The Archive

OpenAI updates mental health safety work: parental controls, distress detection, trusted contacts.

Dario Amodei discusses Anthropic's Department of War coordination regarding national security uses of Claude.

Google releases Nano Banana 2 image generation model with advanced world knowledge and subject consistency at fast inference speeds.

DraftNEPABench benchmark evaluates AI coding agents on federal permitting; shows 15% NEPA drafting time reduction potential.

OpenAI Codex and Figma integrate to enable seamless code-to-design iteration.

Cohere analysis of AI adoption in financial services: productivity gains, operational efficiency, and implementation pathways.

Cohere webinar on AI applications in financial services; generic promotional content.

Overview of AI use cases in banking sector covering automation and risk management; commentary without new findings.

Generative AI applications in finance sector with implementation guidance; marketing content without technical depth.

Anthropic acquires Vercept to enhance Claude's computer use and agentic capabilities.

LLM context lengths are exploding, and architectures are moving toward complex attention schemes like Multi-Head Latent Attention (MLA) and Grouped Query... LLM context lengths are exploding, and architectures are moving toward complex attention schemes like Multi-Head Latent Attention (MLA) and Grouped Query Attention (GQA). As a result, AI ”speed of thought” is increasingly governed not by the massive throughput of matrix multiplications, but by the transcendental math of the softmax function. Transcendentals refer to functions that cannot be… Source

OpenAI threat report examines malicious AI use combining models with websites and social platforms.

Anthropic releases Responsible Scaling Policy v3.0, detailing AI safety protocols for model development.

OpenAI appoints Arvind KC as Chief People Officer to lead organizational scaling and culture.

Anthropic publishes research on detecting and preventing model distillation attacks.

As the sizes of AI models and datasets continue to increase, relying only on higher-precision BF16 training is no longer sufficient. Key challenges such as... As the sizes of AI models and datasets continue to increase, relying only on higher-precision BF16 training is no longer sufficient. Key challenges such as training throughput expectations, memory limits, and rising costs are becoming the primary barriers to scaling transformer models. Using lower-precision training can address these challenges. By reducing the numeric precision used during… Source

Import AI 446 covers nuclear power LLMs, China's comprehensive AI benchmark suite, and measurement for policy.

OpenAI discontinues SWE-bench Verified due to contamination and training leakage; recommends SWE-bench Pro.

OpenAI launches Frontier Alliance Partners program to help enterprises deploy AI agents from pilots to production with secure, scalable infrastructure.

Claude Code Security, integrated into Claude Code, scans codebases for vulnerabilities and patches in limited preview.

OpenAI submits proof attempts to First Proof math challenge, demonstrating research-grade reasoning on expert-level problems.

NVIDIA flagship data center GPUs in the NVIDIA Ampere, NVIDIA Hopper, and NVIDIA Blackwell families all feature non-uniform memory access (NUMA) behaviors, but... NVIDIA flagship data center GPUs in the NVIDIA Ampere, NVIDIA Hopper, and NVIDIA Blackwell families all feature non-uniform memory access (NUMA) behaviors, but expose a single memory space. Most programs therefore do not have an issue with memory non-uniformity. However, as bandwidth increases in newer generation GPUs, there are significant performance and power gains to be had when taking into… Source

Google DeepMind releases Gemini 3.1 Pro, an upgraded model for complex reasoning and multi-step problem solving.

OpenAI commits $7.5M to The Alignment Project for independent AI alignment research addressing AGI safety and security.

OpenAI expands into India with local infrastructure, enterprise partnerships, and workforce development initiatives.

As AI workloads scale, achieving high throughput, efficient resource usage, and predictable latency becomes essential. NVIDIA Run:ai addresses these challenges... As AI workloads scale, achieving high throughput, efficient resource usage, and predictable latency becomes essential. NVIDIA Run:ai addresses these challenges through intelligent scheduling and dynamic GPU fractioning. GPU fractioning is wholly delivered by NVIDIA Run:ai in any environment—cloud, NCP, and on-premises. This post presents the joint benchmarking effort between NVIDIA and AI… Source

Python dominates machine learning for its ergonomics, but writing truly fast GPU code has historically meant dropping into C++ to write custom kernels and to... Python dominates machine learning for its ergonomics, but writing truly fast GPU code has historically meant dropping into C++ to write custom kernels and to maintain bindings back to Python. For most Python developers and researchers, this is a significant barrier to entry. Frameworks like PyTorch address this by implementing kernels in CUDA C++—either handwritten or by leveraging libraries… Source