ToolsService provider doubled usage limits for individual and team subscription packagesAccording to the provider's announcement, they doubled the usage limits of subscription packages for individuals and teams, affecting Grok, Composer and the new Cursor models; the change is effective…1 min read
ToolsNVIDIA's Vera CPU and Olympus Core Optimized for Single‑Thread Agentic AI PerformanceNVIDIA designed the Vera CPU around the Olympus core to address agentic AI workloads that demand strong per‑thread progress under high concurrency.6 min read
ToolsPrince Canuma releases Nativ: a macOS app to run MLX models locallyPrince Canuma, author of the MLX-VLM Python library, has released Nativ — a native macOS desktop application that wraps MLX to run AI models locally.2 min read
ToolsAnthropic on Claude Code, Claude Tag and Fable: changing engineering, review and safety practicesAnthropic engineers described how Claude Code, Claude Tag and Fable have reshaped day-to-day product work, review processes and internal collaboration.5 min read
ToolsJack Dorsey launches Buzz, an open-source group chat that integrates human and AI agentsJack Dorsey and his company Block have introduced Buzz, an open-source group chat platform that embeds AI agents alongside people and aims to combine workflows now split across tools like Slack and GitHub.2 min read
ToolsGrabette: an open handheld gripper system to record robot manipulation datasetsGrabette is an open, low-cost handheld gripper system that lets anyone record real-world manipulation demonstrations and convert them into robot-ready datasets via an online processing pipeline.5 min read
ToolsAnthropic offers percentage-based support in Claude credits to researchers of rare diseasesAnthropic has launched an application round under the AI for Science program: it is offering researchers up to $50,000 in Claude usage credits to support projects that accelerate cures for rare diseases.1 min read
ToolsNVLink: purpose-built scale-up networking for low-latency, high-bandwidth AI factoriesNVIDIA’s NVLink scale-up networking fabric is designed to enable large, low-latency GPU domains needed by modern AI workloads such as trillion-parameter models and mixture-of-experts (MoE) architectures.6 min read
ToolsNVIDIA’s ovrtx brings RTX sensor simulation as an embeddable C/Python SDK for OpenUSD workflowsNVIDIA released ovrtx as a prerelease SDK on GitHub, a lightweight C and Python library that lets developers integrate RTX-based sensor simulation (camera, lidar, radar, semantic outputs) into existing OpenUSD application loops.5 min read
ToolsOpenAI retires standalone ChatGPT Atlas browser, folds features into ChatGPT desktop appOpenAI will discontinue the standalone ChatGPT Atlas browser and migrate its AI-driven browsing and agentic features into the new ChatGPT desktop application and other products.2 min read
ToolsTweet: proposal for a strategy that uses every model at onceA Twitter message briefly proposes: "Bet on every model at once" — that is, to apply all models simultaneously and in parallel.1 min read
ToolsAnthropic's Claude Code switches to a Rust implementation of Bun, improving Linux startup timeDeveloper Jarred Sumner reported that Claude Code v2.1.181 (released June 17) and later use a Rust port of the Bun runtime, yielding roughly a 10% faster startup on Linux.2 min read
