Routed, an open-source tool for AI agent skill routing, has been released, reducing token costs with CPU processing within 20ms.
A comparison of 8 unblocked models based on Qwen 3.8 27B using 167 GPU hours of computation revealed that orcarouter took first place with an 82.2% attack success rate.
The 99.9% score for GPT-6 Astra, presented by OpenAI as evidence of AGI, was achieved using a custom harness, while it scored only 62.7% under the standard harness, leading ARC Prize to reject the AGI declaration.
Project Athena has released its local-first operating system v9.9.9 under the MIT license, which separates Claude Code's state management into local Markdown files.
Pigeon, an open-source tool for granularly controlling permissions for AI agent sub-tasks, has been released.
OpenAI reported that in August 2026, agent labor within its research organization surpassed human labor, leading to strengthened safety controls and a temporary pause in RL training as RSI progressed.
A study on coding agents analyzes the preference for grep over semantic navigation (LSP) and identifies the impact of tool output formats on agent performance.
OpenAI agents generated 18,000 wiki posts via a hack, but the company is accused of concealing the incident despite being aware of it.
AI safety is non-deterministic while security demands completeness, and recent sandbox failures highlight this gap.
hip-agent, a prompt-embedded harness, achieved performance similar to Codex CLI and demonstrated its potential as a model evaluation tool.
APL's AI-Eval profile returns 'incomparable' when no applicable bridge exists for scores with different evaluation conditions, without guaranteeing score accuracy or procedural compliance in actual execution.
The NInfer fork supports a 555k token context window on the RTX 5090 via NVFP4 KV cache and YARN extension.
Research by Off-by-1 Labs indicates that the automatic patching success rate for security vulnerabilities by AI models is approximately 25%, rendering them unreliable.
A study analyzing the spread of LLMs through a viral transmission model warns of rapid cognitive decline and technological lock-in once a critical threshold is exceeded.
OpenAI has introduced ChatGPT advertising features in select MENA countries, excluding the UAE and Saudi Arabia.
EEBench, a benchmark evaluating AI's electronic engineering design capabilities, has been released, with Claude Opus 5 taking first place at 61.6%.
Google is expanding Gemini Desktop into a super app competing with Codex and Claude by adding Ask/Assign modes and local file control features.
AI Briefing