Machine Learning
Ongoing coverage on Machine Learning.
Latest coverage
AMD's FSR 4.1 upscaling is coming to Xbox Series X|S
Microsoft and AMD have announced that AMD's FSR 4.1 upscaling technology will be coming to Xbox Series X|S consoles soon. This machine-learning-based temporal upscaler renders games at a lower resolution and reconstructs them at a higher one. While FSR 4.1 will also support various PC hardware and handhelds, its implementation on Xbox Series X|S will require developers to patch their games, and gains may not match dedicated hardware like the PS5 Pro.
Xbox follows in PlayStation's footsteps, introducing AMD FSR 4.1, targeting better graphics and performance on Series X/S
Microsoft announced that AMD's FSR 4.1 image upscaling technology will be coming to Xbox Series X and Series S consoles, as well as handhelds like the ASUS ROG Ally. This AI-driven technology renders games at a lower resolution and upscales them for improved visuals and performance, potentially extending the life of current-generation hardware. The move follows Sony's recent introduction of its own upscaling technology, QSSR, for the PlayStation 5.
Playing God: The revolutionary AI that made Black & White still unforgettable
This article reflects on the revolutionary machine learning AI in the 2001 game Black & White, developed by Lionhead Studios and directed by Peter Molyneux. It highlights how the AI, coded by Demis Hassabis and Richard Evans (who later founded DeepMind), allowed creatures to learn and imitate player actions, a feat still considered remarkable. The piece contrasts the game's joyful sandbox approach with modern AI anxieties and praises its unique design.
I found a use of AI that doesn't make me cringe and shudder, and it's in the mic everyone's waited over a…
Logitech has released the Yeti 2 microphone, incorporating machine learning-aided features for voice following and noise reduction. Despite initial skepticism towards AI in products, the reviewer found these features genuinely useful and impressive, especially considering the $160 price point. The new microphone builds upon the classic Yeti design with an OLED screen and enhanced customization options.
Mecha Mayhem – Beta Sign Up | Alpha Beta Gamer
Mecha Mayhem, a physics-based robot fighting game, is opening beta sign-ups. Players control a customizable maintenance droid that can salvage and equip enemy parts, exploring a persistent roguelike world map solo or in co-op. The game also features a freeplay mode with AI opponents trained using machine learning.
Microsoft file patent for a machine learning system that would "train" itself on your playtime to work out when you'll be least annoyed by in-game ads
Microsoft has filed a patent for a machine learning system designed to optimize in-game advertising delivery. The system aims to analyze player behavior and game state to serve ads during moments least likely to disrupt the user experience, potentially using a credit system to manage ad frequency. This development is seen as a potentially intrusive step in game monetization, contrasting with current practices like ad billboards.
The High-Tech Tools Driving the New Casino Boom
The online casino industry is experiencing a boom driven by technological advancements such as real-time streaming, machine learning, and scalable cloud architecture. These technologies enhance user experience, improve security, and streamline operations, leading to increased player engagement and trust.
Nvidia puts its money where its mouth is on open weight AI, moving to buy Hugging Face for $13 billion
Nvidia has announced its intention to acquire Hugging Face, a leading open platform for machine learning applications, for approximately $13 billion. Nvidia CEO Jensen Huang stated that Hugging Face will remain an open platform, supporting various models, frameworks, and computing platforms, emphasizing Nvidia's commitment to open-source AI development.
AI is 'just a plagiarism machine that makes people lose their jobs' in the eyes of The Talos Principle…
Narrative designers Jonas and Verena Kyratzes from Croteam express strong skepticism about the current applications of AI in game development, calling it a 'plagiarism machine' that harms creativity and leads to job losses. They argue that while the underlying technology is impressive, its current use in profit-driven systems is detrimental, and the focus should be on labor issues rather than the technology itself.
For an unhinged 84 minutes, Slay the Spire 2 expert contemplates 1 damage and explains a fight simulator built to solve…
YouTuber Jorbs, known for his expert Slay the Spire 2 analysis, has developed a complex fight simulator using machine learning to optimize in-game encounters. The tool, built in Sandbox, allows players to input cards and enemy data to calculate optimal play lines, though its complexity makes it impractical for most situations. Jorbs acknowledges the excessive nature of the tool but finds joy in the granular analysis and the pursuit of optimizing gameplay.
