Artificial Intelligence

OpenAI says roughly 30% of SWE-Bench Pro’s public coding tasks are defective, forcing engineering teams to rethink how they evaluate AI coding agents.

A newly disclosed symlink weakness across six AI coding assistants shows why agent approvals, filesystem isolation and patch discipline now belong in the DevOps threat model.

Europe’s new advanced-AI cybersecurity plan links model evaluation, structured access and secure testing. Here is what platform, cloud and DevOps teams should prepare now.

OpenAI has made GPT‑5.6 Sol, Terra and Luna generally available, adding multi-agent execution, programmatic tool calling and new caching economics for production AI teams.

OpenAI’s GPT-5.6 family is now broadly available across the API, Codex and ChatGPT. Here is what Sol, Terra and Luna change for model routing, cost controls and production AI operations.

OpenAI's ChatGPT Work launch moves GPT-5.6 agents closer to everyday enterprise workflows, raising practical governance questions for DevOps, platform and cloud teams.

Meta's public preview of the Meta Model API gives U.S. developers access to Muse Spark 1.1, adding another paid agentic coding model for teams weighing cost, routing, safety, and production governance.

OpenAI's GPT-Live brings full-duplex conversation to ChatGPT Voice, giving DevOps and platform teams a fresh set of realtime AI reliability, safety, and observability questions.

Varonis says Google has patched a Dialogflow CX weakness that could let an authorized editor abuse Code Blocks across agents in the same project. The practical lesson is least-privilege access, change review, logging and isolation for AI agents.