YouTube video covers GPT-6 and GPT Sites changes

HuatuoGPT-3-27B is a medical LLM built on Qwen3.8-27B using One-stage Policy Optimization (OnePO), which adapts models to medicine in a single reinforcement-learning stage without a preceding domain-specific stage.

The DoD budget request would fund "Polygraph+" over five years, adding AI/ML scoring algorithms and "standoff sensing" that reads physiology without attaching a device. Run by the Defense Counterintelligence and Security Agency, it targets vetting and insider threat detection; Congress has not approved it.

A Reddit user announced CatBench at catbench.noip.ws, pitched as a safer alternative to AssBench for testing LLM content restrictions. No methodology, model results, or scoring details were provided.

A Reddit user ran Qwen3.8 27B on an RTX 5090 and Qwen Image 2.1 on a 10GB RTX 3080 with ComfyUI to produce practical 3D-printing work for a Bambu H2C printer, spending about a day and a half on the attempt.

OpenVINO backend for Laya runs 40 ms per question on CPU, a 3.4x speedup over PyTorch. Code and a Flappy Bird demo are published on GitHub.