[OpenAI Tests Codex Persistent Mode: Tasks Completed but Can Continue Following Up] Beating AI Newsflash: OpenAI is testing Codex's Persistent mode, which allows continuous operation. Normally, an agent stops after completing a task, but this mode enables Codex to continue checking and handling subsequent matters after finishing the current task. Relevant code has already appeared in Codex's public repository, but OpenAI stated there are no plans for a rollout in the near future. [Original Link]