OpenAI Unveils Independent Cursor and Background Tasks for Developers.

OpenAI Power-ups Codex: Introducing Background Multitasking and Persistent Workflows

OpenAI has announced a major update to Codex, significantly expanding its ability to interact with computer environments. This update focuses on enhancing developer productivity by allowing the AI to handle complex, long-running tasks without interrupting the user’s primary workspace.

Seamless Background Automation

One of the standout features is Codex’s new ability to operate in the background. The AI now utilizes a dedicated, independent cursor, allowing it to run test scripts, manage environments, or execute deployments while the user continues to work on other tasks. This "dual-cursor" approach effectively turns Codex into a true digital co-pilot that can multitask alongside the developer.

Contextual Memory and Scheduled Tasks

The updated Codex now features Preference Awareness, allowing it to remember:

• Selected Tech Stacks: Your preferred frameworks and libraries.

• Custom Workflows: Specific sequences of operations tailored to your project.

• Resume Capability: The ability to pick up exactly where it left off after an interruption or system restart.

• Task Scheduling: Users can now set advanced schedules for automated jobs to run at specific times.

Furthermore, OpenAI has integrated the gpt-image-1.5 model into the workflow, providing high-fidelity image simulation and generation whenever a visual component is required during the development or testing process.

Availability and Phased Rollout

The update is rolling out starting today for ChatGPT users. However, the Computer Control features are initially exclusive to macOS. Deployment for Enterprise and Education sectors, as well as users in the European Union (EU) and the United Kingdom (UK), will follow at a later date due to regulatory and compliance scheduling.

Having an independent cursor is a key aspect of the agentic workflow concept. It transcends from being just a chatbot to becoming a "virtual worker" capable of handling tedious tasks like regression testing or organizing large documents while you're in a meeting.

The fact that the Computer Control feature started on macOS is likely due to the deep collaboration between OpenAI and Apple in leveraging macOS's secure and stable Screen Accessibility API. This allows Codex to accurately "see" and control the screen without compromising overall system security.

The upgrade to version 1.5 of the visualization tool isn't just about aesthetics; it's about UI/UX mockup simulation. Developers can instruct Codex to "create a new login screen following this project's guidelines," and Codex will instantly simulate it within the workflow, bridging the gap between design and coding.

The reason the EU and UK have to wait is due to the strict AI Act regarding AI's direct control of computers (autonomous systems). OpenAI needs to improve its security and authentication systems. (Human-in-the-loop) must comply with the legal standards of those regions first.

 

 

Now Creating AI Images Directly from Your Google Photos. 

 

Source: OpenAI 

💬 AI Content Assistant

Ask me anything about this article. No data is stored for your question.

OpenAI Power-ups Codex: Introducing Background Multitasking and Persistent Workflows

OpenAI has announced a major update to Codex, significantly expanding its ability to interact with computer environments. This update focuses on enhancing developer productivity by allowing the AI to handle complex, long-running tasks without interrupting the user’s primary workspace.

Seamless Background Automation

One of the standout features is Codex’s new ability to operate in the background. The AI now utilizes a dedicated, independent cursor, allowing it to run test scripts, manage environments, or execute deployments while the user continues to work on other tasks. This "dual-cursor" approach effectively turns Codex into a true digital co-pilot that can multitask alongside the developer.

Contextual Memory and Scheduled Tasks

The updated Codex now features Preference Awareness, allowing it to remember:

• Selected Tech Stacks: Your preferred frameworks and libraries.

• Custom Workflows: Specific sequences of operations tailored to your project.

• Resume Capability: The ability to pick up exactly where it left off after an interruption or system restart.

• Task Scheduling: Users can now set advanced schedules for automated jobs to run at specific times.

Furthermore, OpenAI has integrated the gpt-image-1.5 model into the workflow, providing high-fidelity image simulation and generation whenever a visual component is required during the development or testing process.

Availability and Phased Rollout

The update is rolling out starting today for ChatGPT users. However, the Computer Control features are initially exclusive to macOS. Deployment for Enterprise and Education sectors, as well as users in the European Union (EU) and the United Kingdom (UK), will follow at a later date due to regulatory and compliance scheduling.

Having an independent cursor is a key aspect of the agentic workflow concept. It transcends from being just a chatbot to becoming a "virtual worker" capable of handling tedious tasks like regression testing or organizing large documents while you're in a meeting.

The fact that the Computer Control feature started on macOS is likely due to the deep collaboration between OpenAI and Apple in leveraging macOS's secure and stable Screen Accessibility API. This allows Codex to accurately "see" and control the screen without compromising overall system security.

The upgrade to version 1.5 of the visualization tool isn't just about aesthetics; it's about UI/UX mockup simulation. Developers can instruct Codex to "create a new login screen following this project's guidelines," and Codex will instantly simulate it within the workflow, bridging the gap between design and coding.

The reason the EU and UK have to wait is due to the strict AI Act regarding AI's direct control of computers (autonomous systems). OpenAI needs to improve its security and authentication systems. (Human-in-the-loop) must comply with the legal standards of those regions first.

 

 

Now Creating AI Images Directly from Your Google Photos. 

 

Source: OpenAI