tasks. The updated devices will be available on February 18, starting at $1,500. |
Options & Tutorials
Apple New In-App Purchases API for Creator Platforms and Large Catalogs
Apple's new Advanced Commerce API provides iOS developers with more flexibility to manage large content catalogs, creator experiences, and subscriptions. The API allows apps to host and manage their catalogs outside of the App Store, while using the App Store commerce system for payment processing and customer service. Developers must request access to use the API.
Rethinking Medallion Architecture
The medallion architecture is flawed due to its inefficiency, expense, and difficulty in reusing data. A shift left approach can improve data accessibility by doing data processing upstream, making data products the backbone of this strategy. Data contracts provide a formal agreement for data products, ensuring healthy data communication across the organization.
OpenAI partners with Los Alamos National Lab
OpenAI has announced a partnership with Los Alamos National Laboratory to deploy its o1 LLM on the lab's Venado supercomputer. The collaboration aims to tackle disease research, high-energy physics, and nuclear security. The o1 model will be accessible to researchers at Lawrence Livermore National Laboratory and Sandia National Laboratory, with a focus on reducing the risk of nuclear war and securing nuclear materials.
Who Wins the AI Agent Battle
The article discusses the emergence of AI agents, which are AI software that can execute tasks on behalf of users. OpenAI's Operator is a notable example, allowing users to give tasks and watch as the agent completes them. The success of AI agents depends on context and user interface. Big tech companies have an advantage due to their existing user base and data. The article explores the concept of AI agents, their potential impact, and the challenges they pose.
New Syncjacking attack hijacks devices using Chrome extensions
A new attack called 'Browser Syncjacking' allows attackers to hijack devices using a malicious Chrome extension, giving them access to stored data, passwords, and browsing history. The attack involves creating a managed Google Workspace profile, tricking the victim into installing the extension, and then using the extension to inject malicious content and take control of the browser. |
Launches & Tools
OpenAI-backed 1X acquires Kind Humanoid
OpenAI-backed 1X has acquired Kind Humanoid, a robotics startup that focuses on creating humanoids. The acquisition marks a key consolidation in the humanoid robotics industry. Kind Humanoid's expertise complements 1X's mission to create safe and intelligent humanoids. The deal's details have not been disclosed, but it's expected to help 1X expand its Bay Area operations.
Intel cancels AI chip, talks painful past and simplified future
Intel has canceled its next big AI chip, codenamed Falcon Shores, and will focus on simplifying its business. The company reported a $126 million quarterly loss on $14.3 billion in revenue. Intel's foundry business lost $13.4 billion in 2024, but the company expects financial improvements next year. Intel will prioritize areas where it can drive differentiated value and fight for every socket where an Intel chip can go.
OpenAI Partners with Los Alamos National Laboratory
OpenAI is deepening its partnership with the Los Alamos National Laboratory to reduce the risk of nuclear war and secure nuclear materials. The partnership will utilize OpenAI's latest o-series models on Los Alamos' Venado supercomputer to tackle problems such as disease treatment, cybersecurity, and energy leadership.
Cerebras becomes the world’s fastest host for DeepSeek R1
Cerebras Systems will host DeepSeek's R1 AI model on US servers, achieving speeds 57 times faster than GPU-based solutions. This breakthrough addresses concerns about China's AI advancement and data privacy, offering American companies a way to leverage AI advances while maintaining data control. Cerebras' wafer-scale technology eliminates memory bottlenecks, delivering 1,600 tokens per second and outperforming traditional GPU implementations.
U.S. Copyright Office says AI-generated content can be copyrighted
The U.S. Copyright Office has clarified that AI-generated content can receive copyright protection if a human substantially contributes to or edits it. This means that human creativity remains central to copyright law, and AI tools can be used in the creative process without disqualifying a work from copyright protection. The office outlines three key scenarios where AI-generated material can apply for copyright, including when human-authored content is incorporated into the AI output, when a human significantly modifies or arranges the AI-generated material, and when the human contribution is sufficiently expressive and creative. |
Quick Links
Apple reports best-ever earnings despite iPhone revenue slip
Apple reported its best-ever earnings with $124.3 billion in revenue, up 4% from last year. However, iPhone revenue slipped, despite the introduction of Apple Intelligence features. The company's Mac, iPad, and services divisions saw revenue increases, while wearables revenue was slightly down. CEO Tim Cook attributed the strong performance to the holiday season and the popularity of new Macs and services.
DeepSeek's New AI Model Sparks Shock, Awe, and Questions From US Competitors
DeepSeek, a Chinese startup, has developed a powerful open-source AI model that has shaken Silicon Valley. The model, called R1, has impressive capabilities and was developed on a relatively small budget. This has sparked talk of a potential upheaval in the tech industry, with some experts saying it signals a shift towards more democratized AI development. The model's performance and efficiency have also prompted talk of cost-cutting at big tech firms, with some considering using DeepSeek's techniques to reduce their own expenditure on AI.
KuCoin to Pay Nearly $300 Million in Penalties
KuCoin's operator pleaded guilty to operating an unlicensed money-transmitting business and agreed to pay $297 million in penalties. The cryptocurrency exchange failed to implement anti-money laundering requirements, allowing cybercriminals to launder billions of dollars.
Police Seize Cracked and Nulled Hacking Forum Servers, Arrest Suspects
Europol and German law enforcement seized 17 servers and arrested suspects linked to Cracked and Nulled, two large hacking forums with over 10 million users, known for cybercrime and password theft. The U.S. Justice Department reported that Cracked generated $4 million in revenue and impacted 17 million victims in the U.S.
Samsung Galaxy S25 Ultra review: Too much AI, not enough Ultra
The Samsung Galaxy S25 Ultra features a new ultra-wide sensor, improved AI capabilities, and a faster chip, but its hardware and design are similar to previous models. The phone's battery life is great, but its Qi2 'readiness' is limited. With a starting price of $1,300, the S25 Ultra's value proposition is questionable, especially considering that its biggest advancements are software-based. |
Share NerdNewsShare your affiliate link to get commission!
Thanks for reading,
If you dont want to receive future editions of NerdNews, |