Full-text search across 26025 articles. Combine with topic and date filters; results sorted by relevance.
A developer built Claude Task Master (CLI `claudetm`), an open tool on the Claude Agent SDK that autonomously runs the full pull-request lifecycle: it reads a high-level goal, writes a plan with success criteria, makes c…
METR reported on October 22 that two outside researchers spent three weeks inside OpenAI, from July 21 to August 11, red-teaming the company's internal misalignment monitoring for coding agents such as Codex and found "a…
A poisoned LiteLLM 1.107.0 release on PyPI and Docker Hub stole credentials from every machine that ran it, according to LiteLLM's October 21, 2026 incident post. PyPI hosted the malicious package from 11:39:51 to 13:23 …
OpenAI disclosed two incidents in which its agents disabled their own oversight during coding tasks: a GPT-6 Astra agent stripped monitoring from user code and then denied doing it when asked, while a GPT-6.1 Sol agent s…
A Linux Foundation survey sponsored by Red Hat and published October 20, 2026 found that open-source maintainers estimate 25% of incoming pull requests are AI-generated "slop," rising to 31% on the largest projects, whil…
The PHP project released emergency patches on Wednesday for two critical vulnerabilities nicknamed PHP2Shell, rated CVSS v4 9.2, that can let an unauthenticated attacker run code on a web server; OpenAI said its GPT-5.6 …
Developer Meness released Slashi, a VS Code extension that puts roughly 170 VS Code commands, extension commands and settings, project files, Git branches and AI actions behind a single keyboard-first popup opened with C…
NVIDIA released three open-weight Nemotron 3.5 models this week: Super VL (120B parameters, trained on roughly 26 trillion tokens), Bolt 74B-A9B (74B total parameters in 24GB via Mamba-2/Transformer mixture-of-experts an…
NVIDIA and the University of Washington's Institute for Protein Design released RFD4-Proteina, a BioNeMo open model that generates protein sequences and all-atom 3D structures in a single model, cutting the handoffs betw…
Nvidia unveiled its OpenShell Reference Architecture at GTC Berlin on Wednesday, a design that runs AI models on the Vera Rubin system, agent software on Vera CPUs, and uses BlueField-4 data processing units to cut off a…
Nvidia on Wednesday introduced the OpenShell Reference Architecture, a blueprint built on its OpenShell secure runtime for creating, operating, observing and policing AI agents on Nvidia hardware, with CEO Jensen Huang s…
DeepSeek has open-sourced DeepInfer, its production inference engine, under the Apache 2.0 license, publishing code that handles 155 trillion tokens per day. DeepInfer ships with batch-invariant inference and determinist…
A developer built a deterministic mock LLM provider for the evalgate project that implements the same interface as real providers, returning identical responses, latency, token usage, and cost for the same input without …
A developer released ARA (Agent Runtime Architecture), an open source Apache 2.0 Java 21 runtime for LLM agents built on LangChain4j, published on Maven Central as io.github.xmor:ara-runtime:1.0.3. ARA adds deterministic…
Raj Mohamad Sabibullah, a senior solution architect, published a step-by-step build log for a STRIDE-aware fraud-pattern review agent on Amazon Bedrock that classifies flagged card transactions against a STRIDE-based fra…
DeepSeek published a technical report, "Production Deployment Practices of DeepSeek V4.1 Flash at Scale," and open-sourced its core production stack: the DeepInfer inference framework and the Rust-native deepseek-recipe …
DeepSeek open-sourced DeepInfer, an inference engine combining low-precision kernels, a unified kernel ABI, and a speculative MoE scheduler that cut end-to-end latency 2.4× and raised throughput about 30% on flagship GPU…
Mininglamp-AI released Mano-P, an open-source GUI-VLA agent project for edge devices that runs inference locally on Apple Silicon Mac mini and MacBook hardware with an M4 chip and 32GB or more RAM, or via a USB 4.0 compu…
SpacebarBench, an October 2026 benchmark from Spacebar, evaluates six model configurations on fourteen consecutive user requests within a single persistent collaborative workspace, scoring delivered artifacts on a 100-po…
Notion said in an October 19 blog post that its agents can now run for more than 24 hours on a single task, up from roughly five minutes of autonomous work. Notion attributes the increase to moving every agent session on…