2 to 3 months from now, running Codex on a local computer will feel incredibly outdated. Tibo's words reveal the key evolution happening with next-gen cutting-edge AI. Thinking about OpenAI's two recent moves, the strategic roadmap behind it is becoming crystal clear: Acquiring Ona: Moving Agents completely to the cloud OpenAI's acquisition of Ona has kicked off the next phase for Codex. This means Agents are no longer limited by local hardware and battery life. Even if you close your laptop, the cloud-based Agent will continue executing tasks seamlessly. Hiring Cloud Agents Engineers OpenAI is building systems for scheduling, sandboxing, storage, authentication, and cost control for long-running Agents, aiming to achieve "large-scale Agent orchestration" across Codex, ChatGPT, and APIs. This marks a paradigm shift in AI usage: Next-gen cutting-edge models will no longer rely on your laptop as a computational carrier. Instead, they’ll require a cloud-based, temporary organization to accomplish complex goals—independent compute nodes, clearly defined Agent roles, shared memory repositories, tool access permissions, and uninterrupted autonomous execution for days on end. The laptop of the future won’t be a local machine running AI but rather a control screen for orchestrating a temporary AI company in the cloud.