Alibaba Opens Wan3.0 Beta: 30-Second AI Video Built From Text, Images — Even PDFs
Alibaba has launched Wan3.0, a new AI video generation model available in beta, capable of producing 30-second videos from text, images, PDFs, and web pages. The model aims to improve detail preservation and consistency compared to its predecessor, Wan2.5, and offers a Standard tier with a 30 percent discount. This development is part of Alibaba's significant investment in AI, despite a recent decline in quarterly profits.
Netflix pitted an LLM against its own feature-engineered recommender — and the LLM won
Netflix has developed a new recommendation system called GenRec, which utilizes a large language model (LLM) and has outperformed its previous feature-engineered system by 1.6 percent in offline ranking quality. This new system requires significantly less training data and is more adaptable to new content formats like games and live programming, though it still requires a separate component to handle hallucinations and ensure only existing catalog items are recommended.
Borrowed From IQ Testing, A New Method Finds Deep Cracks In AI Safety Benchmarks
A new study by researchers, including those from the UK AI Security Institute, has revealed significant flaws in current AI safety benchmarks for language models. Borrowing methods from psychological testing, the analysis found that models can improve their safety scores by indiscriminately refusing requests, rather than becoming genuinely safer. The study also highlighted issues with benchmark efficiency and the potential for models to 'sandbag' or intentionally perform differently during evaluations.
WhatsApp Adds On-Device AI Scam Warnings, and Nothing Leaves Your Phone
WhatsApp is rolling out an optional Scam Alert tool that uses on-device machine learning to identify suspicious messages. The feature warns users of potential scams without sending message content off the device for classification. Users can mark chats as trusted to disable alerts or opt-in to share recent messages to help improve the tool's accuracy.
Sony is Testing PS5 Pro Update That Enables Enhanced PSSR By Default – Rumor
Sony is reportedly testing a beta system software update for the PlayStation 5 Pro that will enable the "Enhance PSSR Image Quality" feature by default for all supported games. While users can still disable the feature, it will be automatically applied to games that support even the older version of PSSR. This move suggests Sony's growing confidence in its AI-based image upscaling technology, which has seen improvements and is being developed in collaboration with AMD.
AI model distillation, explained: the cheap way to inherit an expensive model’s brain
AI model distillation, also known as knowledge distillation, is a technique where a smaller "student" model learns from a larger "teacher" model. This process allows for the creation of cheaper, more efficient AI systems by inheriting the capabilities of expensive, pre-trained models. However, concerns exist regarding the potential loss of safety knowledge and the inheritance of biases during this distillation process.
Two AI Commercials, Two Futures: Anthropic Sells Fear, Alibaba Sells Free Time
Alibaba's new Qwen3.8-Max AI model, with 2.4 trillion parameters, demonstrates impressive long-horizon agentic task capabilities, operating autonomously for nearly five days on an experiment. Priced significantly lower than competitors like Anthropic's Fable 5 and OpenAI's GPT-5.6 Sol, Alibaba's model is poised to disrupt the market, with some users already switching from Western AI subscriptions. The article contrasts Alibaba's commercial, which emphasizes free time and leisure, with Anthropic's fear-based advertising, reflecting differing public sentiments towards AI in China and the US.
The Pokemon Company Announces Pokemon TCG AI Battle Challenge, Fully Exploring Controversial Technology
The Pokemon Company has announced the Pokemon TCG AI Battle Challenge, a competition where players train AI bots to compete in the Pokemon Trading Card Game for cash prizes. Organized in partnership with HEROZ and the Matsuo Institute, with support from Google, NVIDIA, and Kaggle, the event aims to explore AI's strategic capabilities in complex card games. The competition has sparked controversy due to its public endorsement of AI technology, which contrasts with Nintendo's human-first approach, and concerns about using the event for AI training.
New Arm tech enables the use of Unreal Engine MegaLights on mobile
Arm has unveiled its new neural graphics technology, demonstrated in the mobile game Neural Dawn developed by Sumo Digital. This technology enables the use of Unreal Engine MegaLights and ray-tracing on mobile devices for the first time, previously only available on PC and consoles. The Arm Mali GPU, incorporating machine learning, aims to deliver desktop-class visuals while preserving battery life, allowing for longer gaming sessions and dynamic lighting previously unachievable on mobile platforms.
AMD bringing FSR 4.1 ML-powered upscaling to RX 7000-series in July, RX 6000-series in 2027
AMD announced that FSR Upscaling 4.1 will be released for Radeon RX 7000-series graphics cards in July, with support for RX 6000-series cards planned for 2027. This update brings machine learning-powered upscaling to a wider range of AMD hardware, though full ML-powered features like Frame Generation remain exclusive to newer RDNA 4 cards.
