NVIDIA released OpenShell, an open source runtime that isolates autonomous AI agents in per-agent sandboxes with kernel-level file and syscall controls and a policy gateway that checks every outbound …
AeroJAX, a JAX-native computational fluid dynamics framework released on GitHub by developer arriemeijer-creator, lets researchers modify solver parameters, boundary conditions, and immersed geometrie…
OpenAI, Red Hat and NVIDIA are backing OpenClaw Enterprise, an MIT-licensed, vendor-neutral open-source control plane for managing persistent AI agents in enterprises, announced September 29, 2026, wi…
NVIDIA researchers, with MIT and the University of Oxford, introduced Physis-Lang, a self-evolving physical-language framework that ranks first on the Physics-IQ Verified leaderboard snapshot dated Se…
The OpenClaw Foundation launched OpenClaw Enterprise (OCE) on September 29, a free, open-source control plane for running persistent AI agents in multi-user enterprise environments, adding multi-tenan…
On NVIDIA Blackwell B200 GPUs, tensor-core throughput exceeds special-function exponential throughput by more than two orders of magnitude, exposing exponential evaluation as a bottleneck in fused att…
NVIDIA has released OpenShell, an alpha-stage runtime (version 0.1.1) that places the security boundary outside autonomous AI agents rather than inside them. The system combines a containerized sandbo…
NaiveAI's Naive-N0.5-Flash, a 309B-parameter Mixture-of-Experts model with 15.5B active parameters per token, requires FP8-capable NVIDIA GPUs and roughly 315GB of weights to run locally through Huggi…
The OpenClaw Foundation released OpenClaw Enterprise on September 29th, an open-source control plane for deploying and governing persistent AI agents on an organization's own infrastructure, and said …
Baseten put Carbon, a new agent-execution platform built on the Blaxel sandbox infrastructure it acquired, into private preview on September 28th, adding support for NVIDIA's OpenShell runtime. Carbon…
A developer released TurboGPT, an MIT-licensed byte-level GPT trainer written in CUDA C++ that trains a 22KiB transformer in 13 seconds. The tool builds on Windows with Visual Studio 2022 C++ tools an…
NVIDIA released TensorRT Model Connect, an open source collection of C++ AI model reference implementations built on NVIDIA TensorRT, covering 128 model families tested on NVIDIA GB300 as of its publi…
NVIDIA released VSS Blueprint 3.3, adding a Build Vision Agent skill (vss-build-vision-ai) and Adaptive Efficient Video Sampling (EVS) that cut VLM input tokens by 80% for a 60-minute summary and rais…
NVIDIA released Kumo Tabular, an open foundation model for tabular classification and regression, on Hugging Face under the OpenMDW-1.1 commercial license, claiming first place on the TabArena, Beyond…
Off Grid AI has released OGAD (Off Grid AI Desktop), a local AI app that runs smaller models on existing consumer hardware for bounded tasks like rewriting, summarizing and document Q&A. The company's…
NVIDIA published its Open Agent Safety Platform on 28 September 2026, a two-part system pairing the Apache 2.0 OpenShell runtime — built by the Gretel team NVIDIA acquired in 2025 — with NVIDIA Sentry…
NVIDIA released Isaac ROS 5.0, adding AI agent skills for stereo calibration, pose estimation, and code migration while removing the NITROS framework. The release changes what robotics developers must…
ZML released zml/llmd 20260929.0, adding support for deepseek-ai/DeepSeek-V4.1-Flash on NVIDIA and AMD GPUs, the project's first release supporting a frontier model. The release adds 2 model families,…
NVIDIA announced the Open Agent Safety Platform on 28 September 2026, comprising NVIDIA OpenShell, an open-source runtime that sandboxes AI agents, and NVIDIA Sentry, an out-of-band watchdog on BlueFi…
Independent developer Brendan Long released gpuc, an open-source single-user GPU queue that coordinates jobs across local, SSH-accessible remote, and rented RunPod GPUs without requiring sudo, publish…