OpenAI declares 'code red' as Google catches up in AI race
OpenAI's 'code red' signals a critical shift as CEO Sam Altman prioritizes enhancements to ChatGPT, reflecting the urgency of competition from Google and Anthropic.
Mistral 3 family of models released
Mistral 3 introduces a new generation of AI models, including Mistral Large 3, which features a sparse mixture-of-experts architecture with 41B active and 675B total parameters, enhancing performance across multimodal and multilingual tasks.
Amazon launches Trainium3
Amazon's new Trainium3 chip boasts a 3 nanometer architecture, delivering over 4x performance and 40% energy efficiency improvements compared to its predecessor, enhancing AI training and inference capabilities.
Apple Releases Open Weights Video Model
STARFlow-V is the first normalizing flow-based causal video generator, achieving visual quality comparable to diffusion models while enabling end-to-end training and multi-task support across various video generation tasks, including text-to-video, image-to-video, and video-to-video.
Anthropic: AI agents find $4.6M in blockchain smart contract exploits
AI agents have demonstrated the ability to exploit smart contracts, collectively uncovering vulnerabilities worth $4.6 million in simulated environments, highlighting the economic risks posed by these technologies.
Published paper uses hardcoded seed and collapsed model to report fraudulent results
A published paper misrepresented its findings by using a hardcoded seed and collapsed models, leading to inflated accuracy claims that do not reflect true model performance.
Claude 4.5 Opus' Soul Document
Claude 4.5 Opus' "Soul Document" is a 14,000 token overview that was used to train the model's personality during its training run, confirmed by Anthropic's Amanda Askell as a real document integral to Claude's development.
Comparing AWS Lambda ARM64 vs. x86_64 Performance Across Runtimes in Late 2025
AWS Lambda's arm64 architecture outperforms x86_64 in cost and performance, with Rust on arm64 being the most efficient combination, achieving up to 40% lower compute costs while maintaining superior speed across workloads.
Arcee Trinity Mini: US-Trained Moe Model
Arcee AI's Trinity Mini is a compact, end-to-end trained MoE model designed in the U.S., featuring open weights and advanced reasoning capabilities, with Trinity Large set to launch in January 2026, utilizing 2048 B300 GPUs for training.
Qwen3-VL can scan two-hour videos and pinpoint nearly every detail
Qwen3-VL by Alibaba can accurately analyze two-hour videos and identify specific frames with 99.5% accuracy, showcasing its advanced capabilities in handling extensive data loads.
Pose-free 3D Gaussian splatting via shape-ray estimation
SHARE introduces a pose-free framework for 3D Gaussian splatting, enabling efficient rendering without the need for precise camera poses, thus addressing common challenges in real-world applications.
Show HN: RunMat – runtime with auto CPU/GPU routing for dense math
RunMat offers a high-performance runtime for MATLAB-style math, automatically optimizing operations for CPU or GPU without requiring kernel code or rewrites, achieving performance that can surpass hand-tuned CUDA in many scenarios.
Gated Attention, a bit of schmidhubering/sociology of science
Gated Attention has emerged as a significant concept, particularly through its headwise gating mechanism, which applies a dense
[0,1] coefficient to each attention head, enhancing output mixing efficiency.
At NeurIPS, NVIDIA Advances Open Model Development for Digital and Physical AI
NVIDIA introduces the DRIVE Alpamayo-R1, the first open industry-scale reasoning vision language action model for autonomous driving, enhancing vehicle decision-making in complex scenarios through advanced AI reasoning techniques.
Polymathic release new scientific foundation model - paper shows it learns general abstract laws of physics
Polymathic AI's new model, Walrus, demonstrates an ability to grasp abstract physical concepts such as speed and diffusion, indicating a significant leap toward creating general-purpose scientific AI.