## 🔍 Key Insights **Gemini Embedding 2** establishes a unified multimodal embedding space; **Claude Code** introduces the revolutionary `/btw` side-conversation mechanism; and **Lingchu Intelligence** secures ¥2 billion in funding, with its valuation surging sevenfold in just one year—embodied AI and AI-agent infrastructure are rapidly transitioning from experimental prototypes to large-scale deployment. ## 🚀 Highlights - **Gemini Embedding 2 Launch**: Google unveils a native, unified multimodal embedding space supporting text, images, audio, video, and documents—dramatically simplifying RAG and AI-agent development. - **Claude Code Adds `/btw` Command**: Enables users to initiate **non-interruptive side conversations** while the AI executes long-running tasks—significantly enhancing interaction naturalness and debugging efficiency. - **Lingchu Intelligence Closes ¥2 Billion Funding Round**: Leveraging its proprietary **Psi-SynEngine**, the company pioneers a “human-centric” embodied AI data collection paradigm and a universal operational brain—earning strong backing from national-level capital. - **Andrej Karpathy Demonstrates AI Self-Optimization**: His autonomous research agent identified over 20 improvements, boosting **GPT-2 training performance by 11%**, validating the feasibility of agent-driven model evolution. - **Browser Use Launches Native Slack Bot**: Enables end-to-end **autonomous workflows and scheduled tasks directly within Slack**, deepening AI-agent integration into enterprise communication layers. - **Anthropic Skills Generator Undergoes Major Upgrade**: Introduces four industrial-grade capabilities—**evaluation systems, benchmarking, parallel testing, and description tuning**—accelerating scalable AI skill production. - **Veeso AI Turns URLs into Editable Design Files**: Generates layered HTML + image structures, enabling pixel-perfect editing of text and UI elements—defining a **new paradigm for AI-powered design**. - **Fish Audio Releases Open-Source S2 TTS Model**: Delivers **ultra-low latency, multi-speaker support, and fine-grained voice control**, elevating open-source TTS to production-ready capability.