AgentsAugust 27, 2026via Wired AI

OpenAI Is Developing a ‘Persistent’ AI Agent

Why it matters

OpenAI is building persistent agent capabilities into Codex—autonomous systems that continue executing tasks proactively rather than responding to individual requests. This signals a shift toward always-on agent architectures and raises questions about control, cost, and deployment patterns for enterprise AI.

Key signals

  • OpenAI developing 'persistent' agent feature for Codex
  • Agent continues working proactively until 'put to sleep'
  • Code review evidence from WIRED
  • Suggests shift from request-response to continuous-autonomy model
  • Published Aug 27, 2026

The hook

OpenAI's next move: agents that keep working until you tell them to stop.

Code reviewed by WIRED reveals the company is developing a feature that enables Codex to continue working proactively until it is “put to sleep.”

The week's key stories, every Friday.

For practitioners and enthusiasts — free, in your inbox.

Free forever. No spam.