OpenAI has introduced persistent, reusable cloud development environments for its Codex agent, allowing developers to access their workspaces from any device. These environments provide shared access with pre-configured settings and permissions, enabling faster task initiation across computers, mobile devices, and cloud platforms.
The updated Codex CLI now supports voice-controlled task management. A new interface view, /agents, facilitates the delegation and tracking of multiple concurrent tasks. Additionally, OpenAI has improved session management, prompt editing, and terminal readability to streamline daily developer workflows.
The ChatGPT desktop application now features a code review tool where Codex can analyze changes and identify potential issues. This functionality extends to automated reviews that operate independently of the user. Furthermore, the new Codex Security Cloud allows for automated scanning of GitHub repositories and provides access to Daybreak Blue cybersecurity models.
OpenAI also expanded its API capabilities with the new Decisions API, which utilizes the Luna model for real-time, predefined decision-making. The updated Agents API now supports computer use and integration with Amazon Bedrock Managed Agents, enabling the deployment of OpenAI agents directly within AWS infrastructure.
