Pulse

Apple has unveiled a preview of its revamped AI-powered Siri assistant, demonstrating its cross-device functionality on iPhone, iPad, and Mac. This update aims to enhance Siri's capabilities and integration across the Apple ecosystem. AI

GitHub Copilot is deprecating its older GPT-5.2 and GPT-5.2-Codex models. This change indicates a move towards newer, likely more capable AI architectures within the Copilot ecosystem. Users relying on these specific models should prepare for the transition to updated versions. AI

VASO, an agentic security scanner, has been released with new updates. This tool utilizes AI agents to perform security scans. The project has also published a new article detailing its capabilities, marking the third part of a series on the scanner. AI

CivicRadar is a new AI-powered tool designed to help legal teams analyze legislation. It automatically reviews bills to identify potential legal conflicts, constitutional issues, and fiscal impacts. The goal is to reduce review time and enhance transparency in the legislative process. AI

Apple is positioning its new AI features, branded as Apple Intelligence, as "Domesticated AI." This approach emphasizes practical, personalized responses using an individual's Apple ID context, akin to conversing with a trusted family member. In contrast to the more unpredictable and potentially costly AI offerings from competitors like Google, Apple aims for secure, easily understandable, and family-shareable AI services. AI

NVIDIA's Isaac Lab is enabling RAI Lab to conduct millions of virtual simulations for humanoid robot development, specifically for the Atlas robot. This approach bypasses the need for direct physical mastery of robot locomotion, suggesting a potential pathway for Europe to regain a leading position in the AI and robotics industry. AI

Researchers have developed a new method for generating 3D CAD models from text descriptions using large language models. This approach allows for controllable and faithful conversion of textual prompts into precise, usable 3D designs. The system aims to bridge the gap between natural language instructions and the complex requirements of computer-aided design. AI

OpenAIRE has outlined its strategic priorities for 2026-2028, emphasizing the need for practical implementation of Open Science in the face of evolving research landscapes. Key focus areas include developing trusted infrastructures for scholarly communication, promoting responsible AI within research, and enhancing Open Science monitoring. The strategy also addresses skills development, capacity building, and navigating legal and policy challenges. AI

A new paper explores how agent harnesses can revolutionize agentic search, drawing parallels to the impact of grep on text processing. The research proposes that these harnesses can significantly enhance the efficiency and effectiveness of agent-based search systems. This approach aims to redefine how agents interact with and retrieve information. AI

Meta has introduced "Workforce Academy," a new five-week training program designed to equip individuals with the skills needed for data center construction. This initiative, a collaboration with CBRE and Associated Builders and Contractors, offers free training and guarantees employment at a Meta data center construction site upon completion. AI

Artificial intelligences have demonstrated notable advancements in solving complex mathematical problems. Research from Ruhr University indicates that AI models are now capable of tackling previously intractable mathematical challenges. AI

Large language models have reportedly solved several difficult Erdős mathematical problems, with mathematician Terence Tao verifying the solutions. This development challenges the notion that LLMs are limited to non-original tasks and suggests a potential future where AI can be relied upon for complex intellectual endeavors. The success raises questions about the difficulty of these problems and the evolving capabilities of AI in creative and analytical domains. AI

The article explores the necessity of non-linearity in neural networks, arguing that it is crucial for handling the complex, non-straightforward nature of real-world data. It posits that activation functions like Softmax are essential for introducing this non-linearity, enabling models to learn intricate patterns and make sophisticated decisions. AI

AI models possess a form of statistical taste, enabling them to generate aesthetically pleasing and well-proportioned designs like landing pages. However, this reliance on existing data patterns means AI-generated content tends to represent the average or AI

Microsoft has developed a new AI system that can generate detailed captions for images, significantly improving efficiency in computer vision tasks. This advancement focuses on transforming raw pixel data into meaningful knowledge by leveraging context and sophisticated models. The system aims to make image understanding more accessible and powerful. AI

The Claude Code Explore sub-agent is effective for initial context gathering in AI development, but struggles with the complexities of the Spring framework. This article details how to train the agent to better understand Spring applications, enabling more accurate initial analysis. The goal is to improve the quality of generated code and solutions by accounting for Spring's specific features and ecosystem. AI

A new service called 'Splitting Stitches' has been launched, offering to analyze user arguments and generate passive-aggressive haiku responses. This unique tool aims to provide an indirect and AI

An experimental Python library called groundy has been developed to help Large Language Models (LLMs) avoid hallucination. Groundy works by posing a question in multiple ways and analyzing the semantic agreement of the responses. If the agreement is high, the LLM provides an answer; however, if the agreement is low, it signals a potential hallucination and refuses to answer. This tool is designed as a drop-in replacement and also includes a command-line interface for easy use, though it measures self-consistency rather than absolute truth. AI

ALEPH is a new AI runtime designed for embedded hardware, drawing inspiration from biological immune systems for security. It features a SHA256 whitelist, stateful iptables, and an anomaly classifier to differentiate between inference loads and denial-of-service attacks. The system operates without cloud connectivity, pre-trained weights, or large language models, and has reportedly run for over 407,000 ticks without any crashes. AI

AI benchmarks are rapidly becoming outdated, with new, more challenging benchmarks emerging approximately every 18 months. This cycle is driven by the intense competition in AI research and model development, which continuously demands updated evaluation metrics. The observation highlights the fast consumption rate of AI evaluation standards. AI

Nigel Farage has urged X (formerly Twitter) to take down deepfake advertisements featuring him and the Bank of England governor. The AI-generated videos depict confrontational scenes between Farage and Andrew Bailey, mimicking the style of the BBC's "Question Time" program. These misleading ads have recently appeared on the social media platform. AI

Seeed Studio has released the SenseCAP T1000-E, a tracker card that offers an alternative to Apple's AirTag. Unlike most trackers, it does not rely on Apple's Find My network or Google's network. Instead, it utilizes a different approach to location tracking, though the specifics of its network are not detailed in this item. AI

Jan.ai is a free and open-source application designed for running local large language models. It offers functionality comparable to LM Studio, allowing users to select from a wide variety of free models. The app aims to provide an alternative to larger, proprietary AI platforms. AI

A new Python CLI tool called denoise has been developed to remove background noise from videos without re-encoding the video stream. This approach extracts the audio, processes it with an AI model for noise reduction, and then remuxes the cleaned audio with the original video. The tool supports multiple denoising backends, including DeepFilterNet3, noisereduce, and FFmpeg's RNNoise, offering users flexibility based on their specific needs and available hardware. AI

Anthropic's Claude Code is a new tool designed to assist developers in various software development tasks. It can generate code, fix bugs, run tests, refactor projects, and understand entire codebases using natural language commands. The tool aims to streamline the software development process. AI

A judge has dismissed a legal case after discovering that both legal teams involved were using AI tools. The judge expressed concern that the AI was not present when they attended law school, implying a lack of familiarity or trust in the technology. This situation led to the cancellation of the trial and the removal of all parties from the case. AI