# Jun 30, 2026 ## Claude Sonnet 5 - **Claude Sonnet 5** is the most agentic model yet, capable of autonomous planning and tool use, outperforming its predecessor Sonnet 4.6 while being more cost-effective than Opus 4.8. ## US Supreme Court Just Blew Up EU-US Data Transfers - The **US Supreme Court's ruling** in _Trump v. Slaughter_ undermines the **independence of the FTC**, which has been crucial for EU-US data transfer agreements, prompting calls for the EU to withdraw its adequacy decision on US data protection. ## Ornith-1.0: Self-improving Open-source Models for Agentic Coding - **Ornith-1.0** is a **self-improving open-source model** for agentic coding, featuring architectures like **9B-Dense**, **31B-Dense**, **35B-MoE**, and **397B-MoE**, which excel in coding benchmarks such as **Terminal-Bench 2.1** and **SWE-Bench**. ## South Korea to Spend $1T on More Memory Chip Production and Humanoid Robots - South Korea plans to invest **$1 trillion** in memory chip production and humanoid robots, aiming for a **physical AI lead** and commercial deployment by **2028**. ## Nano Banana 2 Lite - **Gemini 3.1 Flash-Lite** offers **dramatically reduced latency**, enhancing workflow efficiency for users in real-time applications. ## Apple Neural Engine: Architecture, Programming, and Performance - The **Apple Neural Engine (ANE)** serves as a **fixed-function matrix accelerator** in Apple devices, with insights derived from reverse-engineering revealing its architecture and performance metrics across the A11 to A18 and M1 to M5 chip families, as detailed in the [arXiv article](https://arxiv.org/abs/2606.22283). ## Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding - **Ornith-1.0** introduces a **self-improving training framework** that enables models to generate their own scaffolds for coding tasks, enhancing solution quality through joint optimization. ## Popping the GPU Bubble - **Photon's inference engine** achieves near-realtime VLM inference (~33ms on NVIDIA B200) by optimizing GPU utilization, resulting in **up to 35% higher decode throughput** through techniques like _pipelined decoding_ that minimize idle time. ## Cerebras OpenAI Deal Capacity Has Effectively Killed the Waitlist for Everyone Else - **Cerebras' deal with OpenAI** has effectively monopolized its inference capacity, allocating **$20 billion** worth of chips to a single customer, leaving smaller startups in the lurch. ## I Built a mmWave Material Classification Radar - The author developed a **mmWave radar** capable of classifying materials, specifically targeting **asbestos detection**, to address the high costs and health risks associated with traditional testing methods. ## Micro-Agent: Beat Frontier Models with Collaboration Inside Model API - **Micro-Agent** technology enhances AI model performance by enabling **collaboration** within the serving layer, allowing multiple models to work together without altering their individual identities. ## Google's Agentic Peer-Reviewer Handled ~10K Papers at ICML/STOC — Formal Research Paper Now Out - **Google's agentic AI peer-reviewer** efficiently reviewed **~10,000 papers** at ICML/STOC, demonstrating a **34% increase** in catching mathematical errors compared to traditional methods. ## Claude Meets Blackwell Ultra: Anthropic’s Models Now Run on NVIDIA GB300 in Azure - **Claude models** are now available on **NVIDIA GB300 Blackwell Ultra** in Microsoft Azure, enabling enterprises to create **autonomous AI agents** tailored to specific domains. ## EML Trees are Universal Approximators - **EML trees serve as universal approximators**, enabling the representation of all elementary functions through composition, thus allowing for the approximation of any function in a general space with controlled size and depth. ## Open Memory Protocol – One Memory Store for Claude, ChatGPT, Curso - **Open Memory Protocol (OMP)** offers a **vendor-neutral standard** for AI tools to share user memory, addressing the **AI memory silo problem** by enabling seamless context transfer across different platforms and sessions.