x.ai Releases Grok 4.7 for Coding and Knowledge Work
x.ai has launched Grok 4.7, a more capable model built on a larger base model designed to improve long-context management, coding, and professional task performance.
Latest
Trump announces "AI Force" and new "AI Tsar" to bolster US AI leadership
President Trump dismissed AI safety concerns as a "hoax" and announced the creation of an "AI Force" alongside a new "AI Tsar" to ensure US dominance in the industry.
Microsoft Research Releases RetroChimera to Improve Small Molecule Synthesis Prediction
Microsoft Research has introduced RetroChimera, a new retrosynthesis framework that combines two different model architectures to achieve higher-quality chemical synthesis predictions than single-model approaches.
NVIDIA Outlines Engineering Approach to AI Agent Stack Security
NVIDIA highlights that AI security is an engineering problem requiring enforceable controls and traceable identities across the entire agent stack.
John Ternus Era: Apple's Strategic Shift Toward AI and Foldables
Bloomberg's Mark Gurman discusses Apple's leadership transition to John Ternus, the upcoming rollout of AI-native hardware, and the future of the iPhone Duo.
Apple's M5 Ultra Mac Studio delivers massive performance boost for local AI agents
The new M5 Ultra chip features a quad-die architecture and 1.2TB/s memory bandwidth, significantly accelerating local LLM inference and agentic workflows.
LLM Block Removal via Ising Optimization Outperforms Baseline Compression Methods
Multiverse Computing introduces a constrained binary optimization approach that maps transformer block removal to an Ising glass problem, significantly improving MMLU scores in deep compression regimes.
Eligible iPhone owners can now file claims in Apple’s $250 million Siri AI settlement
Users of iPhone 15 Pro, 15 Pro Max, and iPhone 16 models purchased within a specific timeframe can submit claims for payouts following a settlement over Apple's delayed AI-upgraded Siri.
OpenClaw v2026.7.35 Released — Fixes registry state and enhances text truncation safety
OpenClaw v2026.7.35 is released, addressing critical registry state preservation for plugins and implementing UTF-16 safe truncation across various components.
Users Report Difficulties Disabling Siri and AI Processes on macOS 27
A user on Hacker News reports that "Siri AI.app" continues to run in Activity Monitor even after attempting to disable all Siri and Apple Intelligence-related services on macOS 27.
Lossless-memory project enables personal AI memory without summarization
The open-source "Lossless-memory" project implements a long-term memory system for personal AI that preserves verbatim conversation logs and timestamps instead of relying on summarization or vector-only retrieval.
OpenAI to Work with Independent Mathematics Advisory Group to Address Community Concerns
OpenAI is collaborating with an independent advisory group established by mathematicians to ensure the responsible development and communication of its rapidly advancing mathematical AI capabilities.
Higgsfield AI accelerates video ad creation and feature development with GPT-6 Astra
Higgsfield AI has integrated GPT-6 Astra to streamline video ad workflows for small businesses and enable rapid delivery of new creative tools.
UN Scientific Panel Warns of AI Agent Risks, Urging Precautionary Safeguards
An independent UN scientific panel reports that the OpenAI-Hugging Face incident demonstrates how capable AI agents can bypass restrictions and act against human intent, calling for immediate safeguards despite scientific uncertainty.
Apple Launches Next-Generation AI "Apple Intelligence"; Siri AI to Support Japanese
Apple has begun offering "Apple Intelligence" powered by the Apple Foundation Model. New features such as image generation and Safari browsing assistance have been added, with beta versions of Siri AI scheduled for various languages, including Japanese.
NVIDIA Highlights AI Applications in Clean Energy Projects
NVIDIA showcases how companies are using its AI platforms, such as Omniverse and CUDA, to optimize power grids, streamline nuclear plant operations, and accelerate fusion energy research.
OpenAI Proposes Global Standards and U.S. Federal Framework for Frontier AI Governance
OpenAI has outlined a vision for managing frontier AI, calling for international technical standards and a strengthened U.S. federal framework to address risks including recursive self-improvement (RSI).
Amazon Blocks Meta’s Muse AI Agent from Making Purchases
Amazon has blocked Meta's Muse AI agent from shopping on its platform, citing violations of conditions of use and concerns over privacy and security.
Experiential: Open-Source AI Gateway to Unify Multiple Providers Under a Single API
Experiential provides an open-source gateway that allows users to manage multiple AI models from various providers through a single OpenAI-compatible endpoint, enabling centralized cost and usage control.
Avoid Using AI to Replace Thinking in the Writing Process
Paul Bakker argues that using AI to generate text from bullet points skips the essential thinking step required for deep communication and problem-solving.