AI agents and LLM-powered XR prototyping tools unveiled

OpenAI released a white‑paper outlining how to build business‑focused AI agents that use large language models (LLMs) to execute autonomous workflows. The guide defines agents, highlights use cases such as complex decision‑making and unstructured data handling, and details the three core components—model, tools, and prompts—required for construction, orchestration, and safe operation.

Google announced a new "Vibe Coding XR" workflow that pairs its Gemini LLM with the web‑based XR Blocks framework to translate natural‑language descriptions into functional Android XR applications in under a minute. The system leverages Gemini’s long‑context reasoning and specialized system prompts to automatically configure spatial logic, perception, and interaction. A live demo will be shown at the Google booth during ACM CHI 2026, and developers can test prototypes on headsets or via a desktop simulator.