RoguePilot flaw let GitHub Copilot leak GITHUB_TOKEN, while new studies expose LLM side channels, ShadowLogic backdoors, and promptware risks.

Enter large language model (LLM) evaluation. The purpose of LLM evaluation is to analyze and refine GenAI outputs to improve their accuracy and reliability while avoiding bias. The evaluation process ...

Researchers from the University of Maryland, Lawrence Livermore, Columbia and TogetherAI have developed a training technique that triples LLM inference speed without auxiliary models or infrastructure ...

The company open sourced an 8-billion-parameter LLM, Steerling-8B, trained with a new architecture designed to make its ...

Exposed endpoints quietly expand attack surfaces across LLM infrastructure. Learn why endpoint privilege management is important to AI security.

You can even self-host it!

As AI deployments scale and start to include packs of agents autonomously working in concert, organizations face a naturally amplified attack surface.

AI agents are powerful, but without a strong control plane and hard guardrails, they’re just one bad decision away from chaos.

Abstract: Large Language Models (LLMs) are widely adopted for automated code generation with promising results. Although prior research has assessed LLM-generated code and identified various quality ...

Despite near-perfect exam scores, large language models falter when real people rely on them for medical advice, exposing a critical gap between AI knowledge and safe patient decision-making. Study: ...

Multilingual coding and tool use see boosts, with support for agent teams in Claude Code's research preview for parallel workflows. Product integrations expand its reach: an upgraded Claude in Excel ...

Share your favorite time-tested AI prompts and coding workflows from the Prompt Potluck @ Homebrew: Coding Edition. This repository collects prompts, templates, debugging tactics, evaluation ...