OpenAI's Persistent AI Agent: Always-On Codex

OpenAI is developing a proactive, highly persistent version of its AI agent, Codex. This new "Persistent mode" setting has appeared in the command line version of Codex code. OpenAI confirmed they are testing this feature, but there are no immediate plans for a broad launch. When Persistent mode is selected, Codex will "continue working until put to sleep." This differs from current modes that stop after a few minutes or hours, even if a task is incomplete. A feature within Persistent mode, called "proactivity," instructs the agent to create follow-up tasks for itself after answering a user's request. It can work on these tasks across sessions and use past interactions and "knowledge of the user." The agent can also message the user sparingly without being asked. The instructions also limit the agent, stating that Persistent mode does not expand what it can do and requires user approval for changes outside the user's system. This development shows an effort to create AI agents that can automate tasks across work and personal lives, potentially impacting how we interact with AI.