AI news for platform & agent engineers
Ranked signal · finite reading
One shared ranking. Scan what changed, save what matters, and stop when the finish line appears.
Loading...
The first known runaway AI agent - or a very bad marketing stunt? Martin Alderson's commentary on the OpenAI accidental cyberattack against Hugging Face includes a couple of details I hadn't considered. First, Hugging... Context & related coverage →
We worked with Andon Labs on Drone-Bench, a new benchmark testing whether AI models can autonomously fly a drone to locate and follow a person. Context & related coverage →
We revamped how we benchmark Deep Agents. Here's the eval setup we run in Harbor across coding, conversation, and retrieval, and how we use it to ship changes. Context & related coverage →
Modern AI agents rely on elaborate inference harnesses such as Claude Code, Codex, and OpenClaw to drive multi-turn reasoning, tool use, and access to external systems. While powerful, these complex harnesses also mak... Context & related coverage →
Large Language Models (LLMs) excel at natural language understanding and generation but remain unreliable for multi-step logical reasoning, especially in safety-critical or compliance-sensitive domains. Recent neuro-s... Context & related coverage →
Production AI agents' failures are less often due to an inability to reason well and more often because they cannot manage what is in their reasoning context: conversation histories, large prompts, large tool definiti... Context & related coverage →
Opus 5 is a step change improvement for the Opus tier powering long-running agents while delivering improvements in coding and professional work. Context & related coverage →
Cognitive impairment (CI) is a growing public health concern. Early and accurate diagnosis is critical for enabling timely intervention and improving patient outcomes. Speech-based CI detection has emerged as a promis... Context & related coverage →
Added Claude Opus 5 ( claude-opus-5 ), now the default Opus model — 1M context, fast mode at $10/$50 per Mtok · Added sandbox.network.strictAllowlist setting to deny non-allowlisted hosts for sandboxed commands withou... Context & related coverage →
Fixed premature stdin closure when background tasks are in flight : query() no longer closes stdin on the first result frame when background tasks (e.g. run_... · Updated bundled Claude CLI to version 2.1.219 Context & related coverage →
This post covers Opus 5’s improvements and practical guidance for AI engineers integrating the model into agentic systems and production inference workloads on Amazon Bedrock. See the documentation for Claude Platform... Context & related coverage →
NVIDIA, Microsoft Lead Push for Open AI Models After Kimi K3 Launch Yahoo! Finance Canada Context & related coverage →