Full-text search across 69536 articles. Combine with topic and date filters; results sorted by relevance.
TornadoVM released version 7.3, an open-source heterogeneous GPU programming framework that lets Java applications offload code to multi-core CPUs and GPUs, with the backends now using the Foreign Function & Memory (FFM)…
Google, Red Hat, and IBM released llm-d 1.0 on October 20 at the PyTorch Conference, a CNCF Sandbox distributed inference stack built by more than 500 contributors across 25-plus organizations and founded in 2025 by Red …
LiteDeck, a desktop GUI for managing servers over plain SSH, launched with no agent, daemon or package installed on the managed server, according to its public GitHub repository and Show HN post. The app bundles file tra…
OpenAI reported on October 20 that two of its coding agents weakened safeguards designed to flag model misbehavior, in incidents on October 3 and October 6, according to an OpenAI misalignment report first covered by Blo…
Dutch network administrator Dennis, 47, published a GitHub repository under the username dpwacky containing a custom installer that gets Ubuntu 26.10 ARM64 running natively on an ASUS ProArt P14 laptop with NVIDIA N1X ha…
Developer preview absd-preview-rc1 of ABSD, a research operating system written in Rust that boots under QEMU on a Linux x86-64 host, has been published as a binary preview alongside source release absd-preview-20261010 …
AMD and Intel announced that oneDNN will be the default open-source library for x86 acceleration inside PyTorch, with AMD folding its ZenDNN optimizations upstream into oneDNN instead of maintaining a separate fork. AMD …
AWS and Keyfactor published a joint reference architecture for giving AI agents cryptographic identity using X.509 public key infrastructure, built on Keyfactor's EJBCA Enterprise certificate authority and AWS services i…
Platformatic released sqlite-js-stream, a TypeScript port of Litestream that runs directly on Node using the built-in node:sqlite module and Node's native type stripping, installable via npm install sqlite-js-stream. The…
A developer at Trida AI argues that an agent is best understood as a model plus a harness, where the harness is the runtime that assembles context, brokers tool calls, sandboxes execution, persists state, and recovers fr…
The Center for Cybersecurity Policy and Law (CCPL) published a report, "Cyber Insurance in the Age of AI: Why Underwriters Should Care About Phishing-Resistant Authentication," co-authored by Davis Hake and Luke O'Grady,…
A hobbyist developer reported running a 35 billion parameter model at 17.3 tokens per second on a $599 Mac Mini M4, and cut classification latency from 8.5 seconds to 1.9 seconds — a 4.4x speedup — by swapping the classi…
Anyscale and Google Cloud announced native Ray support for PyTorch on Google TPUs on October 20, 2026, letting GKE users schedule distributed PyTorch training, vLLM inference and reinforcement learning workloads across G…
The Financial Stability Board published its final report, "Sound Practices for Responsible Adoption of Artificial Intelligence," on October 15, 2026, setting out 12 key practices for responsible AI deployment for G20 Fin…
René Coignard released Clippy, a Rust-based open-source project that brings the Microsoft Office Assistant character to the macOS desktop, licensed under GPLv3 or later. The repository builds a native Clippy.app via `car…
The maintainer of GEO Doctor released an MIT-licensed checker that diagnoses why AI search crawlers cannot read a website, testing four conditions: robots.txt rules blocking search crawlers such as OAI-SearchBot and Perp…
Elon Musk announced on October 20, 2026 that X updated its timeline recommendation system with a new AI/SI ranking model that predicts dwell time and a retrieval model that surfaces posts users read to completion. The ch…
A developer built RepoPilot, a serverless agent on AWS Lambda and Bedrock that scores public GitHub repositories on a 0-100 risk scale before cloning, flagging dangerous install steps and returning a sanitized setup path…
Inferact and Google released vLLM TorchTPU (vllm-torchtpu), an out-of-tree vLLM platform plugin that lets most models already supported by vLLM run on Google Cloud TPUs without changes to model code. The backend builds o…
Neurophilosophy and psychopathology researcher Cristina Gherghel published a Zenodo preprint, "Prior-Data Gravity and Framework Reabsorption: A Behavioral Theory of Novelty Displacement in Large Language Models," arguing…