A forensic analysis reveals that several leading Chinese AI firms, including DeepSeek and Moonshot AI, have accessed NVIDIA's Blackwell GPUs despite export restrictions. Six recently released models show distinct patterns indicating training on this advanced hardware, raising questions about compliance with export controls in the AI sector.

Edge0-145B and Edge0-120B have been open-sourced, offering powerful AI models that run locally on personal devices. These models enhance coding, reasoning, and task completion while reducing the need for cloud connectivity and associated costs. The initiative aims to empower developers with new tools for innovation.

MIT researchers have developed a collaborative system that enhances algorithmic design through human feedback, improving manufacturability and structural integrity. By integrating human input into the topology optimization process, this approach aims to make complex designs more practical and accessible for engineers, fostering trust and creativity in design.

Jot, a Mac capture app, successfully transitioned from an Electron framework to Rust, reducing its size from 675 MB to just 91 MB. This shift not only streamlined performance but also retained cross-platform compatibility. The port includes improved speech recognition features and maintains file compatibility with previous versions.

This article explores the challenges of writing correct software, emphasizing the gap between theoretical concepts and practical programming. It reviews key findings from computability and complexity theories, highlighting that achieving software correctness is more complex than often suggested, involving significant trade-offs and limitations inherent to computation.

This article discusses the progress made in adapting the JIT compiler yk for the CPython interpreter. The author shares performance benchmarks demonstrating notable speed improvements over standard CPython and its experimental JIT. It reflects on challenges encountered and the process of enhancing interpreter efficiency without sacrificing compatibility.

Xing Mobility has unveiled its Immersio CTP battery pack, offering potential energy density increases for electric ships. With its immersion cooling design, the CTP claims to outperform traditional marine batteries. However, questions about safety, longevity, and real-world performance remain as the company seeks certification and conducts further testing.

This article reflects on the evolution of software engineering from simple systems to complex AI-driven environments. The author shares personal experiences in transitioning from coding to leveraging AI tools, emphasizing the need for engineers to adapt while enabling others to harness this technology effortlessly.

This article explores configuring a Broadcom PCIe switch on Linux to efficiently utilize four AMD Radeon R9700 GPUs with Resizable BARs. It discusses the challenges faced in achieving optimal memory mapping, shares practical solutions for BAR resizing, and highlights resulting performance improvements in data transfer rates among the GPUs.

After testing AWS Lambda Web Functions, Easel found that their own compute solution significantly outperformed it. With faster response times and reduced idle costs, they opted to manage their own infrastructure for greater control, better performance, and optimized resource allocation.

Gemini 4 Hydrogen is designed for rapid coding and agentic workflows, offering users speed and efficiency. This model outperforms its predecessors and competitors in various tasks, enhancing software development and everyday assistance with improved task execution and adaptability across multiple thinking levels, making it a versatile tool for developers and consumers alike.

This article details the author's journey of uncovering a long-standing vulnerability in PostgreSQL while exploring memory-corruption exploits. Focusing on the nuances of PostgreSQL's data types, the narrative illustrates how seemingly minor questions led to significant discoveries, shedding light on system vulnerabilities and their implications.

Singularity-Bench introduces an open-ended benchmark designed to evaluate AI agents' autonomy in developing systems. It assesses their performance on six tasks while revealing vulnerabilities in existing frameworks. Results indicate significant potential for certain agents to match or surpass human engineering capabilities, raising questions about their future role in AI development.

This article presents DashBench, a benchmark designed to evaluate Video Large Language Models on corner cases in autonomous driving. It highlights significant performance gaps in these models when analyzing rare but critical driving events, underlining the need for improvements in safety and explainability for autonomous systems.

This article explores a new algorithm called quantum tunneling search (QTS) designed for non-convex optimization challenges in machine learning. By utilizing quantum tunneling on photonic chips, the research demonstrates how QTS can efficiently navigate complex landscapes to identify global minima, outperforming traditional methods.